Bitcoin ETL

Join Telegram Group

Install Bitcoin ETL:

pip install bitcoin-etl

Export blocks and transactions (Schema, Reference):

> bitcoinetl export_blocks_and_transactions --start-block 0 --end-block 500000 \
--provider-uri http://user:pass@localhost:8332 --chain bitcoin \
 --blocks-output blocks.json --transactions-output transactions.json

Supported chains:

• bitcoin
• bitcoin_cash
• bitcoin_gold
• dogecoin
• litecoin
• dash
• zcash

Stream blockchain data continually to console (Reference):

> pip install bitcoin-etl[streaming]
> bitcoinetl stream -p http://user:pass@localhost:8332 --start-block 500000

Stream blockchain data continually to Google Pub/Sub (Reference):

> export GOOGLE_APPLICATION_CREDENTIALS=/path_to_credentials_file.json
> bitcoinetl stream -p http://user:pass@localhost:8332 --start-block 500000 --output projects/your-project/topics/crypto_bitcoin

For the latest version, check out the repo and call

> pip install -e .[streaming]
> python bitcoinetl.py

Table of Contents

• Bitcoin ETL
  • Table of Contents
  • Schema
    • blocks.json
    • transactions.json
    • transaction_input
    • transaction_output
  • Exporting the Blockchain
    • Running in Docker
    • Command Reference
      • export_blocks_and_transactions
      • enrich_transactions
      • get_block_range_for_date
      • export_all
      • stream
    • Running Tests
    • Running Tox Tests
    • Public Datasets in BigQuery

Schema

blocks.json

transactions.json

transaction_input

transaction_output

You can find column descriptions in schemas

Notes:

1. Output values returned by Dogecoin API had precision loss in the clients prior to version 1.14.It's caused by this issue https://github.com/dogecoin/dogecoin/issues/1558The explorers that used older versions to export the data may show incorrect address balances and transaction amounts.

2. For Zcash, vjoinsplit and valueBalance fields are converted to inputs and outputs with type 'shielded'https://zcash-rpc.github.io/getrawtransaction.html, https://zcash.readthedocs.io/en/latest/rtd_pages/zips/zip-0243.html

Exporting the Blockchain

1. Install python 3.5.3+ https://www.python.org/downloads/

2. Install Bitcoin node https://hackernoon.com/a-complete-beginners-guide-to-installing-a-bitcoin-full-node-on-linux-2018-edition-cb8e384479ea

3. Start Bitcoin.Make sure it downloaded the blocks that you need by executing $ bitcoin-cli getblockchaininfo in the terminal.You can export blocks below blocks, there is no need to wait until the full sync

4. Install Bitcoin ETL:

   > pip install bitcoin-etl

5. Export blocks & transactions:

   > bitcoinetl export_all --start 0 --end 499999  \
   --partition-batch-size 100 \
   --provider-uri http://user:pass@localhost:8332 --chain bitcoin

   The result will be in the output subdirectory, partitioned in Hive style:

   output/blocks/start_block=00000000/end_block=00000099/blocks_00000000_00000099.csv
   output/blocks/start_block=00000100/end_block=00000199/blocks_00000100_=00000199.csv
   ...
   output/transactions/start_block=00000000/end_block=00000099/transactions_00000000_00000099.csv
   ...

   In case bitcoinetl command is not available in PATH, use python -m bitcoinetl instead.

Running in Docker

1. Install Docker https://docs.docker.com/install/

2. Build a docker image

   > docker build -t bitcoin-etl:latest .
   > docker image ls

3. Run a container out of the image

   > docker run -v $HOME/output:/bitcoin-etl/output bitcoin-etl:latest export_blocks_and_transactions --start-block 0 --end-block 500000 \
       --provider-uri http://user:pass@localhost:8332 --blocks-output output/blocks.json --transactions-output output/transactions.json

4. Run streaming to console or Pub/Sub

   > docker build -t bitcoin-etl:latest-streaming -f Dockerfile_with_streaming .
   > echo "Stream to console"
   > docker run bitcoin-etl:latest-streaming stream -p http://user:pass@localhost:8332 --start-block 500000
   > echo "Stream to Pub/Sub"
   > docker run -v /path_to_credentials_file/:/bitcoin-etl/ --env GOOGLE_APPLICATION_CREDENTIALS=/bitcoin-etl/credentials_file.json bitcoin-etl:latest-streaming stream -p http://user:pass@localhost:8332 --start-block 500000 --output projects/your-project/topics/crypto_bitcoin

5. Refer to https://github.com/blockchain-etl/bitcoin-etl-streaming for deploying the streaming app toGoogle Kubernetes Engine.

Command Reference

• export_blocks_and_transactions
• enrich_transactions
• get_block_range_for_date
• export_all
• stream

All the commands accept -h parameter for help, e.g.:

> bitcoinetl export_blocks_and_transactions --help
Usage: bitcoinetl.py export_blocks_and_transactions [OPTIONS]

  Export blocks and transactions.

Options:
  -s, --start-block INTEGER   Start block
  -e, --end-block INTEGER     End block  [required]
  -b, --batch-size INTEGER    The number of blocks to export at a time.
  -p, --provider-uri TEXT     The URI of the remote Bitcoin node
  -w, --max-workers INTEGER   The maximum number of workers.
  --blocks-output TEXT        The output file for blocks. If not provided
                              blocks will not be exported. Use "-" for stdout
  --transactions-output TEXT  The output file for transactions. If not
                              provided transactions will not be exported. Use
                              "-" for stdout
  --help                      Show this message and exit.

For the --output parameters the supported type is json. The format type is inferred from the output file name.

export_blocks_and_transactions

> bitcoinetl export_blocks_and_transactions --start-block 0 --end-block 500000 \
  --provider-uri http://user:pass@localhost:8332 \
  --blocks-output blocks.json --transactions-output transactions.json

Omit --blocks-output or --transactions-output options if you want to export only transactions/blocks.

You can tune --batch-size, --max-workers for performance.

Note that required_signatures, type, addresses, and value fields will be empty in transactions inputs.Use enrich_transactions to populate those fields.

enrich_transactions

You need to run bitcoin daemon with option txindex=1 for this command to work.

> bitcoinetl enrich_transactions  \
  --provider-uri http://user:pass@localhost:8332 \
  --transactions-input transactions.json --transactions-output enriched_transactions.json

You can tune --batch-size, --max-workers for performance.

get_block_range_for_date

> bitcoinetl get_block_range_for_date --provider-uri http://user:pass@localhost:8332 --date=2017-03-01

This command is guaranteed to return the block range that covers all blocks with block.time on the specifieddate. However the returned block range may also contain blocks outside the specified date, because block times are notmonotonic https://twitter.com/EvgeMedvedev/status/1073844856009576448. You can filterblocks.json/transactions.json with the below command:

> bitcoinetl filter_items -i blocks.json -o blocks_filtered.json \
-p "datetime.datetime.fromtimestamp(item['timestamp']).astimezone(datetime.timezone.utc).strftime('%Y-%m-%d') == '2017-03-01'"

export_all

> bitcoinetl export_all --provider-uri http://user:pass@localhost:8332 --start 2018-01-01 --end 2018-01-02

You can tune --export-batch-size, --max-workers for performance.

stream

> bitcoinetl stream --provider-uri http://user:pass@localhost:8332 --start-block 500000

• This command outputs blocks and transactions to the console by default.
• Use --output option to specify the Google Pub/Sub topic where to publish blockchain data,e.g. projects/your-project/topics/crypto_bitcoin. Blocks and transactions will be pushed toprojects/your-project/topics/crypto_bitcoin.blocks and projects/your-project/topics/crypto_bitcoin.transactionstopics.
• The command saves its state to last_synced_block.txt file where the last synced block number is saved periodically.
• Specify either --start-block or --last-synced-block-file option. --last-synced-block-file should point to thefile where the block number, from which to start streaming the blockchain data, is saved.
• Use the --lag option to specify how many blocks to lag behind the head of the blockchain. It's the simplest way tohandle chain reorganizations - they are less likely the further a block from the head.
• Use the --chain option to specify the type of the chain, e.g. bitcoin, litecoin, dash, zcash, etc.
• You can tune --period-seconds, --batch-size, --max-workers for performance.

Running Tests

> pip install -e .[dev]
> echo "The below variables are optional"
> export BITCOINETL_BITCOIN_PROVIDER_URI=http://user:pass@localhost:8332
> export BITCOINETL_LITECOIN_PROVIDER_URI=http://user:pass@localhost:8331
> export BITCOINETL_DOGECOIN_PROVIDER_URI=http://user:pass@localhost:8330
> export BITCOINETL_BITCOIN_CASH_PROVIDER_URI=http://user:pass@localhost:8329
> export BITCOINETL_DASH_PROVIDER_URI=http://user:pass@localhost:8328
> export BITCOINETL_ZCASH_PROVIDER_URI=http://user:pass@localhost:8327
> pytest -vv

Running Tox Tests

> pip install tox
> tox

Public Datasets in BigQuery

https://cloud.google.com/blog/products/data-analytics/introducing-six-new-cryptocurrencies-in-bigquery-public-datasets-and-how-to-analyze-them