Digest Generator

Overview

Digest Generator is a Python pipeline that aggregates articles from RSS feeds you define, generates fact-dense per-article summaries via an LLM, classifies them with zero-shot NLI, and produces a Markdown digest via Ollama. Feeds, sections, and prompts are all user-supplied; the tool ships generic baselines so it runs on any topic out of the box.

How It Works

flowchart LR
    Feeds[("RSS")]:::storage

    subgraph S["run --no-digest"]
        direction TB
        Fetch:::compute --> Summarize:::compute --> Classify:::compute
    end

    JSON[("JSON")]:::storage

    subgraph D["digest"]
        direction TB
        Writer:::compute --> Editor:::compute
        Editor --> Framer:::compute --> Watcher:::compute --> Composer:::compute
        Editor --> Composer
        Framer --> Composer
    end

    MD[("digest.md")]:::storage

    Feeds --> Fetch
    Classify --> JSON
    JSON --> Writer
    Composer --> MD

    classDef compute fill:#fed7aa,stroke:#9a3412,color:#0f172a
    classDef storage fill:#e5e7eb,stroke:#374151,color:#0f172a

digest-generator run does both halves end to end, writing the JSON corpus and the final Markdown digest into the same run directory. You can also run the halves separately: run --no-digest stops after building the corpus, and digest <run_dir> turns an existing corpus into a digest.

For full usage details, see docs/usage.md.

Installation

pip install digest-generator          # or: uv tool install digest-generator
digest-generator init                 # write a starter feeds.yaml
# edit ~/.config/digest-generator/feeds.yaml to add your categories and feeds

init creates ~/.config/digest-generator/feeds.yaml from a starter template. Edit it to define your own sections (categories:) and the feeds in each, then run digest-generator feeds to check it. The digest stages need a running Ollama; the topic classifier downloads a public model on first use.

Working from a clone instead (for development or audio/GPU extras):

git clone https://github.com/laplacef/digest-generator.git
cd digest-generator
uv sync --extra dev

Configuration

Every setting has a sensible default, so most setups need no environment variables. Override via the environment or a .env file in the working directory. The common ones:

Every field in digest_generator/shared/settings.py maps to an uppercase env var. Full setup (prerequisites, optional audio rendering, optional GPU acceleration) is in docs/setup.md.

Usage

digest-generator init                 # write a starter feeds.yaml
digest-generator run                  # full pipeline (fetch + summarize + classify + digest)
digest-generator run --no-digest      # corpus build only (skip digest generation)
digest-generator run --audio          # full pipeline + Piper TTS rendition
digest-generator digest <run_dir>     # regenerate the digest from an existing run directory
digest-generator audio <run_dir>      # render audio for an existing digest (no LLM cost)
digest-generator lint <digest.md>     # check a digest before publishing (exits 1 on errors)
digest-generator feeds                # list available feeds

Each run lands in its own timestamped directory under output/, containing the per-stage caches, the final Markdown digest, run metadata, and a log of the run. See docs/usage.md for the full CLI reference, programmatic API, and output layout.

Contributing

Bug reports, feature requests, and pull requests are all welcome. See CONTRIBUTING.md for development setup, coding standards, and the contribution workflow.

This project follows a Code of Conduct. By participating, you are expected to uphold it.

License

This project is licensed under the Apache License 2.0. You are free to use, modify, and distribute this project, provided you include proper attribution. See the NOTICE file for details.