status-go/multiaccounts/migrations/bindata.go

536 lines
26 KiB
Go
Raw Normal View History

2022-03-24 09:35:56 +00:00
// Code generated by go-bindata. DO NOT EDIT.
// sources:
2022-03-24 09:35:56 +00:00
// 0001_accounts.down.sql (21B)
// 0001_accounts.up.sql (163B)
// 1605007189_identity_images.down.sql (29B)
// 1605007189_identity_images.up.sql (268B)
// 1606224181_drop_photo_path_from_accounts.down.sql (892B)
// 1606224181_drop_photo_path_from_accounts.up.sql (866B)
// 1648646095_image_clock.down.sql (939B)
// 1648646095_image_clock.up.sql (69B)
// 1649317600_add_color_hash.up.sql (201B)
// 1660238799_accounts_kdf.up.sql (115B)
2023-03-22 17:48:42 +00:00
// 1679505708_add_customization_color.up.sql (78B)
2022-03-24 09:35:56 +00:00
// doc.go (74B)
package migrations
import (
"bytes"
"compress/gzip"
2022-03-24 09:35:56 +00:00
"crypto/sha256"
"fmt"
"io"
"io/ioutil"
"os"
"path/filepath"
"strings"
"time"
)
func bindataRead(data []byte, name string) ([]byte, error) {
gz, err := gzip.NewReader(bytes.NewBuffer(data))
if err != nil {
return nil, fmt.Errorf("read %q: %v", name, err)
}
var buf bytes.Buffer
_, err = io.Copy(&buf, gz)
clErr := gz.Close()
if err != nil {
return nil, fmt.Errorf("read %q: %v", name, err)
}
if clErr != nil {
return nil, err
}
return buf.Bytes(), nil
}
type asset struct {
2022-03-24 09:35:56 +00:00
bytes []byte
info os.FileInfo
digest [sha256.Size]byte
}
type bindataFileInfo struct {
name string
size int64
mode os.FileMode
modTime time.Time
}
func (fi bindataFileInfo) Name() string {
return fi.name
}
func (fi bindataFileInfo) Size() int64 {
return fi.size
}
func (fi bindataFileInfo) Mode() os.FileMode {
return fi.mode
}
func (fi bindataFileInfo) ModTime() time.Time {
return fi.modTime
}
func (fi bindataFileInfo) IsDir() bool {
return false
}
func (fi bindataFileInfo) Sys() interface{} {
return nil
}
var __0001_accountsDownSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x72\x09\xf2\x0f\x50\x08\x71\x74\xf2\x71\x55\x48\x4c\x4e\xce\x2f\xcd\x2b\x29\xb6\xe6\x02\x04\x00\x00\xff\xff\x96\x1e\x13\xa1\x15\x00\x00\x00")
func _0001_accountsDownSqlBytes() ([]byte, error) {
return bindataRead(
__0001_accountsDownSql,
"0001_accounts.down.sql",
)
}
func _0001_accountsDownSql() (*asset, error) {
bytes, err := _0001_accountsDownSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "0001_accounts.down.sql", size: 21, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0xd2, 0x61, 0x4c, 0x18, 0xfc, 0xc, 0xdf, 0x5c, 0x1f, 0x5e, 0xd3, 0xbd, 0xfa, 0x12, 0x5e, 0x8d, 0x8d, 0x8b, 0xb9, 0x5f, 0x99, 0x46, 0x63, 0xa5, 0xe3, 0xa6, 0x8a, 0x4, 0xf1, 0x73, 0x8a, 0xe9}}
return a, nil
}
var __0001_accountsUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x1c\xcc\xb1\x6e\x83\x30\x14\x46\xe1\xdd\x4f\xf1\x8f\xad\xe4\x37\xe8\x64\xa8\x5b\xae\x42\x00\x99\x4b\x80\xd1\x02\x04\x56\x82\x8d\xc0\x19\xf2\xf6\x51\x58\x8f\xf4\x9d\xd4\x68\xc5\x1a\xac\x92\x5c\x83\xfe\x50\x94\x0c\xdd\x51\xcd\x35\xec\x30\x84\xa7\x8f\x07\xbe\xc4\x7d\x7a\x35\x6e\xc4\x4d\x99\x34\x53\x06\x95\xa1\xab\x32\x3d\x2e\xba\x97\xc2\xdb\x75\x02\xeb\x8e\x4f\x5b\x34\x79\x2e\xc5\x23\xcc\xce\xb3\x5b\xa7\x23\xda\x75\x43\x42\xff\xa0\x82\xa5\xd8\x96\x10\x43\x65\xe3\x72\x02\xf9\xf9\x0e\x76\x1f\x2b\xeb\x76\xe7\xe7\x33\x8a\x6f\xb4\xc4\x59\xd9\x30\x4c\xd9\xd2\xef\x8f\x78\x07\x00\x00\xff\xff\xab\xcf\xa2\xbd\xa3\x00\x00\x00")
func _0001_accountsUpSqlBytes() ([]byte, error) {
return bindataRead(
__0001_accountsUpSql,
"0001_accounts.up.sql",
)
}
func _0001_accountsUpSql() (*asset, error) {
bytes, err := _0001_accountsUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "0001_accounts.up.sql", size: 163, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0xf2, 0xfa, 0x99, 0x8e, 0x96, 0xb3, 0x13, 0x6c, 0x1f, 0x6, 0x27, 0xc5, 0xd2, 0xd4, 0xe0, 0xa5, 0x26, 0x82, 0xa7, 0x26, 0xf2, 0x68, 0x9d, 0xed, 0x9c, 0x3d, 0xbb, 0xdc, 0x37, 0x28, 0xbc, 0x1}}
return a, nil
}
var __1605007189_identity_imagesDownSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x72\x09\xf2\x0f\x50\x08\x71\x74\xf2\x71\x55\xc8\x4c\x49\xcd\x2b\xc9\x2c\xa9\x8c\xcf\xcc\x4d\x4c\x4f\x2d\xb6\xe6\xe5\x02\x04\x00\x00\xff\xff\xa1\x22\x72\x37\x1d\x00\x00\x00")
func _1605007189_identity_imagesDownSqlBytes() ([]byte, error) {
return bindataRead(
__1605007189_identity_imagesDownSql,
"1605007189_identity_images.down.sql",
)
}
func _1605007189_identity_imagesDownSql() (*asset, error) {
bytes, err := _1605007189_identity_imagesDownSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1605007189_identity_images.down.sql", size: 29, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0x2f, 0xcf, 0xa7, 0xae, 0xd5, 0x4f, 0xcd, 0x14, 0x63, 0x9, 0xbe, 0x39, 0x49, 0x18, 0x96, 0xb2, 0xa3, 0x8, 0x7d, 0x41, 0xdb, 0x50, 0x5d, 0xf5, 0x4d, 0xa2, 0xd, 0x8f, 0x57, 0x79, 0x77, 0x67}}
return a, nil
}
var __1605007189_identity_imagesUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x5c\xce\xc1\x6a\xc3\x30\x10\x04\xd0\xbb\xc1\xff\x30\xc7\x04\xf2\x07\x3d\xc9\xaa\x42\x44\x55\x29\x28\x4a\xd3\x9c\x84\x40\x5b\x7b\x69\xe2\x96\x58\xa5\xb8\x5f\x5f\xea\xfa\x60\x72\xdc\xc7\xec\x30\xd2\x2b\x11\x14\x82\x68\x8c\x82\xde\xc2\xba\x00\xf5\xaa\x0f\xe1\x00\xce\xd4\x17\x2e\x63\xe4\x6b\x6a\x69\x58\xd5\x15\x00\xbc\xd3\x18\xbf\x38\xe3\x45\x78\xb9\x13\x7e\xf3\xaf\x7d\xba\xd2\x1d\x4d\x5f\xf1\x33\x8d\x97\x8f\x94\xd1\x18\xd7\x4c\xe5\xf6\x68\xcc\x9c\xf8\xe6\x5c\x3a\x70\x5f\xe6\xbb\x23\x6e\xbb\xb2\x80\x37\xbe\x50\x1c\xf8\x87\x16\x76\xa3\x3f\x88\x25\xdd\x5a\x5a\x66\xf7\x5e\x3f\x0b\x7f\xc6\x93\x3a\x63\x35\x8f\xdc\x4c\xbb\xd6\x70\x16\xd2\xd9\xad\xd1\x32\xc0\xab\xbd\x11\x52\xd5\xd5\x1a\x27\x1d\x76\xee\x18\xe0\xdd\x49\x3f\x3e\xd4\xd5\x6f\x00\x00\x00\xff\xff\x8c\x6a\x0a\x57\x0c\x01\x00\x00")
func _1605007189_identity_imagesUpSqlBytes() ([]byte, error) {
return bindataRead(
__1605007189_identity_imagesUpSql,
"1605007189_identity_images.up.sql",
)
}
func _1605007189_identity_imagesUpSql() (*asset, error) {
bytes, err := _1605007189_identity_imagesUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1605007189_identity_images.up.sql", size: 268, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0x50, 0xb6, 0xc1, 0x5c, 0x76, 0x72, 0x6b, 0x22, 0x34, 0xdc, 0x96, 0xdc, 0x2b, 0xfd, 0x2d, 0xbe, 0xcc, 0x1e, 0xd4, 0x5, 0x93, 0xd, 0xc2, 0x51, 0xf3, 0x1a, 0xef, 0x2b, 0x26, 0xa4, 0xeb, 0x65}}
return a, nil
}
var __1606224181_drop_photo_path_from_accountsDownSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\xd4\x93\xc1\x6e\x9c\x30\x18\x84\xef\x3c\xc5\x1c\xdb\x95\x25\x1e\x60\x4f\x04\x9c\xc4\x2a\x0b\xc8\x98\x66\x73\x6a\x1c\xb0\x82\x15\xb0\x11\x6b\x54\xed\xdb\x57\x6b\x58\x76\x8b\xd4\xaa\xd7\x5e\xff\xf9\x7f\xcf\x68\x3e\x39\xdc\x21\xb6\xc3\x19\xae\x55\x90\x75\x6d\x27\xe3\x4e\x70\xf2\xbd\x53\xd0\xc6\x59\x48\x38\xd5\x0f\xf3\x84\x80\x1e\xe3\xb4\x4a\xa8\xdf\x7e\xd3\x8d\x32\x4e\xd7\xd6\xbc\xa1\xb6\xdd\xd4\x1b\x48\xd3\x80\x65\x77\x2b\x43\x6b\x9d\x2d\xa4\x6b\xd7\x95\x5d\x18\xc4\x9c\x46\x82\x42\xd0\x43\x91\xf3\x88\xbf\x42\x44\x0f\x29\x5d\xdd\x7f\xbc\xcb\xfa\x73\x1a\xbe\x04\x00\xf0\xa9\xce\x95\x6e\xf0\x3d\xe2\xf1\x73\xc4\x51\x70\x76\xb8\x5c\x7c\xa3\xaf\xc4\xeb\x46\xf6\x0a\x82\x1e\x05\xb2\x5c\x20\xab\xd2\x74\x9e\x77\xf6\x43\x1b\xa1\x7b\x75\x72\xb2\x1f\xf0\xc0\x9e\xc0\x32\x31\x6b\x6b\x28\x7f\x48\xae\x3e\xb5\x1c\x9b\x42\xea\x51\x9b\x0f\x2f\x04\x5f\xf1\xc2\xc4\x73\x5e\x09\xf0\xfc\x85\x25\xfb\x80\x65\x25\xe5\xe2\xf2\x50\xbe\x4d\x8b\x92\xa6\x34\x16\x4b\x5e\xe2\x73\x91\x4d\x0a\x82\xb5\x31\xb2\x35\x7c\xe4\xf9\x61\x7d\x73\x1f\x04\xe1\x0e\xc9\x68\x07\xdf\xa2\xed\x9a\x2d\x9a\x4b\xd1\xa3\xaa\x47\x25\x9d\xc2\x4f\xed\x5a\xc8\xae\x5b\x2a\x3e\x2d\x94\x58\xf6\xf4\x1b\xa3\x1b\x1c\xaf\xdc\xa1\xd9\x85\x41\xc2\xf3\x62\xc3\x61\xbf\x82\xf2\x63\xf6\xe8\x2b\xa6\x47\x56\x8a\xf2\x96\xe7\x3f\xc1\xf4\x6f\x7c\x56\xcb\xbf\xf3\x59\x98\xcf\x98\x84\x6e\xce\x98\x06\x82\xe6\xca\xeb\xf6\x61\xfe\xd0\xec\xf5\xfe\x57\x00\x00\x00\xff\xff\x72\x67\x20\xf9\x7c\x03\x00\x00")
func _1606224181_drop_photo_path_from_accountsDownSqlBytes() ([]byte, error) {
return bindataRead(
__1606224181_drop_photo_path_from_accountsDownSql,
"1606224181_drop_photo_path_from_accounts.down.sql",
)
}
func _1606224181_drop_photo_path_from_accountsDownSql() (*asset, error) {
bytes, err := _1606224181_drop_photo_path_from_accountsDownSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1606224181_drop_photo_path_from_accounts.down.sql", size: 892, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0x90, 0x24, 0x17, 0x7, 0x80, 0x93, 0x6f, 0x8d, 0x5d, 0xaa, 0x8c, 0x79, 0x15, 0x5d, 0xb3, 0x19, 0xd7, 0xd8, 0x39, 0xf9, 0x3a, 0x63, 0x8f, 0x81, 0x15, 0xb6, 0xd6, 0x9a, 0x37, 0xa8, 0x8e, 0x9b}}
return a, nil
}
var __1606224181_drop_photo_path_from_accountsUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\xc4\x93\xc1\x8e\x9b\x30\x18\x84\xef\x3c\xc5\x1c\x5b\x64\x89\x07\xc8\x89\x05\x77\xd7\x2a\x01\xe4\x35\xdd\xdd\x53\x71\xc0\x0a\x56\xc0\x46\xc4\xa8\xca\xdb\x57\x01\x42\x54\xd4\x56\xb9\xed\xf5\x1f\xff\xff\x8c\xe6\x93\x03\x1f\x91\xed\x2f\x70\x8d\x82\xac\x2a\x3b\x1a\x77\x86\x93\x87\x56\x41\x1b\x67\x21\xe1\x54\xd7\xcf\x13\x02\xfa\x1e\x25\x45\x4c\xa7\xd7\x65\xdf\x58\x67\x73\xe9\x9a\x12\xd2\xd4\x60\xe9\xac\x95\xba\x56\xc6\xe9\xca\x9a\x12\x95\x6d\xc7\xce\xc0\x0f\xbc\x88\xd3\x50\x50\x08\xba\xcf\x33\x1e\xf2\x0f\x88\xf0\x29\xa1\xab\xe5\xcf\x83\xac\x4e\x63\xff\xc5\x03\x80\x93\xba\x14\xba\xc6\x8f\x90\x47\x2f\x21\x47\xce\xd9\xfe\xba\xf1\x9d\x7e\x90\x49\x37\xb2\x53\x10\xf4\x5d\x20\xcd\x04\xd2\x22\x49\xe6\x79\x6b\x8f\xda\x08\xdd\xa9\xb3\x93\x5d\x8f\x27\xf6\x0c\x96\x8a\x59\x5b\x43\x4d\x8b\xe4\xe6\x53\xc9\xa1\xce\xa5\x1e\xb4\x39\x4e\x82\xf7\x15\x6f\x4c\xbc\x64\x85\x00\xcf\xde\x58\xbc\xf3\x58\xfa\x4a\xb9\xb8\x1e\xca\xb6\x69\xf1\x4a\x13\x1a\x89\x25\x2f\x99\x72\x91\x4d\x0a\x82\xb5\x26\xb2\x35\xfc\xc6\xb3\xfd\x7a\x73\xe7\x79\x81\x8f\x78\xb0\xfd\xd4\xae\x6d\xeb\x2d\x8f\x6b\xc9\x83\xaa\x06\x25\x9d\xc2\x2f\xed\x1a\xc8\xb6\x5d\x2a\x3e\x2f\x68\x58\xfa\xfc\x0f\x30\x93\x72\x47\xe3\x07\x5e\xcc\xb3\x7c\x83\x61\xb7\x72\xfa\x63\xfc\xf9\x58\xf0\x10\x97\xc7\x80\xac\x9e\xff\x07\xb2\x40\x9e\xb9\x08\x5d\x5f\x30\xf6\x04\xf5\x0d\xd0\xfd\x5b\xe0\xef\x5d\xde\xf6\x7f\x07\x00\x00\xff\xff\xc4\x05\x28\x49\x62\x03\x00\x00")
func _1606224181_drop_photo_path_from_accountsUpSqlBytes() ([]byte, error) {
return bindataRead(
__1606224181_drop_photo_path_from_accountsUpSql,
"1606224181_drop_photo_path_from_accounts.up.sql",
)
}
func _1606224181_drop_photo_path_from_accountsUpSql() (*asset, error) {
bytes, err := _1606224181_drop_photo_path_from_accountsUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1606224181_drop_photo_path_from_accounts.up.sql", size: 866, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0xff, 0x4c, 0x97, 0xee, 0xef, 0x82, 0xb8, 0x6c, 0x71, 0xbb, 0x50, 0x7b, 0xe6, 0xd9, 0x22, 0x31, 0x7c, 0x1a, 0xfe, 0x91, 0x28, 0xf6, 0x6, 0x36, 0xe, 0xb1, 0xf1, 0xc8, 0x25, 0xac, 0x7e, 0xd6}}
return a, nil
}
var __1648646095_image_clockDownSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\xdc\x92\xbf\x6e\xdb\x30\x18\xc4\x77\x3e\xc5\x8d\x49\x40\x20\x0f\xe0\x49\x56\x68\x84\x28\x2d\xaa\x14\xd3\x34\x93\xc0\x48\x5f\x25\x22\xfa\x07\x89\x42\xe0\x3e\x7d\x61\xd9\xb0\x6b\xa3\xf5\xd6\xa5\xe3\x77\x3c\x92\x77\xfc\xf1\xf1\x01\xd9\x57\xe5\x03\xa1\xec\x69\x42\xd7\x07\x4c\xf3\x30\xf4\x63\x40\x39\xf6\xc3\xe0\xbb\x0a\x45\xdf\xcc\x6d\x37\x71\xd4\xd4\x15\x84\x4f\x42\x3b\x4f\x01\xc5\x48\x2e\x10\x1c\x02\xb5\x03\x82\x7b\x6f\x08\x0f\x8f\x2c\x36\x22\xb2\x02\x56\x6c\x53\x6d\x22\xf3\x06\x1b\xad\x95\x80\x2f\xa9\x0b\x3e\xec\x72\xdf\xba\x8a\xa6\xfc\xdd\x15\x1f\xf3\x70\xc7\x00\xe0\x83\x76\xf9\xec\x4b\x7c\x8b\x4c\xfc\x1c\x19\xbe\x88\x9d\x6b\xe9\x52\x59\x76\xe6\x83\xdb\x35\xbd\x2b\xb1\x56\x7a\x8d\x44\x5b\x24\x2f\x4a\x1d\x0c\x9f\xbe\x0c\x35\x7c\x17\x0e\x63\x4d\xbe\xaa\xc3\x79\xfe\xe1\x1b\xca\x27\xff\x93\xce\xd2\x48\xfb\x39\x0f\x6e\xac\xe8\x37\x67\x6a\xe4\x76\x1f\xfd\x8b\x78\xc3\xdd\x31\x1d\x5f\x12\xdd\x43\x27\x88\x75\xb2\x51\x32\xb6\x30\x22\x55\x51\x2c\xd8\x3d\x5e\xa5\x7d\xd6\x2f\x16\x46\xbf\xca\xa7\x15\x63\x32\xc9\x84\xb1\x90\x89\xd5\x7f\xa9\x8e\x4c\x28\x11\x5b\x5c\x1c\xcf\x2f\x4b\xf2\x43\x25\x7e\xac\xc2\xcf\x15\xf8\x55\xf4\x8d\xd1\xdb\xeb\x8b\x56\x8c\x3d\x19\x9d\xfe\x19\xc0\x8a\xb1\x13\xaa\x65\x5d\x6e\x96\xd7\x14\xdf\x65\x66\xb3\x6b\xf7\x7f\xcb\xe9\x16\xa8\x7f\x4f\xe8\xf8\x15\x6e\x82\x3a\x79\x7e\x05\x00\x00\xff\xff\xce\x2f\xe3\x37\xab\x03\x00\x00")
func _1648646095_image_clockDownSqlBytes() ([]byte, error) {
return bindataRead(
__1648646095_image_clockDownSql,
"1648646095_image_clock.down.sql",
)
}
func _1648646095_image_clockDownSql() (*asset, error) {
bytes, err := _1648646095_image_clockDownSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1648646095_image_clock.down.sql", size: 939, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0x4d, 0xa8, 0x1f, 0xf, 0xe0, 0xd7, 0xc9, 0x68, 0x98, 0xd8, 0x37, 0xb8, 0xba, 0x9e, 0xb2, 0x19, 0xf3, 0xc4, 0x73, 0x80, 0x3, 0x17, 0x2a, 0x53, 0x68, 0x10, 0x13, 0x54, 0x99, 0xb1, 0xf5, 0x1c}}
return a, nil
}
var __1648646095_image_clockUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x72\xf4\x09\x71\x0d\x52\x08\x71\x74\xf2\x71\x55\xc8\x4c\x49\xcd\x2b\xc9\x2c\xa9\x8c\xcf\xcc\x4d\x4c\x4f\x2d\x56\x70\x74\x71\x51\x70\xf6\xf7\x09\xf5\xf5\x53\x48\xce\xc9\x4f\xce\x56\xf0\xf4\x0b\x51\xf0\xf3\x0f\x51\xf0\x0b\xf5\xf1\x51\x70\x71\x75\x73\x0c\xf5\x09\x51\x30\xb0\xe6\x02\x04\x00\x00\xff\xff\x22\x35\x20\xbf\x45\x00\x00\x00")
func _1648646095_image_clockUpSqlBytes() ([]byte, error) {
return bindataRead(
__1648646095_image_clockUpSql,
"1648646095_image_clock.up.sql",
)
}
func _1648646095_image_clockUpSql() (*asset, error) {
bytes, err := _1648646095_image_clockUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1648646095_image_clock.up.sql", size: 69, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0x98, 0xa6, 0xa4, 0x4e, 0x4e, 0xca, 0x17, 0x56, 0xea, 0xfb, 0xf0, 0xa9, 0x81, 0x95, 0xe, 0x80, 0x52, 0x1, 0x47, 0x9b, 0xde, 0x14, 0xfa, 0x72, 0xc9, 0x62, 0x6f, 0x24, 0xa2, 0xc, 0x32, 0x50}}
return a, nil
}
var __1649317600_add_color_hashUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x72\xf4\x09\x71\x0d\x52\x08\x71\x74\xf2\x71\x55\x48\x4c\x4e\xce\x2f\xcd\x2b\x29\x56\x70\x74\x71\x51\x70\xf6\xf7\x09\xf5\xf5\x53\x48\xce\xcf\xc9\x2f\xf2\x48\x2c\xce\x50\x08\x71\x8d\x08\x51\xf0\xf3\x0f\x51\xf0\x0b\xf5\xf1\x51\x70\x71\x75\x73\x0c\xf5\x09\x51\x50\x52\xb2\xe6\x22\xca\x0c\xcf\x14\x05\x4f\x3f\x2c\x06\x18\x58\x73\x85\x06\xb8\x38\x86\x20\x69\x0d\x76\x0d\x41\xb2\xd7\x16\x6c\x07\x4e\x35\x9e\x29\x0a\xb6\x20\x43\x00\x01\x00\x00\xff\xff\xfa\xaf\xaf\xd9\xc9\x00\x00\x00")
func _1649317600_add_color_hashUpSqlBytes() ([]byte, error) {
return bindataRead(
__1649317600_add_color_hashUpSql,
"1649317600_add_color_hash.up.sql",
)
}
func _1649317600_add_color_hashUpSql() (*asset, error) {
bytes, err := _1649317600_add_color_hashUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1649317600_add_color_hash.up.sql", size: 201, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0x1a, 0xf, 0x37, 0x6d, 0xcf, 0x99, 0xc9, 0x2e, 0xdc, 0x70, 0x11, 0xb4, 0x36, 0x26, 0x4f, 0x39, 0xa8, 0x44, 0xf, 0xcb, 0xcc, 0x81, 0x74, 0x7a, 0x88, 0xaa, 0x54, 0x8c, 0xc4, 0xe, 0x56, 0x4f}}
return a, nil
}
var __1660238799_accounts_kdfUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x72\xf4\x09\x71\x0d\x52\x08\x71\x74\xf2\x71\x55\x48\x4c\x4e\xce\x2f\xcd\x2b\x29\x56\x70\x74\x71\x51\x70\xf6\xf7\x09\xf5\xf5\x53\xc8\x4e\x49\xf3\x2c\x49\x2d\x4a\x2c\xc9\xcc\xcf\x2b\x56\xf0\xf4\x0b\x51\xf0\xf3\x0f\x51\xf0\x0b\xf5\xf1\x51\x70\x71\x75\x73\x0c\xf5\x09\x51\x30\x36\x32\x30\xb0\xe6\x0a\x0d\x70\x71\x0c\x41\x32\x23\xd8\x35\x04\x4d\xb3\x2d\x54\x25\x20\x00\x00\xff\xff\x37\x9c\xbc\xd5\x73\x00\x00\x00")
func _1660238799_accounts_kdfUpSqlBytes() ([]byte, error) {
return bindataRead(
__1660238799_accounts_kdfUpSql,
"1660238799_accounts_kdf.up.sql",
)
}
func _1660238799_accounts_kdfUpSql() (*asset, error) {
bytes, err := _1660238799_accounts_kdfUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1660238799_accounts_kdf.up.sql", size: 115, mode: os.FileMode(0644), modTime: time.Unix(1664889752, 0)}
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0xdf, 0xe6, 0x7a, 0x69, 0x25, 0x42, 0x3b, 0x9c, 0x20, 0xf5, 0xcb, 0xae, 0xb0, 0xb3, 0x1b, 0x66, 0xc2, 0x5d, 0xd0, 0xc1, 0x59, 0xe8, 0xa9, 0xc5, 0x69, 0x58, 0x8f, 0xae, 0xe6, 0xd1, 0x4c, 0x53}}
return a, nil
}
2023-03-22 17:48:42 +00:00
var __1679505708_add_customization_colorUpSql = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x72\xf4\x09\x71\x0d\x52\x08\x71\x74\xf2\x71\x55\x48\x4c\x4e\xce\x2f\xcd\x2b\x29\x56\x70\x74\x71\x51\x70\xf6\xf7\x09\xf5\xf5\x53\x48\x2e\x2d\x2e\xc9\xcf\xcd\xac\x4a\x2c\xc9\xcc\xcf\x73\xce\xcf\xc9\x2f\x52\x08\x73\x0c\x72\xf6\x70\x0c\x52\x70\x71\x75\x73\x0c\xf5\x09\x51\x50\x2a\x28\xca\xcc\x4d\x2c\xaa\x54\xb2\xe6\x02\x04\x00\x00\xff\xff\x08\xb6\x89\xf4\x4e\x00\x00\x00")
func _1679505708_add_customization_colorUpSqlBytes() ([]byte, error) {
return bindataRead(
__1679505708_add_customization_colorUpSql,
"1679505708_add_customization_color.up.sql",
)
}
func _1679505708_add_customization_colorUpSql() (*asset, error) {
bytes, err := _1679505708_add_customization_colorUpSqlBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "1679505708_add_customization_color.up.sql", size: 78, mode: os.FileMode(0644), modTime: time.Unix(1680025955, 0)}
2023-03-22 17:48:42 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0xa9, 0xe1, 0x3d, 0xaa, 0x5d, 0x35, 0x87, 0x8a, 0x8b, 0xe9, 0x4a, 0xa6, 0x7b, 0x85, 0xbc, 0x33, 0x11, 0xc7, 0x7d, 0x61, 0xac, 0x65, 0x59, 0xda, 0x32, 0x59, 0x68, 0x9d, 0xa1, 0x10, 0x7b, 0xa9}}
return a, nil
}
var _docGo = []byte("\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\xff\x2c\xc9\xb1\x0d\xc4\x20\x0c\x05\xd0\x9e\x29\xfe\x02\xd8\xfd\x6d\xe3\x4b\xac\x2f\x44\x82\x09\x78\x7f\xa5\x49\xfd\xa6\x1d\xdd\xe8\xd8\xcf\x55\x8a\x2a\xe3\x47\x1f\xbe\x2c\x1d\x8c\xfa\x6f\xe3\xb4\x34\xd4\xd9\x89\xbb\x71\x59\xb6\x18\x1b\x35\x20\xa2\x9f\x0a\x03\xa2\xe5\x0d\x00\x00\xff\xff\x60\xcd\x06\xbe\x4a\x00\x00\x00")
func docGoBytes() ([]byte, error) {
return bindataRead(
_docGo,
"doc.go",
)
}
func docGo() (*asset, error) {
bytes, err := docGoBytes()
if err != nil {
return nil, err
}
URL unfurling (initial implementation) (#3471) This is the initial implementation for the new URL unfurling requirements. The most important one is that only the message sender will pay the privacy cost for unfurling and extracting metadata from websites. Once the message is sent, the unfurled data will be stored at the protocol level and receivers will just profit and happily decode the metadata to render it. Further development of this URL unfurling capability will be mostly guided by issues created on clients. For the moment in status-mobile: https://github.com/status-im/status-mobile/labels/url-preview - https://github.com/status-im/status-mobile/issues/15918 - https://github.com/status-im/status-mobile/issues/15917 - https://github.com/status-im/status-mobile/issues/15910 - https://github.com/status-im/status-mobile/issues/15909 - https://github.com/status-im/status-mobile/issues/15908 - https://github.com/status-im/status-mobile/issues/15906 - https://github.com/status-im/status-mobile/issues/15905 ### Terminology In the code, I've tried to stick to the word "unfurl URL" to really mean the process of extracting metadata from a website, sort of lower level. I use "link preview" to mean a higher level structure which is enriched by unfurled data. "link preview" is also how designers refer to it. ### User flows 1. Carol needs to see link previews while typing in the chat input field. Notice from the diagram nothing is persisted and that status-go endpoints are essentially stateless. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_getTextURLs Server-->>Client: Normalized URLs Client->>Client: Render cached unfurled URLs Client->>Server: Unfurl non-cached URLs.\nCall wakuext_unfurlURLs Server->>Website: Fetch metadata Website-->>Server: Metadata (thumbnail URL, title, etc) Server->>Website: Fetch thumbnail Server->>Website: Fetch favicon Website-->>Server: Favicon bytes Website-->>Server: Thumbnail bytes Server->>Server: Decode & process images Server-->>Client: Unfurled data (thumbnail data URI, etc) #+end_src ``` ``` ,------. ,------. ,-------. |Client| |Server| |Website| `--+---' `--+---' `---+---' | Call wakuext_getTextURLs | | | ---------------------------------------> | | | | | Normalized URLs | | | <- - - - - - - - - - - - - - - - - - - - | | | | |----. | | | | Render cached unfurled URLs | | |<---' | | | | | | Unfurl non-cached URLs. | | | Call wakuext_unfurlURLs | | | ---------------------------------------> | | | | | | Fetch metadata | | | ------------------------------------> | | | | | Metadata (thumbnail URL, title, etc)| | | <- - - - - - - - - - - - - - - - - - | | | | | Fetch thumbnail | | | ------------------------------------> | | | | | Fetch favicon | | | ------------------------------------> | | | | | Favicon bytes | | | <- - - - - - - - - - - - - - - - - - | | | | | Thumbnail bytes | | | <- - - - - - - - - - - - - - - - - - | | | | |----. | | | | Decode & process images | | |<---' | | | | | Unfurled data (thumbnail data URI, etc)| | | <- - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,---+---. |Client| |Server| |Website| `------' `------' `-------' ``` 2. Carol sends the text message with link previews in the RPC request wakuext_sendChatMessages. status-go assumes the link previews are good because it can't and shouldn't attempt to re-unfurl them. ``` #+begin_src plantuml :results verbatim Client->>Server: Call wakuext_sendChatMessages Server->>Server: Transform link previews to\nbe proto-marshalled Server->DB: Write link previews serialized as JSON Server-->>Client: Updated message response #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | Call wakuext_sendChatMessages| | | -----------------------------> | | | | | |----. | | | | Transform link previews to | | |<---' be proto-marshalled | | | | | | | | | Write link previews serialized as JSON| | | --------------------------------------> | | | | Updated message response | | | <- - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` 3. The message was sent over waku and persisted locally in Carol's device. She should now see the link previews in the chat history. There can be many link previews shared by other chat members, therefore it is important to serve the assets via the media server to avoid overloading the ReactNative bridge with lots of big JSON payloads containing base64 encoded data URIs (maybe this concern is meaningless for desktop). When a client is rendering messages with link previews, they will have the field linkPreviews, and the thumbnail URL will point to the local media server. ``` #+begin_src plantuml :results verbatim Client->>Server: GET /link-preview/thumbnail (media server) Server->>DB: Read from user_messages.unfurled_links Server->Server: Unmarshal JSON Server-->>Client: HTTP Content-Type: image/jpeg/etc #+end_src ``` ``` ,------. ,------. ,--. |Client| |Server| |DB| `--+---' `--+---' `+-' | GET /link-preview/thumbnail (media server)| | | ------------------------------------------> | | | | | | Read from user_messages.unfurled_links| | | --------------------------------------> | | | | |----. | | | | Unmarshal JSON | | |<---' | | | | | HTTP Content-Type: image/jpeg/etc | | | <- - - - - - - - - - - - - - - - - - - - - | ,--+---. ,--+---. ,+-. |Client| |Server| |DB| `------' `------' `--' ``` ### Some limitations of the current implementation The following points will become separate issues in status-go that I'll work on over the next couple weeks. In no order of importance: - Improve how multiple links are fetched; retries on failure and testing how unfurling behaves around the timeout limits (deterministically, not by making real HTTP calls as I did). https://github.com/status-im/status-go/issues/3498 - Unfurl favicons and store them in the protobuf too. - For this PR, I added unfurling support only for websites with OpenGraph https://ogp.me/ meta tags. Other unfurlers will be implemented on demand. The next one will probably be for oEmbed https://oembed.com/, the protocol supported by YouTube, for example. - Resize and/or compress thumbnails (and favicons). Often times, thumbnails are huge for the purposes of link previews. There is already support for compressing JPEGs in status-go, but I prefer to work with compression in a separate PR because I'd like to also solve the problem for PNGs (probably convert them to JPEGs, plus compress them). This would be a safe choice for thumbnails, favicons not so much because transparency is desirable. - Editing messages is not yet supported. - I haven't coded any artificial limit on the number of previews or on the size of the thumbnail payload. This will be done in a separate issue. I have heard the ideal solution may be to split messages into smaller chunks of ~125 KiB because of libp2p, but that might be too complicated at this stage of the product (?). - Link preview deletion. - For the moment, OpenGraph metadata is extracted by requesting data for the English language (and fallback to whatever is available). In the future, we'll want to unfurl by respecting the user's local device language. Some websites, like GoDaddy, are already localized based on the device's IP, but many aren't. - The website's description text should be limited by a certain number of characters, especially because it's outside our control. Exactly how much has not been decided yet, so it'll be done separately. - URL normalization can be tricky, so I implemented only the basics to help with caching. For example, the url https://status.im and HTTPS://status.im are considered identical. Also, a URL is considered valid for unfurling if its TLD exists according to publicsuffix.EffectiveTLDPlusOne. This was essential, otherwise the default Go url.Parse approach would consider many invalid URLs valid, and thus the server would waste resources trying to unfurl the unfurleable. ### Other requirements - If the message is edited, the link previews should reflect the edited text, not the original one. This has been aligned with the design team as well. - If the website's thumbnail or the favicon can't be fetched, just ignore them. The only mandatory piece of metadata is the website's title and URL. - Link previews in clients should be generated in near real-time, that is, as the user types, previews are updated. In mobile this performs very well, and it's what other clients like WhatsApp, Telegram, and Facebook do. ### Decisions - While the user typing in the input field, the client is constantly (debounced) asking status-go to parse the text and extract normalized URLs and then the client checks if they're already in its in-memory cache. If they are, no RPC call is made. I chose this approach to achieve the best possible performance in mobile and avoid the whole RPC overhead, since the chat experience is already not smooth enough. The mobile client uses URLs as cache keys in a hashmap, i.e. if the key is present, it means the preview is readily available (naive, but good enough for now). This decision also gave me more flexibility to find the best UX at this stage of the feature. - Due to the requirement that users should be able to see independent loading indicators for each link preview, when status-go can't unfurl a URL, it doesn't return it in the response. - As an initial implementation, I added the BLOB column unfurled_links to the user_messages table. The preview data is then serialized as JSON before being stored in this column. I felt that creating a separate table and the related code for this initial PR would be inconvenient. Is that reasonable to you? Once things stabilize I can create a proper table if we want to avoid this kind of solution with serialized columns.
2023-05-18 18:43:06 +00:00
info := bindataFileInfo{name: "doc.go", size: 74, mode: os.FileMode(0644), modTime: time.Unix(1662753054, 0)}
2022-03-24 09:35:56 +00:00
a := &asset{bytes: bytes, info: info, digest: [32]uint8{0xde, 0x7c, 0x28, 0xcd, 0x47, 0xf2, 0xfa, 0x7c, 0x51, 0x2d, 0xd8, 0x38, 0xb, 0xb0, 0x34, 0x9d, 0x4c, 0x62, 0xa, 0x9e, 0x28, 0xc3, 0x31, 0x23, 0xd9, 0xbb, 0x89, 0x9f, 0xa0, 0x89, 0x1f, 0xe8}}
return a, nil
}
// Asset loads and returns the asset for the given name.
// It returns an error if the asset could not be found or
// could not be loaded.
func Asset(name string) ([]byte, error) {
2022-03-24 09:35:56 +00:00
canonicalName := strings.Replace(name, "\\", "/", -1)
if f, ok := _bindata[canonicalName]; ok {
a, err := f()
if err != nil {
return nil, fmt.Errorf("Asset %s can't read by error: %v", name, err)
}
return a.bytes, nil
}
return nil, fmt.Errorf("Asset %s not found", name)
}
2022-03-24 09:35:56 +00:00
// AssetString returns the asset contents as a string (instead of a []byte).
func AssetString(name string) (string, error) {
data, err := Asset(name)
return string(data), err
}
// MustAsset is like Asset but panics when Asset would return an error.
// It simplifies safe initialization of global variables.
func MustAsset(name string) []byte {
a, err := Asset(name)
if err != nil {
panic("asset: Asset(" + name + "): " + err.Error())
}
return a
}
2022-03-24 09:35:56 +00:00
// MustAssetString is like AssetString but panics when Asset would return an
// error. It simplifies safe initialization of global variables.
func MustAssetString(name string) string {
return string(MustAsset(name))
}
// AssetInfo loads and returns the asset info for the given name.
// It returns an error if the asset could not be found or
// could not be loaded.
func AssetInfo(name string) (os.FileInfo, error) {
2022-03-24 09:35:56 +00:00
canonicalName := strings.Replace(name, "\\", "/", -1)
if f, ok := _bindata[canonicalName]; ok {
a, err := f()
if err != nil {
return nil, fmt.Errorf("AssetInfo %s can't read by error: %v", name, err)
}
return a.info, nil
}
return nil, fmt.Errorf("AssetInfo %s not found", name)
}
2022-03-24 09:35:56 +00:00
// AssetDigest returns the digest of the file with the given name. It returns an
// error if the asset could not be found or the digest could not be loaded.
func AssetDigest(name string) ([sha256.Size]byte, error) {
canonicalName := strings.Replace(name, "\\", "/", -1)
if f, ok := _bindata[canonicalName]; ok {
a, err := f()
if err != nil {
return [sha256.Size]byte{}, fmt.Errorf("AssetDigest %s can't read by error: %v", name, err)
}
return a.digest, nil
}
return [sha256.Size]byte{}, fmt.Errorf("AssetDigest %s not found", name)
}
// Digests returns a map of all known files and their checksums.
func Digests() (map[string][sha256.Size]byte, error) {
mp := make(map[string][sha256.Size]byte, len(_bindata))
for name := range _bindata {
a, err := _bindata[name]()
if err != nil {
return nil, err
}
mp[name] = a.digest
}
return mp, nil
}
// AssetNames returns the names of the assets.
func AssetNames() []string {
names := make([]string, 0, len(_bindata))
for name := range _bindata {
names = append(names, name)
}
return names
}
// _bindata is a table, holding each asset generator, mapped to its name.
var _bindata = map[string]func() (*asset, error){
"0001_accounts.down.sql": _0001_accountsDownSql,
"0001_accounts.up.sql": _0001_accountsUpSql,
"1605007189_identity_images.down.sql": _1605007189_identity_imagesDownSql,
"1605007189_identity_images.up.sql": _1605007189_identity_imagesUpSql,
"1606224181_drop_photo_path_from_accounts.down.sql": _1606224181_drop_photo_path_from_accountsDownSql,
"1606224181_drop_photo_path_from_accounts.up.sql": _1606224181_drop_photo_path_from_accountsUpSql,
"1648646095_image_clock.down.sql": _1648646095_image_clockDownSql,
"1648646095_image_clock.up.sql": _1648646095_image_clockUpSql,
"1649317600_add_color_hash.up.sql": _1649317600_add_color_hashUpSql,
"1660238799_accounts_kdf.up.sql": _1660238799_accounts_kdfUpSql,
2023-03-22 17:48:42 +00:00
"1679505708_add_customization_color.up.sql": _1679505708_add_customization_colorUpSql,
"doc.go": docGo,
}
// AssetDir returns the file names below a certain
// directory embedded in the file by go-bindata.
// For example if you run go-bindata on data/... and data contains the
// following hierarchy:
// data/
// foo.txt
// img/
// a.png
// b.png
2022-03-24 09:35:56 +00:00
// then AssetDir("data") would return []string{"foo.txt", "img"},
// AssetDir("data/img") would return []string{"a.png", "b.png"},
// AssetDir("foo.txt") and AssetDir("notexist") would return an error, and
// AssetDir("") will return []string{"data"}.
func AssetDir(name string) ([]string, error) {
node := _bintree
if len(name) != 0 {
2022-03-24 09:35:56 +00:00
canonicalName := strings.Replace(name, "\\", "/", -1)
pathList := strings.Split(canonicalName, "/")
for _, p := range pathList {
node = node.Children[p]
if node == nil {
return nil, fmt.Errorf("Asset %s not found", name)
}
}
}
if node.Func != nil {
return nil, fmt.Errorf("Asset %s not found", name)
}
rv := make([]string, 0, len(node.Children))
for childName := range node.Children {
rv = append(rv, childName)
}
return rv, nil
}
type bintree struct {
Func func() (*asset, error)
Children map[string]*bintree
}
2022-03-24 09:35:56 +00:00
var _bintree = &bintree{nil, map[string]*bintree{
"0001_accounts.down.sql": &bintree{_0001_accountsDownSql, map[string]*bintree{}},
"0001_accounts.up.sql": &bintree{_0001_accountsUpSql, map[string]*bintree{}},
"1605007189_identity_images.down.sql": &bintree{_1605007189_identity_imagesDownSql, map[string]*bintree{}},
"1605007189_identity_images.up.sql": &bintree{_1605007189_identity_imagesUpSql, map[string]*bintree{}},
"1606224181_drop_photo_path_from_accounts.down.sql": &bintree{_1606224181_drop_photo_path_from_accountsDownSql, map[string]*bintree{}},
"1606224181_drop_photo_path_from_accounts.up.sql": &bintree{_1606224181_drop_photo_path_from_accountsUpSql, map[string]*bintree{}},
"1648646095_image_clock.down.sql": &bintree{_1648646095_image_clockDownSql, map[string]*bintree{}},
"1648646095_image_clock.up.sql": &bintree{_1648646095_image_clockUpSql, map[string]*bintree{}},
"1649317600_add_color_hash.up.sql": &bintree{_1649317600_add_color_hashUpSql, map[string]*bintree{}},
"1660238799_accounts_kdf.up.sql": &bintree{_1660238799_accounts_kdfUpSql, map[string]*bintree{}},
2023-03-22 17:48:42 +00:00
"1679505708_add_customization_color.up.sql": &bintree{_1679505708_add_customization_colorUpSql, map[string]*bintree{}},
"doc.go": &bintree{docGo, map[string]*bintree{}},
}}
2022-03-24 09:35:56 +00:00
// RestoreAsset restores an asset under the given directory.
func RestoreAsset(dir, name string) error {
data, err := Asset(name)
if err != nil {
return err
}
info, err := AssetInfo(name)
if err != nil {
return err
}
err = os.MkdirAll(_filePath(dir, filepath.Dir(name)), os.FileMode(0755))
if err != nil {
return err
}
err = ioutil.WriteFile(_filePath(dir, name), data, info.Mode())
if err != nil {
return err
}
2022-03-24 09:35:56 +00:00
return os.Chtimes(_filePath(dir, name), info.ModTime(), info.ModTime())
}
2022-03-24 09:35:56 +00:00
// RestoreAssets restores an asset under the given directory recursively.
func RestoreAssets(dir, name string) error {
children, err := AssetDir(name)
// File
if err != nil {
return RestoreAsset(dir, name)
}
// Dir
for _, child := range children {
err = RestoreAssets(dir, filepath.Join(name, child))
if err != nil {
return err
}
}
return nil
}
func _filePath(dir, name string) string {
2022-03-24 09:35:56 +00:00
canonicalName := strings.Replace(name, "\\", "/", -1)
return filepath.Join(append([]string{dir}, strings.Split(canonicalName, "/")...)...)
}