greprag 5.40.1 → 5.42.0

package/dist/crush/adaptive-sizer.js

@@ -0,0 +1,156 @@
+"use strict";
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/adaptive-sizer.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/adaptive-sizer.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Adaptive keep-count via information saturation — port of Headroom's
+ * transforms/adaptive_sizer.rs (itself a port of adaptive_sizer.py).
+ *
+ * Decides HOW MANY items to keep, statistically: find the knee of the
+ * cumulative unique-bigram coverage curve (Kneedle). Coverage stops
+ * growing -> that's the knee -> keep that many.
+ *
+ * Deliberate divergence from Headroom: the zlib-ratio validation tier
+ * is dropped (it nudged k up ~20% on pathological subsets; not worth a
+ * compression dependency in a zero-dep core). Simhash is replaced with
+ * a plain FNV-1a 4-gram sketch — it only feeds the "near-total
+ * redundancy" fast path, where any stable hash works.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.computeUniqueBigramCurve = computeUniqueBigramCurve;
+exports.findKnee = findKnee;
+exports.computeOptimalK = computeOptimalK;
+/** FNV-1a 32-bit over a string, with a seed offset. Stable, dependency-free. */
+function fnv1a(s, seed) {
+    let h = (0x811c9dc5 ^ seed) >>> 0;
+    for (let i = 0; i < s.length; i++) {
+        h ^= s.charCodeAt(i);
+        h = Math.imul(h, 0x01000193);
+    }
+    return h >>> 0;
+}
+/** 64-bit-ish simhash over character 4-grams (two independent FNV seeds). */
+function simhash(s) {
+    const votes1 = new Array(32).fill(0);
+    const votes2 = new Array(32).fill(0);
+    const text = s.length < 4 ? s + ' '.repeat(4 - s.length) : s;
+    for (let i = 0; i + 4 <= text.length; i++) {
+        const gram = text.slice(i, i + 4);
+        const h1 = fnv1a(gram, 0);
+        const h2 = fnv1a(gram, 0x9e3779b9);
+        for (let b = 0; b < 32; b++) {
+            votes1[b] += (h1 >>> b) & 1 ? 1 : -1;
+            votes2[b] += (h2 >>> b) & 1 ? 1 : -1;
+        }
+    }
+    let out = '';
+    for (let b = 0; b < 32; b++)
+        out += votes1[b] >= 0 ? '1' : '0';
+    for (let b = 0; b < 32; b++)
+        out += votes2[b] >= 0 ? '1' : '0';
+    return out;
+}
+/**
+ * Count distinct simhash groups. No early exit — the count also feeds
+ * the diversity ratio, which needs the true cardinality.
+ */
+function countUniqueSimhash(items) {
+    const seen = new Set();
+    for (const item of items)
+        seen.add(simhash(item));
+    return seen.size;
+}
+/**
+ * Cumulative unique word-bigram coverage: curve[i] = distinct bigrams
+ * seen across items[0..=i]. Single-word items emit `(word, "")`;
+ * empty items emit `("", "")` — mirrors the Python/Rust cardinality.
+ */
+function computeUniqueBigramCurve(items) {
+    const seen = new Set();
+    const curve = [];
+    for (const item of items) {
+        const words = item.split(/\s+/).filter((w) => w.length > 0);
+        if (words.length === 0) {
+            seen.add(' ');
+        }
+        else if (words.length === 1) {
+            seen.add(words[0] + ' ');
+        }
+        else {
+            for (let i = 0; i + 1 < words.length; i++) {
+                seen.add(words[i] + ' ' + words[i + 1]);
+            }
+        }
+        curve.push(seen.size);
+    }
+    return curve;
+}
+/**
+ * Kneedle knee finder on a monotonically-increasing curve. Returns the
+ * 1-indexed keep-count, or null when no knee clears the 0.05 deviation
+ * threshold. Flat curve (all items identical) -> 1.
+ */
+function findKnee(curve) {
+    const n = curve.length;
+    if (n < 3)
+        return null;
+    const yMin = curve[0];
+    const yMax = curve[n - 1];
+    if (Math.abs(yMax - yMin) < Number.EPSILON)
+        return 1;
+    const xRange = n - 1;
+    const yRange = yMax - yMin;
+    let maxDiff = -1;
+    let kneeIdx = null;
+    for (let i = 0; i < n; i++) {
+        const xNorm = i / xRange;
+        const yNorm = (curve[i] - yMin) / yRange;
+        const diff = yNorm - xNorm;
+        if (diff > maxDiff) {
+            maxDiff = diff;
+            kneeIdx = i;
+        }
+    }
+    if (maxDiff < 0.05)
+        return null;
+    return kneeIdx === null ? null : kneeIdx + 1;
+}
+/**
+ * Compute the optimal number of items to keep.
+ *
+ * @param items string representations of items in importance order
+ * @param bias  multiplier on the knee (>1 keep more, <1 compress harder)
+ * @param minK  lower bound
+ * @param maxK  upper bound (undefined = items.length)
+ */
+function computeOptimalK(items, bias, minK, maxK) {
+    const n = items.length;
+    const effectiveMax = maxK ?? n;
+    // Tier 1: trivial.
+    if (n <= 8)
+        return n;
+    // Near-total redundancy: <=3 unique groups -> keep that many.
+    const uniqueCount = countUniqueSimhash(items);
+    if (uniqueCount <= 3) {
+        return Math.min(Math.max(minK, uniqueCount), effectiveMax);
+    }
+    // Tier 2: Kneedle on the bigram coverage curve.
+    const curve = computeUniqueBigramCurve(items);
+    let knee = findKnee(curve);
+    const diversityRatio = uniqueCount / n;
+    if (knee === null) {
+        // No saturation — scale keep-fraction with diversity.
+        const keepFraction = 0.3 + 0.7 * diversityRatio;
+        knee = Math.max(minK, Math.floor(n * keepFraction));
+    }
+    else if (diversityRatio > 0.7) {
+        const floor = Math.max(minK, Math.floor(n * (0.3 + 0.7 * diversityRatio)));
+        knee = Math.max(knee, floor);
+    }
+    let k = Math.max(minK, Math.floor(knee * bias));
+    k = Math.min(k, effectiveMax);
+    return Math.max(minK, Math.min(k, effectiveMax));
+}
+//# sourceMappingURL=adaptive-sizer.js.map

package/dist/crush/adaptive-sizer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"adaptive-sizer.js","sourceRoot":"","sources":["../../src/crush/adaptive-sizer.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;;;;;;;GAkBG;;AA+CH,4DAiBC;AAOD,4BAqBC;AAUD,0CAmCC;AAvID,gFAAgF;AAChF,SAAS,KAAK,CAAC,CAAS,EAAE,IAAY;IACpC,IAAI,CAAC,GAAG,CAAC,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;IAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAClC,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;QACrB,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;IAC/B,CAAC;IACD,OAAO,CAAC,KAAK,CAAC,CAAC;AACjB,CAAC;AAED,6EAA6E;AAC7E,SAAS,OAAO,CAAC,CAAS;IACxB,MAAM,MAAM,GAAG,IAAI,KAAK,CAAS,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC7C,MAAM,MAAM,GAAG,IAAI,KAAK,CAAS,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC7C,MAAM,IAAI,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC1C,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QAClC,MAAM,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAC1B,MAAM,EAAE,GAAG,KAAK,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;QACnC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,EAAE,EAAE,CAAC;YAC5B,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YACrC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IACD,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,EAAE;QAAE,GAAG,IAAI,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;IAC/D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,EAAE;QAAE,GAAG,IAAI,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;IAC/D,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;GAGG;AACH,SAAS,kBAAkB,CAAC,KAAwB;IAClD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,KAAK,MAAM,IAAI,IAAI,KAAK;QAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;IAClD,OAAO,IAAI,CAAC,IAAI,CAAC;AACnB,CAAC;AAED;;;;GAIG;AACH,SAAgB,wBAAwB,CAAC,KAAwB;IAC/D,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QAC5D,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAChB,CAAC;aAAM,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC9B,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC;QAC3B,CAAC;aAAM,CAAC;YACN,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC1C,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,GAAG,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YAC1C,CAAC;QACH,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACxB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;;GAIG;AACH,SAAgB,QAAQ,CAAC,KAAwB;IAC/C,MAAM,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;IACvB,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACvB,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;IACtB,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAC1B,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,GAAG,IAAI,CAAC,GAAG,MAAM,CAAC,OAAO;QAAE,OAAO,CAAC,CAAC;IACrD,MAAM,MAAM,GAAG,CAAC,GAAG,CAAC,CAAC;IACrB,MAAM,MAAM,GAAG,IAAI,GAAG,IAAI,CAAC;IAC3B,IAAI,OAAO,GAAG,CAAC,CAAC,CAAC;IACjB,IAAI,OAAO,GAAkB,IAAI,CAAC;IAClC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC;QACzB,MAAM,KAAK,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,MAAM,CAAC;QACzC,MAAM,IAAI,GAAG,KAAK,GAAG,KAAK,CAAC;QAC3B,IAAI,IAAI,GAAG,OAAO,EAAE,CAAC;YACnB,OAAO,GAAG,IAAI,CAAC;YACf,OAAO,GAAG,CAAC,CAAC;QACd,CAAC;IACH,CAAC;IACD,IAAI,OAAO,GAAG,IAAI;QAAE,OAAO,IAAI,CAAC;IAChC,OAAO,OAAO,KAAK,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,GAAG,CAAC,CAAC;AAC/C,CAAC;AAED;;;;;;;GAOG;AACH,SAAgB,eAAe,CAC7B,KAAwB,EACxB,IAAY,EACZ,IAAY,EACZ,IAAa;IAEb,MAAM,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;IACvB,MAAM,YAAY,GAAG,IAAI,IAAI,CAAC,CAAC;IAE/B,mBAAmB;IACnB,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC;IAErB,8DAA8D;IAC9D,MAAM,WAAW,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;IAC9C,IAAI,WAAW,IAAI,CAAC,EAAE,CAAC;QACrB,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,WAAW,CAAC,EAAE,YAAY,CAAC,CAAC;IAC7D,CAAC;IAED,gDAAgD;IAChD,MAAM,KAAK,GAAG,wBAAwB,CAAC,KAAK,CAAC,CAAC;IAC9C,IAAI,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC3B,MAAM,cAAc,GAAG,WAAW,GAAG,CAAC,CAAC;IAEvC,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;QAClB,sDAAsD;QACtD,MAAM,YAAY,GAAG,GAAG,GAAG,GAAG,GAAG,cAAc,CAAC;QAChD,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,YAAY,CAAC,CAAC,CAAC;IACtD,CAAC;SAAM,IAAI,cAAc,GAAG,GAAG,EAAE,CAAC;QAChC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,GAAG,GAAG,GAAG,GAAG,cAAc,CAAC,CAAC,CAAC,CAAC;QAC3E,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC/B,CAAC;IAED,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC;IAChD,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;IAC9B,OAAO,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC,CAAC;AACnD,CAAC"}

package/dist/crush/crush-keywords.d.ts

@@ -0,0 +1,40 @@
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/crush-keywords.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/crush-keywords.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Keyword registries + word-boundary classification shared by the crush
+ * compressors. Port of Headroom's signals/keyword_detector.rs +
+ * smart_crusher/error_keywords.rs.
+ *
+ * Carries Headroom's documented fixes:
+ * - the error set includes abort/timeout/denied/rejected (the Python
+ *   regex omitted them while the keyword list had them);
+ * - `token` is NOT in the security set (false-positives on every LLM
+ *   token count in agent logs).
+ */
+/** Error/failure keywords used for line scoring and JSON item preservation. */
+export declare const ERROR_KEYWORDS: readonly string[];
+/**
+ * The 12-keyword preservation set for JSON items (smart_crusher's
+ * fallback signal — intentionally broad; over-preserving beats dropping
+ * a real error). Differs from ERROR_KEYWORDS by omitting bare 'fail'
+ * (too noisy inside serialized JSON, where 'failed'/'failure' cover it).
+ */
+export declare const JSON_ERROR_KEYWORDS: readonly string[];
+export declare const WARNING_KEYWORDS: readonly string[];
+export declare const IMPORTANCE_KEYWORDS: readonly string[];
+/**
+ * Case-insensitive whole-word containment check ("errorless" does not
+ * match "error"; "ERROR:" does).
+ */
+export declare function containsKeyword(line: string, keyword: string): boolean;
+export declare function containsAnyKeyword(line: string, keywords: readonly string[]): boolean;
+export type ImportanceCategory = 'error' | 'warning' | 'importance';
+/**
+ * Classify a line's importance category for search-match scoring.
+ * Error outranks warning outranks importance (matches Headroom's
+ * priority-pattern ordering: 0.5 / 0.4 / 0.3 boosts).
+ */
+export declare function classifyImportance(line: string): ImportanceCategory | null;

package/dist/crush/crush-keywords.js

@@ -0,0 +1,110 @@
+"use strict";
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/crush-keywords.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/crush-keywords.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Keyword registries + word-boundary classification shared by the crush
+ * compressors. Port of Headroom's signals/keyword_detector.rs +
+ * smart_crusher/error_keywords.rs.
+ *
+ * Carries Headroom's documented fixes:
+ * - the error set includes abort/timeout/denied/rejected (the Python
+ *   regex omitted them while the keyword list had them);
+ * - `token` is NOT in the security set (false-positives on every LLM
+ *   token count in agent logs).
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.IMPORTANCE_KEYWORDS = exports.WARNING_KEYWORDS = exports.JSON_ERROR_KEYWORDS = exports.ERROR_KEYWORDS = void 0;
+exports.containsKeyword = containsKeyword;
+exports.containsAnyKeyword = containsAnyKeyword;
+exports.classifyImportance = classifyImportance;
+/** Error/failure keywords used for line scoring and JSON item preservation. */
+exports.ERROR_KEYWORDS = [
+    'error',
+    'exception',
+    'fail',
+    'failed',
+    'failure',
+    'fatal',
+    'critical',
+    'crash',
+    'panic',
+    'abort',
+    'timeout',
+    'denied',
+    'rejected',
+];
+/**
+ * The 12-keyword preservation set for JSON items (smart_crusher's
+ * fallback signal — intentionally broad; over-preserving beats dropping
+ * a real error). Differs from ERROR_KEYWORDS by omitting bare 'fail'
+ * (too noisy inside serialized JSON, where 'failed'/'failure' cover it).
+ */
+exports.JSON_ERROR_KEYWORDS = [
+    'error',
+    'exception',
+    'failed',
+    'failure',
+    'critical',
+    'fatal',
+    'crash',
+    'panic',
+    'abort',
+    'timeout',
+    'denied',
+    'rejected',
+];
+exports.WARNING_KEYWORDS = ['warn', 'warning'];
+exports.IMPORTANCE_KEYWORDS = [
+    'important',
+    'note',
+    'todo',
+    'fixme',
+    'hack',
+    'xxx',
+    'bug',
+    'fix',
+];
+function isWordChar(ch) {
+    return /[A-Za-z0-9_]/.test(ch);
+}
+/**
+ * Case-insensitive whole-word containment check ("errorless" does not
+ * match "error"; "ERROR:" does).
+ */
+function containsKeyword(line, keyword) {
+    const haystack = line.toLowerCase();
+    const needle = keyword.toLowerCase();
+    let from = 0;
+    for (;;) {
+        const idx = haystack.indexOf(needle, from);
+        if (idx === -1)
+            return false;
+        const leftOk = idx === 0 || !isWordChar(haystack[idx - 1]);
+        const right = idx + needle.length;
+        const rightOk = right === haystack.length || !isWordChar(haystack[right]);
+        if (leftOk && rightOk)
+            return true;
+        from = idx + 1;
+    }
+}
+function containsAnyKeyword(line, keywords) {
+    return keywords.some((kw) => containsKeyword(line, kw));
+}
+/**
+ * Classify a line's importance category for search-match scoring.
+ * Error outranks warning outranks importance (matches Headroom's
+ * priority-pattern ordering: 0.5 / 0.4 / 0.3 boosts).
+ */
+function classifyImportance(line) {
+    if (containsAnyKeyword(line, exports.ERROR_KEYWORDS))
+        return 'error';
+    if (containsAnyKeyword(line, exports.WARNING_KEYWORDS))
+        return 'warning';
+    if (containsAnyKeyword(line, exports.IMPORTANCE_KEYWORDS))
+        return 'importance';
+    return null;
+}
+//# sourceMappingURL=crush-keywords.js.map

package/dist/crush/crush-keywords.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"crush-keywords.js","sourceRoot":"","sources":["../../src/crush/crush-keywords.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;;;;;GAeG;;;AA6DH,0CAaC;AAED,gDAEC;AASD,gDAKC;AA1FD,+EAA+E;AAClE,QAAA,cAAc,GAAsB;IAC/C,OAAO;IACP,WAAW;IACX,MAAM;IACN,QAAQ;IACR,SAAS;IACT,OAAO;IACP,UAAU;IACV,OAAO;IACP,OAAO;IACP,OAAO;IACP,SAAS;IACT,QAAQ;IACR,UAAU;CACX,CAAC;AAEF;;;;;GAKG;AACU,QAAA,mBAAmB,GAAsB;IACpD,OAAO;IACP,WAAW;IACX,QAAQ;IACR,SAAS;IACT,UAAU;IACV,OAAO;IACP,OAAO;IACP,OAAO;IACP,OAAO;IACP,SAAS;IACT,QAAQ;IACR,UAAU;CACX,CAAC;AAEW,QAAA,gBAAgB,GAAsB,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;AAE1D,QAAA,mBAAmB,GAAsB;IACpD,WAAW;IACX,MAAM;IACN,MAAM;IACN,OAAO;IACP,MAAM;IACN,KAAK;IACL,KAAK;IACL,KAAK;CACN,CAAC;AAEF,SAAS,UAAU,CAAC,EAAU;IAC5B,OAAO,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACjC,CAAC;AAED;;;GAGG;AACH,SAAgB,eAAe,CAAC,IAAY,EAAE,OAAe;IAC3D,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;IACpC,MAAM,MAAM,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC;IACrC,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,SAAS,CAAC;QACR,MAAM,GAAG,GAAG,QAAQ,CAAC,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;QAC3C,IAAI,GAAG,KAAK,CAAC,CAAC;YAAE,OAAO,KAAK,CAAC;QAC7B,MAAM,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;QAC3D,MAAM,KAAK,GAAG,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC;QAClC,MAAM,OAAO,GAAG,KAAK,KAAK,QAAQ,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC;QAC1E,IAAI,MAAM,IAAI,OAAO;YAAE,OAAO,IAAI,CAAC;QACnC,IAAI,GAAG,GAAG,GAAG,CAAC,CAAC;IACjB,CAAC;AACH,CAAC;AAED,SAAgB,kBAAkB,CAAC,IAAY,EAAE,QAA2B;IAC1E,OAAO,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,eAAe,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;AAC1D,CAAC;AAID;;;;GAIG;AACH,SAAgB,kBAAkB,CAAC,IAAY;IAC7C,IAAI,kBAAkB,CAAC,IAAI,EAAE,sBAAc,CAAC;QAAE,OAAO,OAAO,CAAC;IAC7D,IAAI,kBAAkB,CAAC,IAAI,EAAE,wBAAgB,CAAC;QAAE,OAAO,SAAS,CAAC;IACjE,IAAI,kBAAkB,CAAC,IAAI,EAAE,2BAAmB,CAAC;QAAE,OAAO,YAAY,CAAC;IACvE,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/crush/crush-types.d.ts

@@ -0,0 +1,38 @@
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/crush-types.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/crush-types.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Shared types for the crush engine — pure tool-output compressors
+ * (TypeScript ports of Headroom's search/log/smart-crusher transforms).
+ *
+ * Every compressor is a pure function: string in → { output, stats } out.
+ * No I/O, no deps. CCR storage and CLI plumbing live in packages/cli.
+ */
+export type CrushSourceType = 'search' | 'log' | 'json';
+export interface CrushStats {
+    /** Source type the compressor handled. */
+    sourceType: CrushSourceType;
+    /** Input size in characters. */
+    originalChars: number;
+    /** Output size in characters. */
+    compressedChars: number;
+    /** compressedChars / originalChars (1.0 = no compression). */
+    compressionRatio: number;
+    /** Units the compressor reasons in (matches/lines/items) before selection. */
+    originalUnits: number;
+    /** Units surviving selection. */
+    keptUnits: number;
+    /** True when the compressor passed input through unchanged. */
+    passthrough: boolean;
+    /** Why passthrough happened (when it did). */
+    passthroughReason?: string;
+}
+export interface CrushResult {
+    output: string;
+    stats: CrushStats;
+}
+/** Rough token estimate shared by validate-or-revert checks: 1 token ≈ 4 chars. */
+export declare function estimateTokens(text: string): number;
+export declare function makePassthrough(content: string, sourceType: CrushSourceType, reason: string, units?: number): CrushResult;

package/dist/crush/crush-types.js

@@ -0,0 +1,36 @@
+"use strict";
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/crush-types.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/crush-types.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Shared types for the crush engine — pure tool-output compressors
+ * (TypeScript ports of Headroom's search/log/smart-crusher transforms).
+ *
+ * Every compressor is a pure function: string in → { output, stats } out.
+ * No I/O, no deps. CCR storage and CLI plumbing live in packages/cli.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.estimateTokens = estimateTokens;
+exports.makePassthrough = makePassthrough;
+/** Rough token estimate shared by validate-or-revert checks: 1 token ≈ 4 chars. */
+function estimateTokens(text) {
+    return Math.ceil(text.length / 4);
+}
+function makePassthrough(content, sourceType, reason, units = 0) {
+    return {
+        output: content,
+        stats: {
+            sourceType,
+            originalChars: content.length,
+            compressedChars: content.length,
+            compressionRatio: 1.0,
+            originalUnits: units,
+            keptUnits: units,
+            passthrough: true,
+            passthroughReason: reason,
+        },
+    };
+}
+//# sourceMappingURL=crush-types.js.map

package/dist/crush/crush-types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"crush-types.js","sourceRoot":"","sources":["../../src/crush/crush-types.ts"],"names":[],"mappings":";AAAA;;;;;;;;;;;GAWG;;AA6BH,wCAEC;AAED,0CAmBC;AAxBD,mFAAmF;AACnF,SAAgB,cAAc,CAAC,IAAY;IACzC,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AACpC,CAAC;AAED,SAAgB,eAAe,CAC7B,OAAe,EACf,UAA2B,EAC3B,MAAc,EACd,KAAK,GAAG,CAAC;IAET,OAAO;QACL,MAAM,EAAE,OAAO;QACf,KAAK,EAAE;YACL,UAAU;YACV,aAAa,EAAE,OAAO,CAAC,MAAM;YAC7B,eAAe,EAAE,OAAO,CAAC,MAAM;YAC/B,gBAAgB,EAAE,GAAG;YACrB,aAAa,EAAE,KAAK;YACpB,SAAS,EAAE,KAAK;YAChB,WAAW,EAAE,IAAI;YACjB,iBAAiB,EAAE,MAAM;SAC1B;KACF,CAAC;AACJ,CAAC"}

package/dist/crush/index.d.ts

@@ -0,0 +1,16 @@
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/index.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/index.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Crush engine — pure tool-output compressors (Headroom ports).
+ * String in -> { output, stats } out. Zero deps, no I/O.
+ */
+export { CrushResult, CrushStats, CrushSourceType, estimateTokens, makePassthrough } from './crush-types';
+export { crushSearch, parseMatchLine, DEFAULT_SEARCH_CONFIG, SearchCompressorConfig } from './search-compressor';
+export { crushLog, detectFormat, classifyLevel, isSummaryLine, parseLogLines, normalizeForDedupe, DEFAULT_LOG_CONFIG, LogCompressorConfig, LogFormat, LogLevel, LogLine, } from './log-compressor';
+export { crushJson, computeKSplit, DEFAULT_JSON_CONFIG, JsonCrusherConfig } from './json-crusher';
+export { classifyArray, detectErrorItems, detectRareStatusValues, detectSequentialPattern, detectStructuralOutliers, ArrayType, JsonValue, } from './json-detectors';
+export { computeOptimalK, computeUniqueBigramCurve, findKnee } from './adaptive-sizer';
+export { ERROR_KEYWORDS, JSON_ERROR_KEYWORDS, containsKeyword, containsAnyKeyword, classifyImportance, } from './crush-keywords';

package/dist/crush/index.js

@@ -0,0 +1,48 @@
+"use strict";
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/index.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/index.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * Crush engine — pure tool-output compressors (Headroom ports).
+ * String in -> { output, stats } out. Zero deps, no I/O.
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.classifyImportance = exports.containsAnyKeyword = exports.containsKeyword = exports.JSON_ERROR_KEYWORDS = exports.ERROR_KEYWORDS = exports.findKnee = exports.computeUniqueBigramCurve = exports.computeOptimalK = exports.detectStructuralOutliers = exports.detectSequentialPattern = exports.detectRareStatusValues = exports.detectErrorItems = exports.classifyArray = exports.DEFAULT_JSON_CONFIG = exports.computeKSplit = exports.crushJson = exports.DEFAULT_LOG_CONFIG = exports.normalizeForDedupe = exports.parseLogLines = exports.isSummaryLine = exports.classifyLevel = exports.detectFormat = exports.crushLog = exports.DEFAULT_SEARCH_CONFIG = exports.parseMatchLine = exports.crushSearch = exports.makePassthrough = exports.estimateTokens = void 0;
+var crush_types_1 = require("./crush-types");
+Object.defineProperty(exports, "estimateTokens", { enumerable: true, get: function () { return crush_types_1.estimateTokens; } });
+Object.defineProperty(exports, "makePassthrough", { enumerable: true, get: function () { return crush_types_1.makePassthrough; } });
+var search_compressor_1 = require("./search-compressor");
+Object.defineProperty(exports, "crushSearch", { enumerable: true, get: function () { return search_compressor_1.crushSearch; } });
+Object.defineProperty(exports, "parseMatchLine", { enumerable: true, get: function () { return search_compressor_1.parseMatchLine; } });
+Object.defineProperty(exports, "DEFAULT_SEARCH_CONFIG", { enumerable: true, get: function () { return search_compressor_1.DEFAULT_SEARCH_CONFIG; } });
+var log_compressor_1 = require("./log-compressor");
+Object.defineProperty(exports, "crushLog", { enumerable: true, get: function () { return log_compressor_1.crushLog; } });
+Object.defineProperty(exports, "detectFormat", { enumerable: true, get: function () { return log_compressor_1.detectFormat; } });
+Object.defineProperty(exports, "classifyLevel", { enumerable: true, get: function () { return log_compressor_1.classifyLevel; } });
+Object.defineProperty(exports, "isSummaryLine", { enumerable: true, get: function () { return log_compressor_1.isSummaryLine; } });
+Object.defineProperty(exports, "parseLogLines", { enumerable: true, get: function () { return log_compressor_1.parseLogLines; } });
+Object.defineProperty(exports, "normalizeForDedupe", { enumerable: true, get: function () { return log_compressor_1.normalizeForDedupe; } });
+Object.defineProperty(exports, "DEFAULT_LOG_CONFIG", { enumerable: true, get: function () { return log_compressor_1.DEFAULT_LOG_CONFIG; } });
+var json_crusher_1 = require("./json-crusher");
+Object.defineProperty(exports, "crushJson", { enumerable: true, get: function () { return json_crusher_1.crushJson; } });
+Object.defineProperty(exports, "computeKSplit", { enumerable: true, get: function () { return json_crusher_1.computeKSplit; } });
+Object.defineProperty(exports, "DEFAULT_JSON_CONFIG", { enumerable: true, get: function () { return json_crusher_1.DEFAULT_JSON_CONFIG; } });
+var json_detectors_1 = require("./json-detectors");
+Object.defineProperty(exports, "classifyArray", { enumerable: true, get: function () { return json_detectors_1.classifyArray; } });
+Object.defineProperty(exports, "detectErrorItems", { enumerable: true, get: function () { return json_detectors_1.detectErrorItems; } });
+Object.defineProperty(exports, "detectRareStatusValues", { enumerable: true, get: function () { return json_detectors_1.detectRareStatusValues; } });
+Object.defineProperty(exports, "detectSequentialPattern", { enumerable: true, get: function () { return json_detectors_1.detectSequentialPattern; } });
+Object.defineProperty(exports, "detectStructuralOutliers", { enumerable: true, get: function () { return json_detectors_1.detectStructuralOutliers; } });
+var adaptive_sizer_1 = require("./adaptive-sizer");
+Object.defineProperty(exports, "computeOptimalK", { enumerable: true, get: function () { return adaptive_sizer_1.computeOptimalK; } });
+Object.defineProperty(exports, "computeUniqueBigramCurve", { enumerable: true, get: function () { return adaptive_sizer_1.computeUniqueBigramCurve; } });
+Object.defineProperty(exports, "findKnee", { enumerable: true, get: function () { return adaptive_sizer_1.findKnee; } });
+var crush_keywords_1 = require("./crush-keywords");
+Object.defineProperty(exports, "ERROR_KEYWORDS", { enumerable: true, get: function () { return crush_keywords_1.ERROR_KEYWORDS; } });
+Object.defineProperty(exports, "JSON_ERROR_KEYWORDS", { enumerable: true, get: function () { return crush_keywords_1.JSON_ERROR_KEYWORDS; } });
+Object.defineProperty(exports, "containsKeyword", { enumerable: true, get: function () { return crush_keywords_1.containsKeyword; } });
+Object.defineProperty(exports, "containsAnyKeyword", { enumerable: true, get: function () { return crush_keywords_1.containsAnyKeyword; } });
+Object.defineProperty(exports, "classifyImportance", { enumerable: true, get: function () { return crush_keywords_1.classifyImportance; } });
+//# sourceMappingURL=index.js.map

package/dist/crush/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/crush/index.ts"],"names":[],"mappings":";AAAA;;;;;;;;GAQG;;;AAEH,6CAA0G;AAAvD,6GAAA,cAAc,OAAA;AAAE,8GAAA,eAAe,OAAA;AAClF,yDAAiH;AAAxG,gHAAA,WAAW,OAAA;AAAE,mHAAA,cAAc,OAAA;AAAE,0HAAA,qBAAqB,OAAA;AAC3D,mDAY0B;AAXxB,0GAAA,QAAQ,OAAA;AACR,8GAAA,YAAY,OAAA;AACZ,+GAAA,aAAa,OAAA;AACb,+GAAA,aAAa,OAAA;AACb,+GAAA,aAAa,OAAA;AACb,oHAAA,kBAAkB,OAAA;AAClB,oHAAA,kBAAkB,OAAA;AAMpB,+CAAkG;AAAzF,yGAAA,SAAS,OAAA;AAAE,6GAAA,aAAa,OAAA;AAAE,mHAAA,mBAAmB,OAAA;AACtD,mDAQ0B;AAPxB,+GAAA,aAAa,OAAA;AACb,kHAAA,gBAAgB,OAAA;AAChB,wHAAA,sBAAsB,OAAA;AACtB,yHAAA,uBAAuB,OAAA;AACvB,0HAAA,wBAAwB,OAAA;AAI1B,mDAAuF;AAA9E,iHAAA,eAAe,OAAA;AAAE,0HAAA,wBAAwB,OAAA;AAAE,0GAAA,QAAQ,OAAA;AAC5D,mDAM0B;AALxB,gHAAA,cAAc,OAAA;AACd,qHAAA,mBAAmB,OAAA;AACnB,iHAAA,eAAe,OAAA;AACf,oHAAA,kBAAkB,OAAA;AAClB,oHAAA,kBAAkB,OAAA"}

package/dist/crush/json-crusher.d.ts

@@ -0,0 +1,38 @@
+/**
+ * MIRRORED FILE — canonical source: packages/core/src/crush/json-crusher.ts.
+ * Byte-identical copy lives at packages/cli/src/crush/json-crusher.ts (the CLI is a
+ * zero-dep artifact and cannot import @greprag/core). KEEP IN SYNC —
+ * the sync guard in tests/test-crush.cjs fails the build on drift.
+ *
+ * JSON-array crusher — TypeScript port of the SIMPLE core of Headroom's
+ * smart_crusher: statistical sampling + error preservation. Classifies
+ * each array, keeps errors / structural outliers / rare statuses /
+ * first+last blocks, and replaces the dropped bulk with a count marker
+ * string (`"[... N items omitted ...]"`).
+ *
+ * The full Headroom orchestration (TOIN semantics, relevance planning,
+ * lossless compaction, change points) is intentionally not ported.
+ */
+import { CrushResult } from './crush-types';
+export interface JsonCrusherConfig {
+    /** Arrays smaller than this are left alone. */
+    minItemsToAnalyze: number;
+    /** Target maximum items kept per crushed array. */
+    maxItemsAfterCrush: number;
+    /** Fraction of the budget pinned to the array head. */
+    firstFraction: number;
+    /** Fraction of the budget pinned to the array tail. */
+    lastFraction: number;
+    bias: number;
+}
+export declare const DEFAULT_JSON_CONFIG: JsonCrusherConfig;
+/** First/last slot split for a keep-budget. Never exceeds kTotal. */
+export declare function computeKSplit(kTotal: number, firstFraction: number, lastFraction: number): {
+    first: number;
+    last: number;
+};
+/**
+ * Compress JSON containing large arrays. Non-JSON input, or JSON with
+ * no qualifying arrays, passes through unchanged.
+ */
+export declare function crushJson(content: string, config?: Partial<JsonCrusherConfig>): CrushResult;