@shrkcrft/compress 0.1.0-alpha.21 → 0.1.0-alpha.22

package/dist/content/detect-content-type.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"detect-content-type.d.ts","sourceRoot":"","sources":["../../src/content/detect-content-type.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AA0FjD;;;;;GAKG;AACH,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,YAAY,CAuG5D"}
+{"version":3,"file":"detect-content-type.d.ts","sourceRoot":"","sources":["../../src/content/detect-content-type.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAgGjD;;;;;GAKG;AACH,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,GAAG,YAAY,CA4G5D"}

package/dist/content/detect-content-type.js

@@ -17,7 +17,13 @@ const YAML_INDENTED_SEQ = /^\s{2,}-\s/;
 // Log levels must appear as a LINE PREFIX (optionally after a leading
 // timestamp / bracket), not anywhere on the line — otherwise common code
 // identifiers (`const ERROR = 500`, `enum { INFO, DEBUG }`) misroute to logs.
-const LOG_MARKER = /^\s*(?:(?:\[?\d{4}-\d{2}-\d{2}[T ][\d:.,]+\]?|\S+\[\d+\]:)\s+)?\[?(?:ERROR|FATAL|FAIL(?:ED|URE)?|WARN(?:ING)?|INFO|DEBUG|NOTICE|TRACE)\b|^\S+\[\d+\]:\s|^\s*Traceback\b|^\s+at\s+\S+\s*\(|^\d{4}-\d{2}-\d{2}[T ]\d{2}:\d{2}/;
+// A bare clock timestamp (`HH:MM:SS`, optionally bracketed / with millis) is
+// also a valid log prefix: it appears as the optional leading-prefix before a
+// level keyword AND as a standalone line-start (a timestamp-led log line with
+// no level word). Without these, `[10:00:00] INFO` / `10:00:00 ...` lines match
+// SEARCH_LINE (the `HH:MM:` shape looks like `token:digits:`) and the blob
+// misroutes to search-results — a lossy compressor for what is really a log.
+const LOG_MARKER = /^\s*(?:(?:\[?\d{4}-\d{2}-\d{2}[T ][\d:.,]+\]?|\S+\[\d+\]:|\[?\d{1,2}:\d{2}:\d{2}(?:[.,]\d+)?\]?)\s+)?\[?(?:ERROR|FATAL|FAIL(?:ED|URE)?|WARN(?:ING)?|INFO|DEBUG|NOTICE|TRACE)\b|^\S+\[\d+\]:\s|^\s*Traceback\b|^\s+at\s+\S+\s*\(|^\d{4}-\d{2}-\d{2}[T ]\d{2}:\d{2}|^\s*\[?\d{1,2}:\d{2}:\d{2}(?:[.,]\d+)?\]?\s/;
 const MARKDOWN_MARKER = /^(?:#{1,6}\s|\s*[-*]\s|\s*\d+\.\s|```|>\s|\|)/;
 // Code signals: declaration keywords + structural/statement shapes that real
 // code has but prose / markdown / INI / env / TOML / nginx do NOT (each new
@@ -120,8 +126,11 @@ export function detectContentType(text) {
         return EContentType.GitDiff;
     }
     // 3. grep / ripgrep search output (`path:line:` prefix) OR compiler
-    //    diagnostics (`path(line,col):`). Count either shape toward the ratio.
-    if (lineHitRatio(lines, (l) => SEARCH_LINE.test(l) || DIAGNOSTIC_LINE.test(l)) >= 0.6) {
+    //    diagnostics (`path(line,col):`). Count either shape toward the ratio,
+    //    but EXCLUDE log-shaped lines: a `HH:MM:SS` clock satisfies the
+    //    `token:digits:` SEARCH_LINE shape, so timestamped logs would otherwise
+    //    win here (ratio ≥ 0.6) before the BuildLog check below ever runs.
+    if (lineHitRatio(lines, (l) => (SEARCH_LINE.test(l) || DIAGNOSTIC_LINE.test(l)) && !LOG_MARKER.test(l)) >= 0.6) {
         return EContentType.SearchResults;
     }
     // 4. Build / test log (error / warn / timestamp markers dense enough).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@shrkcrft/compress",
-  "version": "0.1.0-alpha.21",
+  "version": "0.1.0-alpha.22",
   "description": "SharkCraft deterministic context-compression engine: content routing, lossless columnar/table compaction, log/search/diff line reduction, and reversible Compress-Cache-Retrieve (CCR). No model inside — every transform is a pure function of its input.",
   "license": "MIT",
   "author": "SharkCraft contributors",
@@ -44,7 +44,7 @@
     "typecheck": "tsc --noEmit -p tsconfig.json"
   },
   "dependencies": {
-    "@shrkcrft/core": "^0.1.0-alpha.21"
+    "@shrkcrft/core": "^0.1.0-alpha.22"
   },
   "publishConfig": {
     "access": "public"