X/LittleWhiteBox
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refine tokenizer preload and metrics ratio

Browse Source
This commit is contained in:
bielie
2026-02-09 20:58:48 +08:00
parent 0a28539b29
commit 36ba66a523
2 changed files with 27 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
modules/story-summary/vector/retrieval/metrics.js
View File

@@ -430,12 +430,18 @@ export function detectIssues(metrics) {
// ─────────────────────────────────────────────────────────────────
if (m.event.considered > 0) {
const selectRatio = m.event.selected / m.event.considered;
if (selectRatio < 0.1) {
issues.push(`Event selection ratio too low (${(selectRatio * 100).toFixed(1)}%) - threshold may be too high`);
// 只统计 Dense 路选中（direct + related），Lexical 是额外补充不计入
const denseSelected =
(m.event.byRecallType?.direct || 0) +
(m.event.byRecallType?.related || 0);
const denseSelectRatio = denseSelected / m.event.considered;
if (denseSelectRatio < 0.1) {
issues.push(`Dense event selection ratio too low (${(denseSelectRatio * 100).toFixed(1)}%) - threshold may be too high`);
}
if (selectRatio > 0.6 && m.event.considered > 10) {
issues.push(`Event selection ratio high (${(selectRatio * 100).toFixed(1)}%) - may include noise`);
if (denseSelectRatio > 0.6 && m.event.considered > 10) {
issues.push(`Dense event selection ratio high (${(denseSelectRatio * 100).toFixed(1)}%) - may include noise`);
}
}