From 14dd763900450574206e7a3f53bc38b985c0ab4c Mon Sep 17 00:00:00 2001
From: Kim De Mey <kim.demey@gmail.com>
Date: Tue, 18 Jan 2022 15:08:02 +0100
Subject: [PATCH] Add populate history db option to fluffy (#929)

---
 fluffy/conf.nim        | 12 ++++++++
 fluffy/fluffy.nim      | 10 +++++--
 fluffy/populate_db.nim | 65 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 85 insertions(+), 2 deletions(-)
 create mode 100644 fluffy/populate_db.nim

diff --git a/fluffy/conf.nim b/fluffy/conf.nim
index 4ced96dc6..608c25722 100644
--- a/fluffy/conf.nim
+++ b/fluffy/conf.nim
@@ -37,6 +37,7 @@ const
 type
   PortalCmd* = enum
     noCommand
+    populateHistoryDb
 
   PortalConf* = object
     logLevel* {.
@@ -163,6 +164,17 @@ type
       defaultValue: noCommand .}: PortalCmd
     of noCommand:
       discard
+    of populateHistoryDb:
+      # Note: we could use the existing data dir here, but it would require
+      # also to properly store the network key and default use the one available
+      dbDir* {.
+        desc: "The directory of the fluffy content database"
+        defaultValue: ""
+        name: "db-dir" }: OutDir
+      dataFile* {.
+        desc: "Specify a json file with a map of k:v pairs representing BlockHash : Rlp encoded block"
+        defaultValue: ""
+        name: "data-file" }: InputFile
 
 proc parseCmdArg*(T: type enr.Record, p: TaintedString): T
     {.raises: [Defect, ConfigurationError].} =
diff --git a/fluffy/fluffy.nim b/fluffy/fluffy.nim
index 4aac4707e..6ae770b11 100644
--- a/fluffy/fluffy.nim
+++ b/fluffy/fluffy.nim
@@ -20,7 +20,7 @@ import
   ./network/state/[state_network, state_content],
   ./network/history/[history_network, history_content],
   ./network/wire/[portal_stream, portal_protocol_config],
-  ./content_db
+  "."/[content_db, populate_db]
 
 proc initializeBridgeClient(maybeUri: Option[string]): Option[BridgeClient] =
   try:
@@ -130,4 +130,10 @@ when isMainModule:
   setLogLevel(config.logLevel)
 
   case config.cmd
-  of noCommand: run(config)
+  of PortalCmd.noCommand:
+    run(config)
+  of PortalCmd.populateHistoryDb:
+    let res = populateHistoryDb(config.dbDir.string, config.dataFile.string)
+    if res.isErr():
+      fatal "Failed populating the history content db", error = $res.error
+      quit 1
diff --git a/fluffy/populate_db.nim b/fluffy/populate_db.nim
new file mode 100644
index 000000000..c6184e87a
--- /dev/null
+++ b/fluffy/populate_db.nim
@@ -0,0 +1,65 @@
+# Nimbus - Portal Network
+# Copyright (c) 2022 Status Research & Development GmbH
+# Licensed and distributed under either of
+#   * MIT license (license terms in the root directory or at https://opensource.org/licenses/MIT).
+#   * Apache v2 license (license terms in the root directory or at https://www.apache.org/licenses/LICENSE-2.0).
+# at your option. This file may not be copied, modified, or distributed except according to those terms.
+
+{.push raises: [Defect].}
+
+import
+  json_serialization, json_serialization/std/tables,
+  stew/[byteutils, io2, results],
+  eth/rlp,
+  ./content_db,
+  ./network/history/history_content
+
+# Offline tool to populate the database with the current existing json files
+# with block data. Might move to some other storage format later on. Perhaps
+# https://github.com/status-im/nimbus-eth2/blob/stable/docs/e2store.md can be
+# interesting.
+
+type
+  BlockData = object
+    rlp: string
+    number: uint64
+
+  BlockDataTable = Table[string, BlockData]
+
+proc populateHistoryDb*(dbDir: string, dataFile: string): Result[void, string] =
+  let db = ContentDB.new(dbDir)
+
+  let blockData = readAllFile(dataFile)
+  if blockData.isErr(): # TODO: map errors
+    return err("Failed reading data-file")
+
+  let decoded =
+    try:
+      Json.decode(blockData.get(), BlockDataTable)
+    except CatchableError as e:
+      return err("Failed decoding json data-file: " & e.msg)
+
+  # This is definitely the slowest part because of the hashing that happens in
+  # toContentId()
+  for k,v in decoded:
+    try:
+      var rlp = rlpFromHex(v.rlp)
+
+      if rlp.enterList():
+        # List that contains 3 items: Block header, body and receipts.
+        # Only store block header for now.
+        # When we want others, can use `rlp.skipElem()` and `rlp.rawData()`.
+
+        # Prepare content key
+        var blockHash: BlockHash
+        blockHash.data = hexToByteArray[sizeof(BlockHash)](k)
+
+        let contentKey = ContentKey(
+          contentType: blockHeader,
+          blockHeaderKey: ContentKeyType(chainId: 1'u16, blockHash: blockHash))
+
+        db.put(contentKey.toContentId(), rlp.rawData())
+    except CatchableError as e:
+      return err("Failed decoding block hash or data: " & e.msg)
+
+  ok()