nim-codex/codex/streams/storestream.nim

109 lines
3.2 KiB
Nim
Raw Permalink Normal View History

## Nim-Dagger
## Copyright (c) 2022 Status Research & Development GmbH
## Licensed under either of
## * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE))
## * MIT license ([LICENSE-MIT](LICENSE-MIT))
## at your option.
## This file may not be copied, modified, or distributed except according to
## those terms.
import std/options
import pkg/upraises
push: {.upraises: [].}
import pkg/libp2p
import pkg/chronos
import pkg/chronicles
import pkg/stew/ptrops
import ../stores
import ../manifest
import ../blocktype
import ./seekablestream
export stores, blocktype, manifest, chronos
logScope:
topics = "codex storestream"
type
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
# Make SeekableStream from a sequence of blocks stored in Manifest
# (only original file data - see StoreStream.size)
StoreStream* = ref object of SeekableStream
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
store*: BlockStore # Store where to lookup block contents
manifest*: Manifest # List of block CIDs
pad*: bool # Pad last block to manifest.blockSize?
proc new*(
T: type StoreStream,
store: BlockStore,
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
manifest: Manifest,
pad = true): T =
result = T(
store: store,
manifest: manifest,
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
pad: pad,
offset: 0)
result.initStream()
method `size`*(self: StoreStream): int =
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
bytes(self.manifest, self.pad)
proc `size=`*(self: StoreStream, size: int)
{.error: "Setting the size is forbidden".} =
discard
method atEof*(self: StoreStream): bool =
self.offset >= self.size
method readOnce*(
self: StoreStream,
pbytes: pointer,
nbytes: int): Future[int] {.async.} =
## Read `nbytes` from current position in the StoreStream into output buffer pointed by `pbytes`.
## Return how many bytes were actually read before EOF was encountered.
## Raise exception if we are already at EOF.
trace "Reading from manifest", cid = self.manifest.cid.get(), blocks = self.manifest.len
if self.atEof:
raise newLPStreamEOFError()
# The loop iterates over blocks in the StoreStream,
# reading them and copying their data into outbuf
var read = 0 # Bytes read so far, and thus write offset in the outbuf
while read < nbytes and not self.atEof:
# Compute from the current stream position `self.offset` the block num/offset to read
# Compute how many bytes to read from this block
let
blockNum = self.offset div self.manifest.blockSize
blockOffset = self.offset mod self.manifest.blockSize
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
readBytes = min([self.size - self.offset, nbytes - read, self.manifest.blockSize - blockOffset])
# Read contents of block `blockNum`
without blk =? await self.store.getBlock(self.manifest[blockNum]), error:
raise newLPStreamReadError(error)
trace "Reading bytes from store stream", blockNum, cid = blk.cid, bytes = readBytes, blockOffset
# Copy `readBytes` bytes starting at `blockOffset` from the block into the outbuf
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
if blk.isEmpty:
zeroMem(pbytes.offset(read), readBytes)
else:
copyMem(pbytes.offset(read), blk.data[blockOffset].addr, readBytes)
# Update current positions in the stream and outbuf
self.offset += readBytes
read += readBytes
return read
method closeImpl*(self: StoreStream) {.async.} =
Download files without padding (#218) The initial goal of this patch was to allow to download of a file via REST API in exactly the same size as it was uploaded, which required adding fields Chunker.offset and Manifest.originalBytes to keep that size. On top of that, we added more integrity checks to operations on Manifest, and reorganized TestNode.nim to test the actual interaction between node.store and node.retrieve operations. Note that the wire format of Manifest was changed, so we need to recreate all BlockStores. * Download without padding * Fixed chunker tests * Chunker: get rid of RabinChunker * Verify offset in the chunker tests * Use manifest.originalBytesPadded in StoreStream.size * StoreStream: replace emptyBlock with zeroMem * Manifest.bytes: compute how many bytes corresponding StoreStream(Manifest, pad) will return * Manifest: verify originalBytes and originalLen on new/encode/decode Also set originalBytes in each Manifest creation/update scenario * Manifest: comments, split code into sections * Reordered parameters to deal with int64 size in 32-bit builds * TestNode.nim: combine Store and Retrieve tests 1. Instead of copy-pasting code from node.nim, new test calls node.store() and node.retrieve() in order to check that they can correctly store and then retrieve data 2. New test compares only file contents, manifest contents considered an implementation detail 3. New test chunks at odd chunkSize=BlockSize/1.618 in order to ensure that data retrieved correctly even when buffer sizes mismatch * TestNode.nim: code refactoring * Manifest.add: one more test * Manifest.verify: return Result instead of raising Defect * Node.store: added blockSize parameter
2022-08-24 12:15:59 +00:00
trace "Closing StoreStream"
self.offset = self.size # set Eof
await procCall LPStream(self).closeImpl()