Research & Development

Scope

• Objective: alembica leverages Large Language Models (LLMs) for systematic synthesis of unstrucutred data sources such as text reported in news corpora.
• Replicability: Addresses the challenge of consistent, unbiased analysis, countering the subjective nature of human information extraction.
• Cost: More economical than custom AI solutions.

Contributing

How to Contribute

We welcome contributions to improve alembica, whether you’re fixing bugs, adding features, or enhancing documentation:

• Branching Strategy: Create a new branch for each set of related changes and submit a pull request via GitHub.
• Code Reviews: All submissions undergo thorough review to maintain code quality.

Repository Overview

Before contributing, it’s helpful to understand how the repository is organized:

• definitions/ – contains JSON schemas and data structures used for validation.
• validation/ – validates input, output, and cost calculations using those schemas.
• llm/ – includes model selection, token counting, and provider-specific query functions.
• extraction/ – the main entry point that sequences prompts and validates responses.
• pricing/ – computes token costs for a request.
• utils/logger/ – provides logging utilities.
• utils/sharedlib/ – exposes C-shared library functions for use from other languages.
• test/ – integration tests with real model providers.

Check the README and documentation under docs/ for additional details on these packages and how they work together.

Guidelines

For detailed contribution guidelines, see our CONTRIBUTING.md and CODE_OF_CONDUCT.md.

Software Stack

alembica is developed in Go, selected for its simplicity and efficiency with concurrent operations. We prioritize the latest stable Go releases to incorporate improvements.

Open Science Support

alembica actively supports Open Science principles.