ethereum-indexer

Personal blockchain indexer to not depend on services like The Graph, which proved to be incredibly unreliable.

Currently, only supports Ethereum mainnet.

How It Works

Extract [Raw Txn Data] All transactions for a given address are downloaded and stored in a db. This process does not terminate when it has downloaded all the data. It runs to poll for any new blocks containing our address.

Transform & Load [Maintaining State] Next, according to the rules of how to parse the above, state starts to build up. Once it catches up with the head of the blockchain, it continues to run checking in the db if there were any new raw txn data entries from the above.

Serve [Serving State] graphql server is spawned up with which you can query all the above state. Each response item will contain the block number, to indicate up to what block number the response state is valid.

Implementation Dependencies

In our implementation, we choose Covalent as the source of historical transactions pertaining to an address. The infrastructure of this code heavily depends on implementing interfaces, thus is very modular and developers can choose to remove this dependency in their extractors.

Conventions

Interfaces are first described before implementation to enforce modularity. All interface functions are described and this description is avoided in implementations.

Tests

In src, run

python -m pytest tests

This should be ran in your poetry environment. To drop into poetry environment, first run poetry install, and then poetry shell. You might need to change your python version to 3.9 for it to install the virtual environment for you.

Implementation Specific Details

To change the network that covalent extracts transactions from, go to extractor/covalent.py

TODO

Dockerfile
pre-commit linter
commit linter
back ups
remove the disable of duplicate code in .pylintrc (i.e. refactor the code such that there is no duplicate code)