Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/emeraldpay/dshackle-archive

ETL for Bitcoin and Ethereum data
https://github.com/emeraldpay/dshackle-archive

analysis bigdata bitcoin ethereum etl

Last synced: about 1 year ago
JSON representation

ETL for Bitcoin and Ethereum data

Awesome Lists containing this project

README 

          
= Dshackle Archive

:version: 0.2.0

:version-short: 0.2



image:https://github.com/emeraldpay/dshackle-archive/workflows/Tests/badge.svg["Unit Tests"]

image:https://codecov.io/gh/emeraldpay/dshackle-archive/branch/master/graph/badge.svg["Coverage",link="https://codecov.io/gh/emeraldpay/dshackle-archive"]

image:https://img.shields.io/docker/pulls/emeraldpay/dshackle-archive?style=flat-square["Docker",link="https://hub.docker.com/r/emeraldpay/dshackle-archive"]

image:https://img.shields.io/github/license/emeraldpay/dshackle-archive.svg?style=flat-square&maxAge=2592000["License",link="https://github.com/emeraldpay/dshackle-archive/blob/master/LICENSE"]

image:https://img.shields.io/discord/1107840420240707704?style=flat-square[Discord,link="https://discord.gg/k9HpF9Jqee"]



Dshackle Archive is a tool that efficiently extracts blockchain data in JSON format and archives it into Avro files for scalable analysis using traditional Big Data tools. It supports Bitcoin and Ethereum-compatible blockchains and can operate in both batch mode for historical data and streaming mode for real-time data extraction.



Dshackle Archive copies JSON data from a blockchain to a plain files archive

(i.e., it's the _Extraction_ of data, as in ETL).

The archive is Avro files which contain blocks and transaction details fetched from blockchain nodes via their API.

The Avro contain the JSON responses _as is_.



Features:



- Extracts data from *Bitcoin and Ethereum* compatible blockchains.

- Produces data in *Avro format*, where one file could be a _range_ of blocks or a _separate_ file per each block

- Archival can be scaled by using multiple blockchain nodes by using https://github.com/emeraldpay/dshackle[Dshackle Load Balancer].

- Runs in batch archive mode, when _historical data_ is archived; or in a _streaming mode_ when only fresh data is added to the archive

- Archive to a Filesystem, Google Storage or S3

- Produces Notification via Google PubSub or Apache Pulsar



The general idea it to use Dshackle Archive to copy data from a blockchain to a plain files, keeping the data structures as is, and then use traditional Big Data tools (Spark, Beam, etc.) to analyse data in a scalable way.



NOTE: It takes several weeks to archive the whole blockchain. It's a very IO intensive operation, but can be scaled by using multiple nodes.



NOTE: It uses https://github.com/emeraldpay/dshackle[Dshackle] protocol to connect to the blockchain, not a plain HTTP/JSON RPC. So a Dshackle instance is required to be running.



== Usage



=== Command Line Options



----

usage: dshackle-archive [options] 

Copy blockchain data into files for further analysis

    --auth.aws.accessKey     AWS / S3 Access Key

    --auth.aws.secretKey     AWS / S3 Secret Key

    --auth.gcp               Path to GCP Authentication JSON

    --aws.endpoint           AWS / S3 endpoint url instead of the

                                  default one

    --aws.region             AWS / S3 region ID to use for requests

    --aws.s3.pathStyle            Enable S3 Path Style access (default is

                                  false)

    --aws.trustTls                Trust any TLS certificate for AWS / S3

                                  (default is false)

 -b,--blockchain             Blockchain

    --back                        Apply the blocks range (--range option)

                                  back from to the current blockchain

                                  height, i.e. process N..M not starting

                                  not from zero by from the current height

 -c,--connection             Connection (host:port)

    --compact.forks               Should accept all blocks including forks

                                  in stream compaction

    --compact.ranges              Process range files also in stream

                                  compaction

    --connection.notls            Disable TLS

    --connection.timeout     Timeout (in seconds) to get data from

                                  blockchain before retrying. Default: 60

                                  seconds

    --continue                    Continue from the last file if set

 -d,--dir                    Target directory

    --deduplicate                 Deduplicate transactions and blocks

                                  (could increase memory footprint)

    --dirBlocks              How many blocks keep per subdirectory

    --dryRun                      Do not modify the storage

 -h,--help                        Show Help

 -i,--inputs                 Input File(s). Accepts a glob pattern

                                  for filename

    --include                Include optional details in JSON (trace,

                                  stateDiff)

    --notify.dir             Write notifications as JSON line to the

                                  specified dir in a file

                                  

    --notify.pubsub          Send notifications as JSON to the

                                  specified Google Pubsub topic

    --notify.pulsar.topic    Send notifications as JSON to the Pulsar

                                  to the specified topic

                                  (notify.pulsar.url must be specified)

    --notify.pulsar.url      Send notifications as JSON to the Pulsar

                                  with specified URL (notify.pulsar.topic

                                  must be specified)

    --parallel               How many blocks to request in parallel.

                                  Range: 1..512. Default: 8

    --prefix                 File prefix

 -r,--range                  Blocks Range (N...M)

    --rangeChunk             Range chunk size (default 1000)

    --tail                   Last T block to ensure are archived in

                                  streaming mode

Available commands:

 archive - the main operation, copies data from a blockchain to archive

 stream  - append fresh blocks one by one to the archive

 compact - merge individual block files into larger range files

 copy    - copy/recover from existing archive by copying into a new one

 report  - show summary on what is in archive for the specified range

 fix     - fix archive by making new archives for missing chunks

 verify  - verify that archive files contains required data and delete incomplete files

----



=== Commands



==== Archive



The main operation, copies historical data from a blockchain to archive.



The data is copied in ranges, and with default range of 1000 blocks it produces two files per range.

One for blocks in that range, and another one with _all_ transactions in all blocks in that range.



See <>.



==== Stream



Continuously append fresh blocks one by one to the archive.

In addition to the copying, Dshackle archive can be configured to notify an external system about new blocks in the archive.



Note that when it's in streaming mode the archives are writen in a per-block basis.

I.e., each block comes with in a separate pair of two files.

One for the block itself, and another one for all transactions in that block.

To merge the individual files into larger ranges use `compact` command.



To notify an external system, there are two options:



- `--notify.dir` - write notifications as JSON line to the specified dir in a file ``

- `--notify.pubsub` - send notifications as JSON to the specified Google Pubsub topic

- `--notify.pulsar.url` + `--notify.pulsar.topic` - send notifications as JSON to the specified Apache Pulsar topic



See <>.



==== Compact



Merge individual block files into larger range files.



==== Copy



Copy from one archive to another.



Technically, you can copy files as is, but the command is useful because by using it you can change the range sizes for the target archive.

Also, it can be used to recover from a corrupted archive, b/c it makes additional checks and so it skips the corrupted data.



==== Report



Show summary on what is in archive for the specified range.



==== Fix



Fixes the archive by checking if there are any missing blocks, and if so, it creates new archives for the missing blocks.



==== Verify



Verify that archive files contains required data and delete incomplete/corrupted files.

The a `fix` command is supposed to run to download missing blocks.



WARNING: This command is destructive, it deletes files from the archive.



=== Archive Size



Dshackle Archive copies and stored data as JSON responses from blockchain nodes the resulting archive is much larger that the node database size, which keeps data in a compact format.

It uses Snappy compression for Avro files, which give a good compression ratio, but still the resulting archive is large.



Average size of a 1000 blocks range (w/o expensive JSON such as `stateDiff` and `trace`):



- ~300Mb for Ethereum

- ~400Mb for Bitcoin



And the whole archive (w/o expensive JSON such as `stateDiff` and `trace`):



- ~2.5Tb for Ethereum

- ~1.9Tb for Bitcoin



=== Related projects:



- Avro structure and Java stubs: https://github.com/emeraldpay/dshackle-archive-avro

- Dshackle load balancer: https://github.com/emeraldpay/dshackle



=== Project Roadmap



- [x] support AWS S3 as a storage

- [x] support Pulsar as a notification system

- [ ] support Kafka as a notification system

- [ ] archive to Cassandra



=== FAQ



==== How to organize the data gathering process?



- First you need to archive the historical data, which may takes several week depending on how many and how fast nodes you have.

- After finishing the initial archive, you run in the Streaming mode which append new blocks to the archive as they are mined.

- Periodically (ex. once a day) you run Compaction to merge individual block files into larger range files.

- Also, periodically (ex. once a day) you run a pair of Verify and Fix commands to ensure the integrity of the archive.



==== What are supported blockchains?



Dshackle requires only compatibility onj JSON RPC level, so technically it can work with any blockchain that uses similar API.

I.e., it's compatible with all major blockchains, including Bitcoin, Ethereum, Binance Smart Chain, Polygon, etc.



==== What blockchain API it uses?



It uses https://github.com/emeraldpay/dshackle[Dshackle] protocol to connect to the blockchain, not a plain HTTP/JSON RPC.

So a Dshackle instance is required to be running.



Dshackle is a Load Balancer for Blockchain APIs, and it can route requests to multiple nodes, which scales up the archival throughput.



==== How does Dshackle Archive ensure the integrity and accuracy?



Dshackle provides two commands to ensure the integrity of the archive:



- first you run `verify` command, which checks the archive and deletes incomplete or corrupted files

- then you run `fix` command, which copies the data again for the blocks deleted in the previous step



You can schedule the execution of these commands to run periodically, e.g. once a day.

To avoid scanning the whole archive every time, you can specify a range to check, e.g. `--back --range 100...1100`.

The option above specifies that is thould verify/fix only the last 1000 blocks, starting from 100 behind the current height.

I.e., it goes backward from the current head block.



[[archive-format]]

=== Archive Format



For a complete descriptions, schema and libs to access Avro files please refer to https://github.com/emeraldpay/dshackle-archive-avro



==== Block



.Fields common between different blockchains

- `blockchainType` - _type of blockchain_, as a definitions of what fields to expect.

One of `ETHEREUM` or `BITCOIN`

- `blockchainId` - actual blockchain id (`ETH`, `BTC`, etc)

- `archiveTimestamp` - when the archive record was created.

Milliseconds since epoch

- `height` - block height

- `blockId` - block hash

- `timestamp` - block timestamp.

Milliseconds since epoch

- `parentId` - parent block hash

- `json` - JSON response for that block



.Ethereum specific fields

- `unclesCount` - number of uncles for the current block

- `uncle0Json` - JSON for first uncle (`eth_getUncleByBlockHashAndIndex(0)`)

- `uncle1Json` - JSON for second uncle (`eth_getUncleByBlockHashAndIndex(1)`)



.Bitcoin specific fields

- none



==== Transaction



.Fields common between different blockchains

- `blockchainType` - _type of blockchain_, as a definitions of what fields to expect. One of `ETHEREUM` or `BITCOIN`

- `blockchainId` - blockchain id (`ETH`, `BTC`, etc)

- `archiveTimestamp` - when the archive record was created. Milliseconds since epoch

- `height` - block height

- `blockId` - block hash

- `timestamp` - block timestamp. Milliseconds since epoch

- `index` - index of the transaction in block

- `txid` - hash or transaction id of the transaction

- `json` - JSON response for that transaction

- `raw` - raw bytes of the transaction



.Ethereum specific fields

- `from` - from address

- `to` - to address

- `receiptJson` - JSON response for `eth_getTransactionReceipt`

- `traceJson` - JSON response for `trace_replayTransaction(trace)`

- `stateDiffJson` - JSON response for `trace_replayTransaction(stateDiff)`



.Bitcoin specific fields

- none



[[notification-format]]

=== Notification format



[source, json]

----

{

  "version":"https://schema.emrld.io/dshackle-archive/notify",

  "ts":"2022-05-20T23:14:24.481327Z",

  "blockchain":"ETH",

  "type":"transactions",

  "run":"stream",

  "heightStart":14813875,

  "heightEnd":14813875,

  "location":"gs://my-bucket/blockchain-archive/eth/014000000/014813000/014813875.txes.avro"

}

----



.Where

- `version` id of the current JSON format

- `ts` timestamp of the archive event

- `blockchain` blockchain

- `type` type of file (`transactions` or `blocks`)

- `run` mode in which the Dshackle Archive is run (`archive`, `stream`, `copy` or `compact`)

- `heightStart` and `heightEnd` range of blocks in the archived files

- `location` a URL to the archived file



== Community



=== Development Chat



Join our Discord chat to discuss development and ask questions:



image:https://img.shields.io/discord/1107840420240707704?style=flat-square[Discord,link="https://discord.gg/k9HpF9Jqee"]



== Commercial Support



Want to support the project, prioritize a specific feature, or get commercial help with using Dshackle in your project?

Please contact splix@emerald.cash to discuss the possibility.



== License



Copyright 2023 EmeraldPay, Inc



Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License.

You may obtain a copy of the License at



http://www.apache.org/licenses/LICENSE-2.0



Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

See the License for the specific language governing permissions and limitations under the License.