Herodotus Data Processor (HDP)

Warning: This codebase is experimental and not audited. Use at your own risk.

HDP enhances off-chain compute capabilities with zkVMs for verifiable on-chain data integration. For more, visit our documentation.

Introduction

The Data Processor CLI serves as an essential tool for developers working with Cairo programs and zkVM environments. Its primary function is to translate human-readable requests into a format compatible with Cairo programs, enabling commands to be executed over the Cairo VM and generating executable outputs. This transformation is a crucial preprocessing step that prepares data for off-chain computations in zkVM environments.

Features

• Development Tools: Encode and decode data lakes and computational tasks.
• Core Processing: Compile data from various sources and compute aggregate functions.
• Extensibility: Support for multiple blockchain integrations and various ZKVM backends is planned.
• Ease of Use: Provides a CLI for easy interaction with the system.

Install HDP

Install with cargo

# Install with cargo
❯ cargo install --git https://github.com/HerodotusDev/hdp/ --tag {TAG} --locked --force hdp-cli

Build from source

# clone repo
❯ git clone https://github.com/HerodotusDev/hdp.git

# install hdp
❯ cargo install --locked -f --path  cli/

Getting Started

To launch the interactive CLI:

# Start the HDP
❯ hdp start
Welcome to Herodotus Data Processor interactive CLI! 🛰️

                _   _   ____    ____
                | | | | |  _ \  |  _ \
                | |_| | | | | | | |_) |
                |  _  | | |_| | |  __/
                |_| |_| |____/  |_|

? Step 1. What's your datalake type?

Usage Examples

First locate .env file like the one in example.

Second, run command like below :

note that this will go though both pre-process -> process step.

hdp run -r ${Request file path} -p ${Program input file path} -b ${Batch proof file path} -c ${PIE file after process}

For a more comprehensive guide on commands available on hdp run:

❯ hdp run --help
Run batch of tasks base on request json file

Usage: hdp run [OPTIONS] --request-file <REQUEST_FILE> --program-input-file <PROGRAM_INPUT_FILE>

Options:
  -r, --request-file <REQUEST_FILE>
          Pass request as json file
      --rpc-url <RPC_URL>
          The RPC URL to fetch the data
      --dry-run-cairo-file <DRY_RUN_CAIRO_FILE>
          dry run contract bootloader program. only used for module task
  -p, --program-input-file <PROGRAM_INPUT_FILE>
          Path to save program input file after pre-processing
      --cairo-format
          Set this boolean to true to generate cairo format program_input_file
  -b, --batch-proof-file <BATCH_PROOF_FILE>
          Path to save batch proof file after pre-processing
      --sound-run-cairo-file <SOUND_RUN_CAIRO_FILE>
          hdp cairo compiled program. main entry point
  -c, --cairo-pie-file <CAIRO_PIE_FILE>
          Path to save pie file
  -h, --help
          Print help (see more with '--help')

Integration Testing

Integration testing in HDP ensures that the functionality of aggregate functions such as SUM, AVG, MIN, MAX, and COUNT operates correctly across various numeric fields within the blockchain data structure. These functions are designed specifically for numeric data types, ensuring accurate and reliable computations.

Integration Test

The core soundness of HDP relies on generating the correct input file and running the Cairo program. To ensure this, a full integration test flow is necessary to link the pre-processor and processor versions. For continuous integration tests, please refer to the hdp-test repository as it contains all the cases of supported features in table below.

Supported Aggregate Functions

• SUM, AVG, MIN, MAX, COUNT: These functions are supported only for fields with numeric values.

Context Required Operation

For a practical example of how to implement context-sensitive operations, refer to the implementation of the COUNT function. This example shows how to pass and utilize additional context for operations, which can be particularly useful for conditional processing or complex calculations.

Function Support Matrix

Here is the support matrix indicating which blockchain elements are tested for each aggregate function. The matrix highlights fields where these functions are applicable.

Note: Fields marked with "-" are not applicable for the specified aggregate functions because they do not contain numeric data or the data type is not suitable for these calculations.

Additional Notes

• Please ensure that the data fields you are applying these functions contain numeric values to avoid computational errors.
• For details on how these tests are performed or to contribute to the further development of tests, please refer to the Integration Test Guide.

Development

# CI check
just run-ci-flow

Local Run

To run HDP in a stable environment locally, you need to have cairo-run installed with the necessary tools in the correct path and locate the compiled Cairo program. If these steps sound tricky to you, just use the Docker image.

To mount in a container environment, you need to create empty input.json, output.json, and cairo.pie files in the root directory of the host machine before running it.

And locate requeset.json file on root that contains intended request format.

docker-compose build

docker-compose up

For those looking for an already built Docker image, you can pull it from here.

License

hdp is licensed under the GNU General Public License v3.0.

Herodotus Dev Ltd - 2024