Dreams Coding Engine

This repository provides a rule-based coding engine for dream narratives. It includes:

• Lexicon-driven tagging rules with negation and idiom guards (src/rules.py).
• A command-line batch analyzer (src/analyze.py).
• A FastAPI layer for integrations with Google Sheets and GitHub webhooks (api/).

Local Setup

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
pip install -r requirements-api.txt  # optional but required for the API
python -m spacy download en_core_web_sm

Command-line Analyzer

python -m src.analyze --in_file data/raw/dreams.csv --text_col text

The analyzer merges the coded fields into a new CSV written to data/processed/ by default.

FastAPI Service

Start the API locally with Uvicorn:

uvicorn api.main:app --reload --port 8000

Configuration

Environment Variable	Default Value	Description
PRESET_DIR	configs/presets	Directory scanned for preset JSON files.
SCHEMA_PATH	schema/ruleset.schema.json	JSON Schema used by /validate_preset.
ENGINE_VERSION	0.3.0	Version string stamped onto /code results.
GITHUB_WEBHOOK_SECRET	CHANGE_ME	Shared secret for /gh/webhook.
CORS_ALLOW_ORIGINS	*	Comma-separated list of allowed origins.

Endpoints

• POST /code — Batch-code rows using the rule engine. Pass an optional preset key to apply a cached preset.
• GET /presets — List available presets (name@version).
• POST /extend_lexicon — Generate deterministic lexicon extension proposals.
• POST /validate_preset — Validate a preset JSON payload against the schema.
• POST /gh/webhook — Refresh the in-memory preset cache when triggered by a GitHub push event.

Docker

Build and run the API using the provided Dockerfile:

docker build -t dreams-engine .
docker run -p 8000:8000 --env GITHUB_WEBHOOK_SECRET=changeme dreams-engine

Google Sheets Add-on

See README_SHEETS_ADDON.md for the Apps Script snippet that integrates the /code endpoint with Google Sheets.

Testing

Run the unit test suite with:

pytest