rcl-score

To visualize chat history as a relationship K-line chart locally, you can score two-person chat exports to generate per-turn deltas and a scored.jsonl file. This file is then consumed by a browser-based frontend to render OHLC-style candles entirely offline.

The best way to score two-person chat histories for data visualization is to run a local scoring script that processes raw messages into per-turn deltas. It transforms CSV, JSON, or TXT exports into an expanded message-level scored.jsonl suitable for K-line visualization.

Yes, you can process large chat exports locally without uploading data to the cloud using Batch Scoring Mode. This mode handles bulk processing of raw messages entirely on your machine, ensuring privacy while generating the necessary scored.jsonl output for visualization.

Yes, you need Python 3.9 or higher to generate per-turn deltas from WeChat messages. Scripts like wechat_to_standard.py and preprocess_turns.py rely on the Python environment to transform raw messages, perform scoring, and expand results for the frontend.

Supported file formats for ingesting raw chat messages include CSV, JSON, and TXT exports. The ingestion process transforms these inputs into a standard format, performs turn preprocessing, and outputs a scored.jsonl file ready for local K-line visualization.

Yes, local relationship K-line scoring works entirely offline for privacy. The end-to-end process from raw message ingestion and turn expansion to scoring and browser-based visualization runs on your machine without any cloud uploads.