Psychic uses local machine learning to rank file search results
What happened
On November 16, Hacker News featured psychic, a terminal file browser similar to fzf. It uses a local LightGBM model to learn which files and directories a user visits and ranks search results with it, updating results as the user types with interface rendering under 16 milliseconds. Search and click records stay in a local sqlite database, and the model retrains in the background each time the tool opens. During cold start it ranks by recent clicks, file modification time and fuzzy matching.
Written by AI from the coverage · updated 5 hours ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pagePsychic: Like fzf but learns from your actions
psychic 是一款终端文件浏览器,通过本地 LightGBM 模型学习文件和目录访问习惯,随输入实时更新结果,界面渲染耗时低于 16ms。搜索与点击记录保存在本机 sqlite 数据库,每次打开工具都会后台重新训练模型,冷启动时使用近期点击、修改时间和模糊匹配进行排序。
Heat over time
Heat now 8·Comparable peak 9(Nov 17 04:00)·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.