@codeledger/selector 0.1.1 → 0.2.1

package/dist/task-type.js

@@ -0,0 +1,115 @@
+const BUG_FIX_PATTERNS = [
+    'fix', 'bug', 'broken', 'crash', 'error', 'issue', 'wrong', 'fail',
+    'incorrect', 'invalid', 'null', 'undefined', 'exception', 'regression',
+    'patch', 'hotfix',
+];
+const FEATURE_ADD_PATTERNS = [
+    'add', 'create', 'implement', 'build', 'new', 'introduce', 'scaffold',
+    'feature', 'endpoint', 'component', 'page', 'route', 'service',
+];
+const REFACTOR_PATTERNS = [
+    'refactor', 'restructure', 'reorganize', 'clean', 'simplify', 'extract',
+    'rename', 'move', 'split', 'merge', 'consolidate', 'optimize',
+    'improve', 'modernize', 'upgrade', 'migrate', 'replace',
+];
+const TEST_UPDATE_PATTERNS = [
+    'test', 'tests', 'testing', 'spec', 'specs', 'coverage', 'assertion',
+    'mock', 'stub', 'fixture', 'jest', 'vitest', 'mocha', 'e2e',
+];
+const CONFIG_PATTERNS = [
+    'config', 'configuration', 'settings', 'setup', 'env', 'environment',
+    'ci', 'pipeline', 'deploy', 'docker', 'workflow', 'lint', 'eslint',
+    'prettier', 'tsconfig', 'webpack', 'vite',
+];
+/**
+ * Infer the task type from the task description.
+ * Returns the most likely type based on keyword pattern matching.
+ */
+export function inferTaskType(taskText) {
+    const lower = taskText.toLowerCase();
+    // Split on non-alphanumeric to strip punctuation (e.g. "fix," → "fix", "test." → "test")
+    const words = lower.split(/[^\p{L}\p{N}]+/u).filter((w) => w.length > 0);
+    // Score each type by keyword hits
+    const scores = {
+        bug_fix: 0,
+        feature_add: 0,
+        refactor: 0,
+        test_update: 0,
+        config: 0,
+        general: 0,
+    };
+    for (const word of words) {
+        if (BUG_FIX_PATTERNS.includes(word))
+            scores.bug_fix += 1;
+        if (FEATURE_ADD_PATTERNS.includes(word))
+            scores.feature_add += 1;
+        if (REFACTOR_PATTERNS.includes(word))
+            scores.refactor += 1;
+        if (TEST_UPDATE_PATTERNS.includes(word))
+            scores.test_update += 1;
+        if (CONFIG_PATTERNS.includes(word))
+            scores.config += 1;
+    }
+    // Find the highest scoring type
+    let bestType = 'general';
+    let bestScore = 0;
+    for (const [type, typeScore] of Object.entries(scores)) {
+        if (typeScore > bestScore) {
+            bestScore = typeScore;
+            bestType = type;
+        }
+    }
+    // Require at least 1 keyword hit to classify
+    return bestScore > 0 ? bestType : 'general';
+}
+/**
+ * Weight adjustment multipliers for each task type.
+ * Values > 1.0 emphasize, < 1.0 de-emphasize, 1.0 = no change.
+ */
+const WEIGHT_ADJUSTMENTS = {
+    bug_fix: {
+        error_infrastructure: 1.5,
+        recent_touch: 1.3,
+        churn: 1.2,
+        keyword: 1.1,
+    },
+    feature_add: {
+        centrality: 1.3,
+        keyword: 1.2,
+        test_relevance: 0.8,
+    },
+    refactor: {
+        churn: 1.4,
+        centrality: 1.3,
+        size_penalty: 0.7,
+    },
+    test_update: {
+        test_relevance: 1.5,
+        keyword: 1.2,
+        churn: 0.8,
+    },
+    config: {
+        keyword: 1.3,
+        churn: 0.7,
+        centrality: 0.8,
+    },
+    general: {},
+};
+/**
+ * Apply task-type weight adjustments to the base weights.
+ * Returns a new weights object with scaled values.
+ */
+export function applyTaskTypeWeights(baseWeights, taskType) {
+    const adjustments = WEIGHT_ADJUSTMENTS[taskType];
+    if (Object.keys(adjustments).length === 0)
+        return baseWeights;
+    const adjusted = { ...baseWeights };
+    for (const [key, multiplier] of Object.entries(adjustments)) {
+        const current = adjusted[key];
+        if (typeof current === 'number') {
+            adjusted[key] = current * multiplier;
+        }
+    }
+    return adjusted;
+}
+//# sourceMappingURL=task-type.js.map

package/dist/task-type.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"task-type.js","sourceRoot":"","sources":["../src/task-type.ts"],"names":[],"mappings":"AA0BA,MAAM,gBAAgB,GAAG;IACvB,KAAK,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM;IAClE,WAAW,EAAE,SAAS,EAAE,MAAM,EAAE,WAAW,EAAE,WAAW,EAAE,YAAY;IACtE,OAAO,EAAE,QAAQ;CAClB,CAAC;AAEF,MAAM,oBAAoB,GAAG;IAC3B,KAAK,EAAE,QAAQ,EAAE,WAAW,EAAE,OAAO,EAAE,KAAK,EAAE,WAAW,EAAE,UAAU;IACrE,SAAS,EAAE,UAAU,EAAE,WAAW,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS;CAC/D,CAAC;AAEF,MAAM,iBAAiB,GAAG;IACxB,UAAU,EAAE,aAAa,EAAE,YAAY,EAAE,OAAO,EAAE,UAAU,EAAE,SAAS;IACvE,QAAQ,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,aAAa,EAAE,UAAU;IAC7D,SAAS,EAAE,WAAW,EAAE,SAAS,EAAE,SAAS,EAAE,SAAS;CACxD,CAAC;AAEF,MAAM,oBAAoB,GAAG;IAC3B,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,WAAW;IACpE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK;CAC5D,CAAC;AAEF,MAAM,eAAe,GAAG;IACtB,QAAQ,EAAE,eAAe,EAAE,UAAU,EAAE,OAAO,EAAE,KAAK,EAAE,aAAa;IACpE,IAAI,EAAE,UAAU,EAAE,QAAQ,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,EAAE,QAAQ;IAClE,UAAU,EAAE,UAAU,EAAE,SAAS,EAAE,MAAM;CAC1C,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,aAAa,CAAC,QAAgB;IAC5C,MAAM,KAAK,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC;IACrC,yFAAyF;IACzF,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAEzE,kCAAkC;IAClC,MAAM,MAAM,GAA6B;QACvC,OAAO,EAAE,CAAC;QACV,WAAW,EAAE,CAAC;QACd,QAAQ,EAAE,CAAC;QACX,WAAW,EAAE,CAAC;QACd,MAAM,EAAE,CAAC;QACT,OAAO,EAAE,CAAC;KACX,CAAC;IAEF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,gBAAgB,CAAC,QAAQ,CAAC,IAAI,CAAC;YAAE,MAAM,CAAC,OAAO,IAAI,CAAC,CAAC;QACzD,IAAI,oBAAoB,CAAC,QAAQ,CAAC,IAAI,CAAC;YAAE,MAAM,CAAC,WAAW,IAAI,CAAC,CAAC;QACjE,IAAI,iBAAiB,CAAC,QAAQ,CAAC,IAAI,CAAC;YAAE,MAAM,CAAC,QAAQ,IAAI,CAAC,CAAC;QAC3D,IAAI,oBAAoB,CAAC,QAAQ,CAAC,IAAI,CAAC;YAAE,MAAM,CAAC,WAAW,IAAI,CAAC,CAAC;QACjE,IAAI,eAAe,CAAC,QAAQ,CAAC,IAAI,CAAC;YAAE,MAAM,CAAC,MAAM,IAAI,CAAC,CAAC;IACzD,CAAC;IAED,gCAAgC;IAChC,IAAI,QAAQ,GAAa,SAAS,CAAC;IACnC,IAAI,SAAS,GAAG,CAAC,CAAC;IAClB,KAAK,MAAM,CAAC,IAAI,EAAE,SAAS,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAA8B,EAAE,CAAC;QACpF,IAAI,SAAS,GAAG,SAAS,EAAE,CAAC;YAC1B,SAAS,GAAG,SAAS,CAAC;YACtB,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;IACH,CAAC;IAED,6CAA6C;IAC7C,OAAO,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC;AAC9C,CAAC;AAED;;;GAGG;AACH,MAAM,kBAAkB,GAAqE;IAC3F,OAAO,EAAE;QACP,oBAAoB,EAAE,GAAG;QACzB,YAAY,EAAE,GAAG;QACjB,KAAK,EAAE,GAAG;QACV,OAAO,EAAE,GAAG;KACb;IACD,WAAW,EAAE;QACX,UAAU,EAAE,GAAG;QACf,OAAO,EAAE,GAAG;QACZ,cAAc,EAAE,GAAG;KACpB;IACD,QAAQ,EAAE;QACR,KAAK,EAAE,GAAG;QACV,UAAU,EAAE,GAAG;QACf,YAAY,EAAE,GAAG;KAClB;IACD,WAAW,EAAE;QACX,cAAc,EAAE,GAAG;QACnB,OAAO,EAAE,GAAG;QACZ,KAAK,EAAE,GAAG;KACX;IACD,MAAM,EAAE;QACN,OAAO,EAAE,GAAG;QACZ,KAAK,EAAE,GAAG;QACV,UAAU,EAAE,GAAG;KAChB;IACD,OAAO,EAAE,EAAE;CACZ,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,oBAAoB,CAClC,WAA4B,EAC5B,QAAkB;IAElB,MAAM,WAAW,GAAG,kBAAkB,CAAC,QAAQ,CAAC,CAAC;IACjD,IAAI,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,WAAW,CAAC;IAE9D,MAAM,QAAQ,GAAG,EAAE,GAAG,WAAW,EAAE,CAAC;IACpC,KAAK,MAAM,CAAC,GAAG,EAAE,UAAU,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,WAAW,CAA2C,EAAE,CAAC;QACtG,MAAM,OAAO,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC;QAC9B,IAAI,OAAO,OAAO,KAAK,QAAQ,EAAE,CAAC;YAC/B,QAAmC,CAAC,GAAG,CAAC,GAAG,OAAO,GAAG,UAAU,CAAC;QACnE,CAAC;IACH,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/dist/todo-scan.d.ts

@@ -0,0 +1,31 @@
+/**
+ * TODO/FIXME Awareness (Feature 14)
+ *
+ * Scans file content for TODO, FIXME, HACK, and XXX comments.
+ * Returns a relevance score (0–1) based on the density and recency
+ * of these markers. Files with many TODOs are more likely to need
+ * attention and should get a small scoring boost.
+ */
+export interface TodoScanResult {
+    /** Total count of TODO-like markers found */
+    count: number;
+    /** Normalized score 0–1 (higher = more TODOs relative to file size) */
+    density: number;
+    /** The specific markers found */
+    markers: string[];
+}
+/**
+ * Scan file content for TODO/FIXME markers.
+ * Returns count, density (normalized by file length), and matched markers.
+ */
+export declare function scanTodos(content: string): TodoScanResult;
+/**
+ * Compute a TODO relevance boost for scoring.
+ * Returns a value between 0 and 1:
+ *   - 0: no TODOs
+ *   - 0.3: a few TODOs
+ *   - 0.6: moderate TODO density
+ *   - 1.0: high TODO density
+ */
+export declare function todoRelevanceScore(content: string): number;
+//# sourceMappingURL=todo-scan.d.ts.map

package/dist/todo-scan.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"todo-scan.d.ts","sourceRoot":"","sources":["../src/todo-scan.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAYH,MAAM,WAAW,cAAc;IAC7B,6CAA6C;IAC7C,KAAK,EAAE,MAAM,CAAC;IACd,uEAAuE;IACvE,OAAO,EAAE,MAAM,CAAC;IAChB,iCAAiC;IACjC,OAAO,EAAE,MAAM,EAAE,CAAC;CACnB;AAED;;;GAGG;AACH,wBAAgB,SAAS,CAAC,OAAO,EAAE,MAAM,GAAG,cAAc,CAyBzD;AAED;;;;;;;GAOG;AACH,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAM1D"}

package/dist/todo-scan.js

@@ -0,0 +1,63 @@
+/**
+ * TODO/FIXME Awareness (Feature 14)
+ *
+ * Scans file content for TODO, FIXME, HACK, and XXX comments.
+ * Returns a relevance score (0–1) based on the density and recency
+ * of these markers. Files with many TODOs are more likely to need
+ * attention and should get a small scoring boost.
+ */
+/** Patterns that indicate incomplete or problematic code */
+const TODO_PATTERNS = [
+    /\bTODO\b/i,
+    /\bFIXME\b/i,
+    /\bHACK\b/i,
+    /\bXXX\b/i,
+    /\bBUG\b(?:\s*#?\d+)/i,
+    /\bWORKAROUND\b/i,
+];
+/**
+ * Scan file content for TODO/FIXME markers.
+ * Returns count, density (normalized by file length), and matched markers.
+ */
+export function scanTodos(content) {
+    const lines = content.split('\n');
+    const markers = [];
+    let count = 0;
+    for (const line of lines) {
+        for (const pattern of TODO_PATTERNS) {
+            const match = pattern.exec(line);
+            if (match) {
+                count++;
+                const marker = match[0].toUpperCase();
+                if (!markers.includes(marker)) {
+                    markers.push(marker);
+                }
+                break; // only count once per line
+            }
+        }
+    }
+    // Density: TODOs per 100 lines, capped at 1.0
+    const density = lines.length > 0
+        ? Math.min(1, (count / lines.length) * 100)
+        : 0;
+    return { count, density, markers };
+}
+/**
+ * Compute a TODO relevance boost for scoring.
+ * Returns a value between 0 and 1:
+ *   - 0: no TODOs
+ *   - 0.3: a few TODOs
+ *   - 0.6: moderate TODO density
+ *   - 1.0: high TODO density
+ */
+export function todoRelevanceScore(content) {
+    const { count, density } = scanTodos(content);
+    if (count === 0)
+        return 0;
+    if (count <= 2)
+        return 0.3;
+    if (density < 0.5)
+        return 0.6;
+    return 1.0;
+}
+//# sourceMappingURL=todo-scan.js.map

package/dist/todo-scan.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"todo-scan.js","sourceRoot":"","sources":["../src/todo-scan.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,4DAA4D;AAC5D,MAAM,aAAa,GAAG;IACpB,WAAW;IACX,YAAY;IACZ,WAAW;IACX,UAAU;IACV,sBAAsB;IACtB,iBAAiB;CAClB,CAAC;AAWF;;;GAGG;AACH,MAAM,UAAU,SAAS,CAAC,OAAe;IACvC,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,KAAK,MAAM,OAAO,IAAI,aAAa,EAAE,CAAC;YACpC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjC,IAAI,KAAK,EAAE,CAAC;gBACV,KAAK,EAAE,CAAC;gBACR,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;gBACtC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;oBAC9B,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gBACvB,CAAC;gBACD,MAAM,CAAC,2BAA2B;YACpC,CAAC;QACH,CAAC;IACH,CAAC;IAED,8CAA8C;IAC9C,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC;QAC9B,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,GAAG,CAAC;QAC3C,CAAC,CAAC,CAAC,CAAC;IAEN,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC;AACrC,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,kBAAkB,CAAC,OAAe;IAChD,MAAM,EAAE,KAAK,EAAE,OAAO,EAAE,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IAC9C,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAC1B,IAAI,KAAK,IAAI,CAAC;QAAE,OAAO,GAAG,CAAC;IAC3B,IAAI,OAAO,GAAG,GAAG;QAAE,OAAO,GAAG,CAAC;IAC9B,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/token-calibration.d.ts

@@ -0,0 +1,27 @@
+/**
+ * Token Calibration (Feature 17)
+ *
+ * Language-aware token estimation that improves on the flat 4-tokens-per-line
+ * estimate. Different languages have different average token densities.
+ *
+ * Calibrated from OpenAI tokenizer empirical data:
+ *   - TypeScript/JavaScript: ~3.5 tokens/line (concise syntax)
+ *   - Python: ~3.2 tokens/line (significant whitespace, shorter lines)
+ *   - Go: ~3.8 tokens/line (verbose types, error handling)
+ *   - Rust: ~4.2 tokens/line (lifetime annotations, verbose syntax)
+ *   - Java: ~4.5 tokens/line (verbose OOP)
+ *   - JSON/YAML/Config: ~2.5 tokens/line (simple key-value)
+ *   - Markdown/Text: ~5.0 tokens/line (natural language)
+ *   - Default: 4.0 tokens/line (conservative estimate)
+ */
+/**
+ * Estimate tokens for a file based on line count and language.
+ * Falls back to the default rate of 4 tokens/line for unknown languages.
+ */
+export declare function estimateTokensCalibrated(lines: number, language: string): number;
+/**
+ * Estimate tokens from a file extension.
+ * Maps common extensions to their language for token rate lookup.
+ */
+export declare function estimateTokensByExtension(lines: number, extension: string): number;
+//# sourceMappingURL=token-calibration.d.ts.map

package/dist/token-calibration.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"token-calibration.d.ts","sourceRoot":"","sources":["../src/token-calibration.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AA6CH;;;GAGG;AACH,wBAAgB,wBAAwB,CACtC,KAAK,EAAE,MAAM,EACb,QAAQ,EAAE,MAAM,GACf,MAAM,CAGR;AAED;;;GAGG;AACH,wBAAgB,yBAAyB,CACvC,KAAK,EAAE,MAAM,EACb,SAAS,EAAE,MAAM,GAChB,MAAM,CAGR"}

package/dist/token-calibration.js

@@ -0,0 +1,113 @@
+/**
+ * Token Calibration (Feature 17)
+ *
+ * Language-aware token estimation that improves on the flat 4-tokens-per-line
+ * estimate. Different languages have different average token densities.
+ *
+ * Calibrated from OpenAI tokenizer empirical data:
+ *   - TypeScript/JavaScript: ~3.5 tokens/line (concise syntax)
+ *   - Python: ~3.2 tokens/line (significant whitespace, shorter lines)
+ *   - Go: ~3.8 tokens/line (verbose types, error handling)
+ *   - Rust: ~4.2 tokens/line (lifetime annotations, verbose syntax)
+ *   - Java: ~4.5 tokens/line (verbose OOP)
+ *   - JSON/YAML/Config: ~2.5 tokens/line (simple key-value)
+ *   - Markdown/Text: ~5.0 tokens/line (natural language)
+ *   - Default: 4.0 tokens/line (conservative estimate)
+ */
+const LANGUAGE_TOKEN_RATES = {
+    // Web languages
+    typescript: 3.5,
+    javascript: 3.5,
+    tsx: 3.5,
+    jsx: 3.5,
+    // Systems languages
+    go: 3.8,
+    rust: 4.2,
+    c: 3.8,
+    cpp: 4.0,
+    // Managed languages
+    java: 4.5,
+    kotlin: 3.8,
+    csharp: 4.2,
+    // Scripting languages
+    python: 3.2,
+    ruby: 3.0,
+    php: 3.8,
+    // Config/data formats
+    json: 2.5,
+    yaml: 2.5,
+    toml: 2.5,
+    xml: 4.0,
+    // Documentation
+    markdown: 5.0,
+    // Shell
+    bash: 3.5,
+    shell: 3.5,
+    // CSS
+    css: 3.0,
+    scss: 3.2,
+    // SQL
+    sql: 4.0,
+    // GraphQL/Proto
+    graphql: 3.0,
+    protobuf: 3.5,
+    prisma: 3.0,
+};
+/** Default rate when language is unknown */
+const DEFAULT_RATE = 4.0;
+/**
+ * Estimate tokens for a file based on line count and language.
+ * Falls back to the default rate of 4 tokens/line for unknown languages.
+ */
+export function estimateTokensCalibrated(lines, language) {
+    const rate = LANGUAGE_TOKEN_RATES[language.toLowerCase()] ?? DEFAULT_RATE;
+    return Math.ceil(lines * rate);
+}
+/**
+ * Estimate tokens from a file extension.
+ * Maps common extensions to their language for token rate lookup.
+ */
+export function estimateTokensByExtension(lines, extension) {
+    const language = extensionToLanguage(extension);
+    return estimateTokensCalibrated(lines, language);
+}
+const EXTENSION_MAP = {
+    '.ts': 'typescript',
+    '.tsx': 'tsx',
+    '.js': 'javascript',
+    '.jsx': 'jsx',
+    '.mjs': 'javascript',
+    '.cjs': 'javascript',
+    '.py': 'python',
+    '.go': 'go',
+    '.rs': 'rust',
+    '.java': 'java',
+    '.kt': 'kotlin',
+    '.cs': 'csharp',
+    '.rb': 'ruby',
+    '.php': 'php',
+    '.c': 'c',
+    '.cpp': 'cpp',
+    '.h': 'c',
+    '.hpp': 'cpp',
+    '.json': 'json',
+    '.yaml': 'yaml',
+    '.yml': 'yaml',
+    '.toml': 'toml',
+    '.xml': 'xml',
+    '.md': 'markdown',
+    '.mdx': 'markdown',
+    '.sh': 'bash',
+    '.bash': 'bash',
+    '.css': 'css',
+    '.scss': 'scss',
+    '.sql': 'sql',
+    '.graphql': 'graphql',
+    '.gql': 'graphql',
+    '.proto': 'protobuf',
+    '.prisma': 'prisma',
+};
+function extensionToLanguage(ext) {
+    return EXTENSION_MAP[ext.toLowerCase()] ?? 'unknown';
+}
+//# sourceMappingURL=token-calibration.js.map

package/dist/token-calibration.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"token-calibration.js","sourceRoot":"","sources":["../src/token-calibration.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,MAAM,oBAAoB,GAA2B;IACnD,gBAAgB;IAChB,UAAU,EAAE,GAAG;IACf,UAAU,EAAE,GAAG;IACf,GAAG,EAAE,GAAG;IACR,GAAG,EAAE,GAAG;IACR,oBAAoB;IACpB,EAAE,EAAE,GAAG;IACP,IAAI,EAAE,GAAG;IACT,CAAC,EAAE,GAAG;IACN,GAAG,EAAE,GAAG;IACR,oBAAoB;IACpB,IAAI,EAAE,GAAG;IACT,MAAM,EAAE,GAAG;IACX,MAAM,EAAE,GAAG;IACX,sBAAsB;IACtB,MAAM,EAAE,GAAG;IACX,IAAI,EAAE,GAAG;IACT,GAAG,EAAE,GAAG;IACR,sBAAsB;IACtB,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,IAAI,EAAE,GAAG;IACT,GAAG,EAAE,GAAG;IACR,gBAAgB;IAChB,QAAQ,EAAE,GAAG;IACb,QAAQ;IACR,IAAI,EAAE,GAAG;IACT,KAAK,EAAE,GAAG;IACV,MAAM;IACN,GAAG,EAAE,GAAG;IACR,IAAI,EAAE,GAAG;IACT,MAAM;IACN,GAAG,EAAE,GAAG;IACR,gBAAgB;IAChB,OAAO,EAAE,GAAG;IACZ,QAAQ,EAAE,GAAG;IACb,MAAM,EAAE,GAAG;CACZ,CAAC;AAEF,4CAA4C;AAC5C,MAAM,YAAY,GAAG,GAAG,CAAC;AAEzB;;;GAGG;AACH,MAAM,UAAU,wBAAwB,CACtC,KAAa,EACb,QAAgB;IAEhB,MAAM,IAAI,GAAG,oBAAoB,CAAC,QAAQ,CAAC,WAAW,EAAE,CAAC,IAAI,YAAY,CAAC;IAC1E,OAAO,IAAI,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,CAAC;AACjC,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,yBAAyB,CACvC,KAAa,EACb,SAAiB;IAEjB,MAAM,QAAQ,GAAG,mBAAmB,CAAC,SAAS,CAAC,CAAC;IAChD,OAAO,wBAAwB,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC;AACnD,CAAC;AAED,MAAM,aAAa,GAA2B;IAC5C,KAAK,EAAE,YAAY;IACnB,MAAM,EAAE,KAAK;IACb,KAAK,EAAE,YAAY;IACnB,MAAM,EAAE,KAAK;IACb,MAAM,EAAE,YAAY;IACpB,MAAM,EAAE,YAAY;IACpB,KAAK,EAAE,QAAQ;IACf,KAAK,EAAE,IAAI;IACX,KAAK,EAAE,MAAM;IACb,OAAO,EAAE,MAAM;IACf,KAAK,EAAE,QAAQ;IACf,KAAK,EAAE,QAAQ;IACf,KAAK,EAAE,MAAM;IACb,MAAM,EAAE,KAAK;IACb,IAAI,EAAE,GAAG;IACT,MAAM,EAAE,KAAK;IACb,IAAI,EAAE,GAAG;IACT,MAAM,EAAE,KAAK;IACb,OAAO,EAAE,MAAM;IACf,OAAO,EAAE,MAAM;IACf,MAAM,EAAE,MAAM;IACd,OAAO,EAAE,MAAM;IACf,MAAM,EAAE,KAAK;IACb,KAAK,EAAE,UAAU;IACjB,MAAM,EAAE,UAAU;IAClB,KAAK,EAAE,MAAM;IACb,OAAO,EAAE,MAAM;IACf,MAAM,EAAE,KAAK;IACb,OAAO,EAAE,MAAM;IACf,MAAM,EAAE,KAAK;IACb,UAAU,EAAE,SAAS;IACrB,MAAM,EAAE,SAAS;IACjB,QAAQ,EAAE,UAAU;IACpB,SAAS,EAAE,QAAQ;CACpB,CAAC;AAEF,SAAS,mBAAmB,CAAC,GAAW;IACtC,OAAO,aAAa,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,IAAI,SAAS,CAAC;AACvD,CAAC"}

package/dist/validation.d.ts

@@ -0,0 +1,7 @@
+import type { BundleFile, BundleWarning, ScoredFile } from '@codeledger/types';
+/**
+ * Validate a generated bundle and produce diagnostic warnings.
+ * Called after the bundle is fully assembled.
+ */
+export declare function validateBundle(files: BundleFile[], taskKeywords: string[], allScored: ScoredFile[], taskText: string): BundleWarning[];
+//# sourceMappingURL=validation.d.ts.map

package/dist/validation.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"validation.d.ts","sourceRoot":"","sources":["../src/validation.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,UAAU,EAAE,aAAa,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAmB/E;;;GAGG;AACH,wBAAgB,cAAc,CAC5B,KAAK,EAAE,UAAU,EAAE,EACnB,YAAY,EAAE,MAAM,EAAE,EACtB,SAAS,EAAE,UAAU,EAAE,EACvB,QAAQ,EAAE,MAAM,GACf,aAAa,EAAE,CAuGjB"}

package/dist/validation.js

@@ -0,0 +1,108 @@
+/**
+ * Bundle Validation — Pre-flight Warnings
+ *
+ * Catches obviously-bad bundles before the agent wastes turns. Validates
+ * bundle quality against the task description and raises warnings for
+ * common failure modes.
+ */
+/** Keywords that suggest the task involves testing */
+const TEST_KEYWORDS = ['test', 'tests', 'testing', 'spec', 'specs', 'jest', 'vitest', 'mocha'];
+/** Keywords that suggest the task involves configuration */
+const CONFIG_KEYWORDS = ['config', 'configuration', 'settings', 'setup', 'env', 'environment'];
+/** Keywords that suggest the task involves database operations */
+const DB_KEYWORDS = ['database', 'migration', 'schema', 'sql', 'query', 'model', 'table'];
+/**
+ * Validate a generated bundle and produce diagnostic warnings.
+ * Called after the bundle is fully assembled.
+ */
+export function validateBundle(files, taskKeywords, allScored, taskText) {
+    const warnings = [];
+    const taskLower = taskText.toLowerCase();
+    // Warning 1: No keyword matches at all
+    const hasKeywordMatch = files.some((f) => !f.is_stub && !f.is_exemplar && f.reasons.includes('keyword_match'));
+    if (!hasKeywordMatch && files.length > 0) {
+        warnings.push({
+            level: 'warning',
+            code: 'no_keyword_match',
+            message: 'No files matched task keywords. Bundle is based on churn/centrality signals only — consider rephrasing the task.',
+        });
+    }
+    // Warning 2: Task mentions tests but no test files in bundle
+    const mentionsTests = taskKeywords.some((kw) => TEST_KEYWORDS.includes(kw)) ||
+        TEST_KEYWORDS.some((kw) => taskLower.includes(kw));
+    const hasTestFile = files.some((f) => f.path.includes('.test.') || f.path.includes('.spec.') || f.path.includes('__tests__'));
+    if (mentionsTests && !hasTestFile) {
+        warnings.push({
+            level: 'warning',
+            code: 'missing_test_files',
+            message: 'Task mentions testing but no test files are in the bundle. Test files may have been filtered or out of budget.',
+        });
+    }
+    // Warning 3: Task mentions config but no config-like files
+    const mentionsConfig = taskKeywords.some((kw) => CONFIG_KEYWORDS.includes(kw));
+    const hasConfigFile = files.some((f) => {
+        const lower = f.path.toLowerCase();
+        return lower.includes('config') || lower.includes('.env') ||
+            lower.includes('tsconfig') || lower.includes('package.json');
+    });
+    if (mentionsConfig && !hasConfigFile) {
+        warnings.push({
+            level: 'info',
+            code: 'missing_config_files',
+            message: 'Task mentions configuration but no config files are in the bundle.',
+        });
+    }
+    // Warning 4: Task mentions DB/schema but no DB files
+    const mentionsDb = taskKeywords.some((kw) => DB_KEYWORDS.includes(kw));
+    const hasDbFile = files.some((f) => {
+        const lower = f.path.toLowerCase();
+        return lower.includes('model') || lower.includes('migration') ||
+            lower.includes('schema') || lower.includes('.sql') ||
+            lower.includes('db/') || lower.includes('database');
+    });
+    if (mentionsDb && !hasDbFile) {
+        warnings.push({
+            level: 'info',
+            code: 'missing_db_files',
+            message: 'Task mentions database/schema but no database files are in the bundle.',
+        });
+    }
+    // Warning 5: Very low score concentration — bundle is diluted
+    const nonStubFiles = files.filter((f) => !f.is_stub && !f.is_exemplar);
+    if (nonStubFiles.length >= 5) {
+        const topScore = nonStubFiles[0]?.score ?? 0;
+        const bottomScore = nonStubFiles[nonStubFiles.length - 1]?.score ?? 0;
+        if (topScore > 0 && bottomScore > 0) {
+            const ratio = bottomScore / topScore;
+            if (ratio > 0.85) {
+                warnings.push({
+                    level: 'info',
+                    code: 'score_plateau',
+                    message: 'File scores are very close together — no strong signal file. The task may be too broad or generic.',
+                });
+            }
+        }
+    }
+    // Warning 6: Bundle is empty
+    if (nonStubFiles.length === 0) {
+        warnings.push({
+            level: 'warning',
+            code: 'empty_bundle',
+            message: 'Bundle contains no source files. The task may not match any files in the repository.',
+        });
+    }
+    // Warning 7: High-scoring files were excluded (budget too tight)
+    if (allScored.length > 0 && nonStubFiles.length > 0) {
+        const lastIncludedScore = nonStubFiles[nonStubFiles.length - 1].score;
+        const excludedHighScorers = allScored.filter((s) => s.score > lastIncludedScore && !files.some((f) => f.path === s.path));
+        if (excludedHighScorers.length > 3) {
+            warnings.push({
+                level: 'info',
+                code: 'budget_squeeze',
+                message: `${excludedHighScorers.length} files with strong scores were excluded by budget constraints. Consider increasing --budget or --max-files.`,
+            });
+        }
+    }
+    return warnings;
+}
+//# sourceMappingURL=validation.js.map

package/dist/validation.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"validation.js","sourceRoot":"","sources":["../src/validation.ts"],"names":[],"mappings":"AAEA;;;;;;GAMG;AAEH,sDAAsD;AACtD,MAAM,aAAa,GAAG,CAAC,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;AAE/F,4DAA4D;AAC5D,MAAM,eAAe,GAAG,CAAC,QAAQ,EAAE,eAAe,EAAE,UAAU,EAAE,OAAO,EAAE,KAAK,EAAE,aAAa,CAAC,CAAC;AAE/F,kEAAkE;AAClE,MAAM,WAAW,GAAG,CAAC,UAAU,EAAE,WAAW,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;AAE1F;;;GAGG;AACH,MAAM,UAAU,cAAc,CAC5B,KAAmB,EACnB,YAAsB,EACtB,SAAuB,EACvB,QAAgB;IAEhB,MAAM,QAAQ,GAAoB,EAAE,CAAC;IACrC,MAAM,SAAS,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC;IAEzC,uCAAuC;IACvC,MAAM,eAAe,GAAG,KAAK,CAAC,IAAI,CAChC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,IAAI,CAAC,CAAC,CAAC,WAAW,IAAI,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,eAAe,CAAC,CAC3E,CAAC;IACF,IAAI,CAAC,eAAe,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzC,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,SAAS;YAChB,IAAI,EAAE,kBAAkB;YACxB,OAAO,EAAE,kHAAkH;SAC5H,CAAC,CAAC;IACL,CAAC;IAED,6DAA6D;IAC7D,MAAM,aAAa,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,aAAa,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;QACzE,aAAa,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC;IACrD,MAAM,WAAW,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CACnC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC,CACvF,CAAC;IACF,IAAI,aAAa,IAAI,CAAC,WAAW,EAAE,CAAC;QAClC,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,SAAS;YAChB,IAAI,EAAE,oBAAoB;YAC1B,OAAO,EAAE,gHAAgH;SAC1H,CAAC,CAAC;IACL,CAAC;IAED,2DAA2D;IAC3D,MAAM,cAAc,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,eAAe,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC;IAC/E,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;QACrC,MAAM,KAAK,GAAG,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;QACnC,OAAO,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC;YACvD,KAAK,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,cAAc,CAAC,CAAC;IACjE,CAAC,CAAC,CAAC;IACH,IAAI,cAAc,IAAI,CAAC,aAAa,EAAE,CAAC;QACrC,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,sBAAsB;YAC5B,OAAO,EAAE,oEAAoE;SAC9E,CAAC,CAAC;IACL,CAAC;IAED,qDAAqD;IACrD,MAAM,UAAU,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,WAAW,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC;IACvE,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;QACjC,MAAM,KAAK,GAAG,CAAC,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;QACnC,OAAO,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,WAAW,CAAC;YAC3D,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,MAAM,CAAC;YAClD,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC;IACxD,CAAC,CAAC,CAAC;IACH,IAAI,UAAU,IAAI,CAAC,SAAS,EAAE,CAAC;QAC7B,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,MAAM;YACb,IAAI,EAAE,kBAAkB;YACxB,OAAO,EAAE,wEAAwE;SAClF,CAAC,CAAC;IACL,CAAC;IAED,8DAA8D;IAC9D,MAAM,YAAY,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,IAAI,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC;IACvE,IAAI,YAAY,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;QAC7B,MAAM,QAAQ,GAAG,YAAY,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,CAAC;QAC7C,MAAM,WAAW,GAAG,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,CAAC;QACtE,IAAI,QAAQ,GAAG,CAAC,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;YACpC,MAAM,KAAK,GAAG,WAAW,GAAG,QAAQ,CAAC;YACrC,IAAI,KAAK,GAAG,IAAI,EAAE,CAAC;gBACjB,QAAQ,CAAC,IAAI,CAAC;oBACZ,KAAK,EAAE,MAAM;oBACb,IAAI,EAAE,eAAe;oBACrB,OAAO,EAAE,oGAAoG;iBAC9G,CAAC,CAAC;YACL,CAAC;QACH,CAAC;IACH,CAAC;IAED,6BAA6B;IAC7B,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC9B,QAAQ,CAAC,IAAI,CAAC;YACZ,KAAK,EAAE,SAAS;YAChB,IAAI,EAAE,cAAc;YACpB,OAAO,EAAE,sFAAsF;SAChG,CAAC,CAAC;IACL,CAAC;IAED,iEAAiE;IACjE,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpD,MAAM,iBAAiB,GAAG,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC,KAAK,CAAC;QACvE,MAAM,mBAAmB,GAAG,SAAS,CAAC,MAAM,CAC1C,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,iBAAiB,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,CAAC,CAC5E,CAAC;QACF,IAAI,mBAAmB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACnC,QAAQ,CAAC,IAAI,CAAC;gBACZ,KAAK,EAAE,MAAM;gBACb,IAAI,EAAE,gBAAgB;gBACtB,OAAO,EAAE,GAAG,mBAAmB,CAAC,MAAM,6GAA6G;aACpJ,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@codeledger/selector",
-  "version": "0.1.1",
+  "version": "0.2.1",
   "type": "module",
   "description": "Deterministic context selection algorithm for CodeLedger",
   "license": "Apache-2.0",
@@ -24,7 +24,7 @@
     }
   },
   "dependencies": {
-    "@codeledger/types": "0.1.1"
+    "@codeledger/types": "0.2.1"
   },
   "devDependencies": {
     "typescript": "^5.4.0"