jatechan/ai-data-extractor

选用指南

AI 根据 README 整理 · 2026/10/19 00:31:14

使用方式

通过命令行使用，支持交互式菜单或直接指定参数运行。

查看 README 依据

python --version # 3.9+ required, 3.10+ recommended # Interactive: pick which sources to extract from a numbered menu python extract.py # Or drive it directly python extract.py --all python extract.py --sources cursor,claude_code,aider python extract.py --list # just show what's installed, don't extract python extract.py --all --merge # also write all_conversations.jsonl # Shorthand for "extract everything" ./extract_all.sh ``` ### CLI reference ``` python extract.py [--all] [--sources ids] [--list] [--output-dir DIR]

部署要求

需要 Python 3.9 或更高版本，推荐 3.10+。

查看 README 依据

python --version # 3.9+ required, 3.10+ recommended # Interactive: pick which sources to extract from a numbered menu python extract.py # Or drive it directly python extract.py --all python extract.py --sources cursor,claude_code,aider python extract.py --list # just show what's installed, don't extract python extract.py --all --merge # also write all_conversations.jsonl # Shorthand for "extract everything" ./extract_all.sh ``` ### CLI reference ``` python extract.py [--all] [--sources ids] [--list] [--output-dir DIR]

已知限制

对于 Cursor、Windsurf 和 Trae，由于存储格式未公开且可能变化，提取可能不完整或失败，需要手动调整。

查看 README 依据

times (Cursor alone has gone through at least three shapes: workspace `ItemTable` chat, inline composer, and split `bubbleId` composer). The `cursor.py` extractor implements all three known shapes explicitly. `windsurf.py` and `trae.py` instead use a **generic heuristic** (`extractors/common.py::heuristic_extract_chat_from_kv`) that scans chat-related keys and walks the parsed JSON looking for objects shaped like a role + text pair. It's honest best-effort, not a documented format - if a future version changes shape and stops matching, that's expected; adjust