Scrub spinner glyphs from window-title bucket keys

Bucket keys are {Class, ScrubTitle(Title)}, so each animated progress frame a
terminal app cycles into its title (Claude Code's braille and dingbat-star
glyphs) became a distinct bucket — fragmenting one task into dozens of rows and
inflating the switch count, since applyEvent counts a switch per key change.

ScrubTitle now drops any non-ASCII symbol or control rune and collapses leftover
whitespace. This is category-based rather than range-based, so future spinner
sets are covered too, while letters and digits of every script survive and ASCII
symbols (e.g. "C++") are left intact.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
2026-06-01 19:23:49 -04:00
parent 1c81eb1b50
commit 71b8c80c69
2 changed files with 34 additions and 9 deletions
+21 -3
View File
@@ -6,6 +6,8 @@ package evidence
import ( import (
"context" "context"
"regexp" "regexp"
"strings"
"unicode"
) )
// EvidenceHealth records whether the sensor could observe the active window. // EvidenceHealth records whether the sensor could observe the active window.
@@ -32,8 +34,24 @@ type Source interface {
// keys (e.g. "1:23", "50.5%", "-3.0"). Ported verbatim from the Rust impl. // keys (e.g. "1:23", "50.5%", "-3.0"). Ported verbatim from the Rust impl.
var titleNoise = regexp.MustCompile(`-?\d+([:.]\d+)+%?`) var titleNoise = regexp.MustCompile(`-?\d+([:.]\d+)+%?`)
// ScrubTitle removes volatile numeric runs so window titles bucket stably. // extraSpace collapses the whitespace runs left behind by the scrubs below.
// The raw event log keeps the unscrubbed title; only bucket keys are scrubbed. var extraSpace = regexp.MustCompile(`\s+`)
// ScrubTitle normalizes a window title into a stable bucket key. It drops
// volatile numeric runs (clocks, ratios, percentages) and any non-ASCII symbol
// or control rune. The latter covers the animated spinner frames terminal apps
// (e.g. Claude Code) cycle through their title — braille and dingbat glyphs are
// both Unicode symbol category, so a single task no longer fragments into dozens
// of buckets. Letters and digits of every script survive, as do ASCII symbols
// (so "C++" is not mangled). Leftover whitespace is collapsed. The raw event log
// keeps the unscrubbed title; only bucket keys (and their display) are scrubbed.
func ScrubTitle(title string) string { func ScrubTitle(title string) string {
return titleNoise.ReplaceAllString(title, "") title = titleNoise.ReplaceAllString(title, "")
title = strings.Map(func(r rune) rune {
if unicode.IsControl(r) || (r > unicode.MaxASCII && unicode.IsSymbol(r)) {
return -1
}
return r
}, title)
return strings.TrimSpace(extraSpace.ReplaceAllString(title, " "))
} }
+8 -1
View File
@@ -11,9 +11,16 @@ func TestScrubTitle(t *testing.T) {
{"Plain title", "Plain title"}, // no digits-with-separator: untouched {"Plain title", "Plain title"}, // no digits-with-separator: untouched
{"Buy 2 eggs", "Buy 2 eggs"}, // bare integer: untouched (group is mandatory) {"Buy 2 eggs", "Buy 2 eggs"}, // bare integer: untouched (group is mandatory)
{"12.5%", ""}, // percent decimal: stripped whole {"12.5%", ""}, // percent decimal: stripped whole
{"1:23:45 remaining", " remaining"}, // clock ratio: stripped {"1:23:45 remaining", "remaining"}, // clock ratio: stripped, space collapsed
{"-3.0 delta", "delta"}, // leading sign + decimal: stripped {"-3.0 delta", "delta"}, // leading sign + decimal: stripped
{"Download 50.5% complete", "Download complete"}, // embedded percent decimal {"Download 50.5% complete", "Download complete"}, // embedded percent decimal
{"⠸ antidrift", "antidrift"}, // braille spinner frame stripped
{"⠼ antidrift", "antidrift"}, // different frame collapses to same key
{"✳ Check latest commit date", "Check latest commit date"}, // dingbat-star indicator
{"⠂ Check latest commit date", "Check latest commit date"}, // same task, different frame
{"✳ ⠼ done", "done"}, // multiple glyphs in one title
{"C++ build", "C++ build"}, // ASCII symbols kept (not all-symbols)
{"日本語 window", "日本語 window"}, // non-ASCII letters preserved
} }
for _, c := range cases { for _, c := range cases {
if got := ScrubTitle(c.in); got != c.want { if got := ScrubTitle(c.in); got != c.want {