@skaft/hamr 0.4.1

package/dist/hamr/providers/repair/json-repair.js

@@ -0,0 +1,315 @@
+/**
+ * JSON repair — bounded auto-recovery for local-model tool-call JSON.
+ *
+ * Local models frequently emit malformed JSON with:
+ * - Trailing commas: {"name": "read", "args": {"path": "x",}}
+ * - Unescaped inner quotes: {"query": "find "foo" in bar"}
+ * - Truncated objects: {"name": "bash", "args": {"command": "npm te
+ * - Missing closing braces: {"name": "edit", "args": {"path": "x", "oldStr": "y"
+ * - Mixed format: prose + partial JSON blocks
+ *
+ * Each repair is recorded in `fixes[]` for debugging. Returns `null` when
+ * the input is unrepairable (garbage, empty, or too broken to guess).
+ */
+/**
+ * Attempt to repair malformed JSON produced by a local model.
+ *
+ * Repairs are applied in increasing order of invasiveness:
+ * 1. Trim whitespace and surrounding noise
+ * 2. Fix trailing commas
+ * 3. Balance braces/brackets
+ * 4. Heuristic inner-quote repair
+ *
+ * Returns `null` if the string is unrepairable.
+ */
+export function repairJson(raw) {
+    const fixes = [];
+    let working = raw.trim();
+    if (!working)
+        return null;
+    // Step 0: Extract the probable JSON region from surrounding text
+    const extracted = extractJsonRegion(working);
+    if (extracted !== working) {
+        fixes.push("extracted JSON from surrounding text");
+        working = extracted;
+    }
+    if (!working)
+        return null;
+    // Step 1: Fix trailing commas
+    const trailingFixed = fixTrailingCommas(working);
+    if (trailingFixed !== working) {
+        fixes.push("removed trailing commas");
+        working = trailingFixed;
+    }
+    // Step 2: Fix unescaped inner quotes (heuristic, conservative)
+    const quotesFixed = fixInnerQuotes(working);
+    if (quotesFixed !== working) {
+        fixes.push("escaped inner quotes");
+        working = quotesFixed;
+    }
+    // Step 3: Balance braces and brackets
+    const balanced = balanceBraces(working);
+    if (balanced !== working) {
+        fixes.push("balanced braces/brackets");
+        working = balanced;
+    }
+    // Step 4: Try to parse. If it works, return the repaired text.
+    if (isValidJson(working)) {
+        return { repaired: working, fixes };
+    }
+    // Step 5: As a last resort, try trimming from the end (for truncated output)
+    const truncated = fixTruncatedObject(working);
+    if (truncated && truncated !== working) {
+        fixes.push("recovered truncated object");
+        working = truncated;
+        if (isValidJson(working)) {
+            return { repaired: working, fixes };
+        }
+    }
+    // If nothing works, return null
+    return null;
+}
+// ─── Internal helpers ─────────────────────────────────────
+/**
+ * Extract a JSON object or array from surrounding text.
+ * Looks for the most likely JSON region.
+ */
+function extractJsonRegion(text) {
+    // If text starts with { or [, it's likely already focused
+    if (text.startsWith("{") || text.startsWith("["))
+        return text;
+    // Try to find JSON object
+    const objStart = text.indexOf("{");
+    const arrStart = text.indexOf("[");
+    const start = objStart === -1 ? arrStart : arrStart === -1 ? objStart : Math.min(objStart, arrStart);
+    if (start === -1)
+        return text; // no JSON found, return as-is
+    // Try to find matching closer
+    const closer = text[start] === "{" ? "}" : "]";
+    const opener = text[start];
+    let depth = 0;
+    let inString = false;
+    let escape = false;
+    for (let i = start; i < text.length; i++) {
+        const ch = text[i];
+        if (escape) {
+            escape = false;
+            continue;
+        }
+        if (ch === "\\" && inString) {
+            escape = true;
+            continue;
+        }
+        if (ch === '"') {
+            inString = !inString;
+            continue;
+        }
+        if (inString)
+            continue;
+        if (ch === opener)
+            depth++;
+        else if (ch === closer) {
+            depth--;
+            if (depth === 0)
+                return text.slice(start, i + 1);
+        }
+    }
+    // If no matching closer found, return from start to end
+    return text.slice(start);
+}
+/**
+ * Fix trailing commas before } or ].
+ */
+function fixTrailingCommas(json) {
+    // Replace comma before } or ] with just the brace
+    return json.replace(/,\s*([}\]])/g, "$1");
+}
+/**
+ * Heuristic: fix unescaped inner double-quotes inside string values.
+ *
+ * Pattern: string contains " like "this" which should be escaped.
+ * This is conservative — only targets patterns where JSON parsing fails
+ * due to quote collision within string values.
+ */
+function fixInnerQuotes(json) {
+    // This is inherently heuristic. We target the common case:
+    // "key": "value with "unescaped" quotes inside"
+    //
+    // Strategy: find string values that contain un-escaped quotes.
+    // We'll use the fact that a well-formed JSON string:
+    // " starts a string, " ends it. Within a string, " is escaped as \".
+    //
+    // When we see a " that's inside what should be a string value
+    // (preceded by :" or ," or {" or ["), and the next unescaped " would
+    // make the JSON invalid, we escape it.
+    const result = [];
+    let inString = false;
+    let inKey = false; // we're inside a key (before the colon)
+    let escape = false;
+    let colonSeen = false; // just saw colon, next value is a string likely
+    for (let i = 0; i < json.length; i++) {
+        const ch = json[i];
+        if (escape) {
+            escape = false;
+            result.push(ch);
+            continue;
+        }
+        if (ch === "\\") {
+            escape = true;
+            result.push(ch);
+            continue;
+        }
+        if (ch === '"') {
+            if (!inString) {
+                inString = true;
+                inKey = !colonSeen; // after colon, we're in a value
+                colonSeen = false;
+                result.push(ch);
+            }
+            else {
+                // Closing a string — peek ahead to see if this is really the end
+                // or if there's more string content that should have been escaped.
+                const peek = json.slice(i + 1);
+                const restTrimmed = peek.trimStart();
+                // If the next non-whitespace is NOT one of: , } ] : (or end of input),
+                // this is likely an inner quote that should have been escaped.
+                // But only fix if we're in a string value (not a key).
+                if (!inKey &&
+                    restTrimmed.length > 0 &&
+                    !restTrimmed.startsWith(",") &&
+                    !restTrimmed.startsWith("}") &&
+                    !restTrimmed.startsWith("]") &&
+                    !restTrimmed.startsWith(":") &&
+                    !restTrimmed.startsWith("\n") &&
+                    !restTrimmed.startsWith("\r")) {
+                    // This is likely an inner quote — escape it
+                    result.push("\\");
+                }
+                inString = false;
+                inKey = false;
+                result.push(ch);
+            }
+            continue;
+        }
+        if (!inString && ch === ":") {
+            colonSeen = true;
+        }
+        else if (!inString && (ch === "," || ch === "{" || ch === "[")) {
+            colonSeen = false;
+        }
+        else if (!inString && /\S/.test(ch)) {
+            colonSeen = false;
+        }
+        result.push(ch);
+    }
+    return result.join("");
+}
+/**
+ * Balance unclosed braces and brackets using a stack.
+ * Adds missing closing chars at the end.
+ */
+function balanceBraces(json) {
+    const stack = [];
+    let inString = false;
+    let escape = false;
+    for (const ch of json) {
+        if (escape) {
+            escape = false;
+            continue;
+        }
+        if (ch === "\\" && inString) {
+            escape = true;
+            continue;
+        }
+        if (ch === '"') {
+            inString = !inString;
+            continue;
+        }
+        if (inString)
+            continue;
+        if (ch === "{")
+            stack.push("}");
+        else if (ch === "[")
+            stack.push("]");
+        else if (ch === "}" || ch === "]") {
+            if (stack.length > 0 && stack[stack.length - 1] === ch) {
+                stack.pop();
+            }
+            // else: extra closer — we don't try to fix that
+        }
+    }
+    // Only repair if the imbalance is reasonable
+    if (stack.length === 0)
+        return json;
+    if (stack.length > 20)
+        return json; // too broken
+    return json + stack.reverse().join("");
+}
+/**
+ * Attempt to recover a truncated JSON object.
+ * Tries progressively removing the last character, then last comma-prefixed
+ * segment, then adds a synthetic closing brace.
+ */
+function fixTruncatedObject(json) {
+    // Strategy: try to close the object by adding missing braces
+    const balanced = balanceBraces(json);
+    if (balanced !== json)
+        return balanced;
+    // If balancing didn't help and the string ends mid-value,
+    // try truncating to the last complete key-value pair
+    if (json.endsWith(",")) {
+        // Already clean — just needs closing
+        const cleaned = `${json}}`;
+        if (isValidJson(cleaned))
+            return cleaned;
+        // Try removing trailing comma and closing
+        const alt = `${json.slice(0, -1)}}`;
+        if (isValidJson(alt))
+            return alt;
+    }
+    // Try appending synthetic closing
+    const synthetic = json.replace(/[^}\]]$/, "").trimEnd();
+    const braces = countUnclosed(synthetic);
+    if (braces > 0 && braces <= 5) {
+        const closed = synthetic + "}".repeat(braces);
+        if (isValidJson(closed))
+            return closed;
+    }
+    return null;
+}
+function countUnclosed(json) {
+    let depth = 0;
+    let inString = false;
+    let escape = false;
+    for (const ch of json) {
+        if (escape) {
+            escape = false;
+            continue;
+        }
+        if (ch === "\\" && inString) {
+            escape = true;
+            continue;
+        }
+        if (ch === '"') {
+            inString = !inString;
+            continue;
+        }
+        if (inString)
+            continue;
+        if (ch === "{")
+            depth++;
+        else if (ch === "}")
+            depth--;
+    }
+    return Math.max(0, depth);
+}
+function isValidJson(text) {
+    try {
+        JSON.parse(text);
+        return true;
+    }
+    catch {
+        return false;
+    }
+}
+//# sourceMappingURL=json-repair.js.map

package/dist/hamr/providers/repair/json-repair.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"json-repair.js","sourceRoot":"","sources":["../../../../src/hamr/providers/repair/json-repair.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAOH;;;;;;;;;;GAUG;AACH,MAAM,UAAU,UAAU,CAAC,GAAW;IACrC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;IAEzB,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAE1B,iEAAiE;IACjE,MAAM,SAAS,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;IAC7C,IAAI,SAAS,KAAK,OAAO,EAAE,CAAC;QAC3B,KAAK,CAAC,IAAI,CAAC,sCAAsC,CAAC,CAAC;QACnD,OAAO,GAAG,SAAS,CAAC;IACrB,CAAC;IAED,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAE1B,8BAA8B;IAC9B,MAAM,aAAa,GAAG,iBAAiB,CAAC,OAAO,CAAC,CAAC;IACjD,IAAI,aAAa,KAAK,OAAO,EAAE,CAAC;QAC/B,KAAK,CAAC,IAAI,CAAC,yBAAyB,CAAC,CAAC;QACtC,OAAO,GAAG,aAAa,CAAC;IACzB,CAAC;IAED,+DAA+D;IAC/D,MAAM,WAAW,GAAG,cAAc,CAAC,OAAO,CAAC,CAAC;IAC5C,IAAI,WAAW,KAAK,OAAO,EAAE,CAAC;QAC7B,KAAK,CAAC,IAAI,CAAC,sBAAsB,CAAC,CAAC;QACnC,OAAO,GAAG,WAAW,CAAC;IACvB,CAAC;IAED,sCAAsC;IACtC,MAAM,QAAQ,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC;IACxC,IAAI,QAAQ,KAAK,OAAO,EAAE,CAAC;QAC1B,KAAK,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAC;QACvC,OAAO,GAAG,QAAQ,CAAC;IACpB,CAAC;IAED,+DAA+D;IAC/D,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;QAC1B,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;IACrC,CAAC;IAED,6EAA6E;IAC7E,MAAM,SAAS,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAC9C,IAAI,SAAS,IAAI,SAAS,KAAK,OAAO,EAAE,CAAC;QACxC,KAAK,CAAC,IAAI,CAAC,4BAA4B,CAAC,CAAC;QACzC,OAAO,GAAG,SAAS,CAAC;QACpB,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;YAC1B,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;QACrC,CAAC;IACF,CAAC;IAED,gCAAgC;IAChC,OAAO,IAAI,CAAC;AACb,CAAC;AAED,6DAA6D;AAE7D;;;GAGG;AACH,SAAS,iBAAiB,CAAC,IAAY;IACtC,0DAA0D;IAC1D,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAE9D,0BAA0B;IAC1B,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IACnC,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IAEnC,MAAM,KAAK,GAAG,QAAQ,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;IACrG,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC,CAAC,8BAA8B;IAE7D,8BAA8B;IAC9B,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;IAC/C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;IAC3B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,IAAI,MAAM,GAAG,KAAK,CAAC;IAEnB,KAAK,IAAI,CAAC,GAAG,KAAK,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC1C,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACnB,IAAI,MAAM,EAAE,CAAC;YACZ,MAAM,GAAG,KAAK,CAAC;YACf,SAAS;QACV,CAAC;QACD,IAAI,EAAE,KAAK,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC7B,MAAM,GAAG,IAAI,CAAC;YACd,SAAS;QACV,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YAChB,QAAQ,GAAG,CAAC,QAAQ,CAAC;YACrB,SAAS;QACV,CAAC;QACD,IAAI,QAAQ;YAAE,SAAS;QACvB,IAAI,EAAE,KAAK,MAAM;YAAE,KAAK,EAAE,CAAC;aACtB,IAAI,EAAE,KAAK,MAAM,EAAE,CAAC;YACxB,KAAK,EAAE,CAAC;YACR,IAAI,KAAK,KAAK,CAAC;gBAAE,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QAClD,CAAC;IACF,CAAC;IAED,wDAAwD;IACxD,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;AAC1B,CAAC;AAED;;GAEG;AACH,SAAS,iBAAiB,CAAC,IAAY;IACtC,kDAAkD;IAClD,OAAO,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,IAAI,CAAC,CAAC;AAC3C,CAAC;AAED;;;;;;GAMG;AACH,SAAS,cAAc,CAAC,IAAY;IACnC,2DAA2D;IAC3D,gDAAgD;IAChD,EAAE;IACF,+DAA+D;IAC/D,qDAAqD;IACrD,qEAAqE;IACrE,EAAE;IACF,8DAA8D;IAC9D,qEAAqE;IACrE,uCAAuC;IAEvC,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,IAAI,KAAK,GAAG,KAAK,CAAC,CAAC,wCAAwC;IAC3D,IAAI,MAAM,GAAG,KAAK,CAAC;IACnB,IAAI,SAAS,GAAG,KAAK,CAAC,CAAC,gDAAgD;IAEvE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAEnB,IAAI,MAAM,EAAE,CAAC;YACZ,MAAM,GAAG,KAAK,CAAC;YACf,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAChB,SAAS;QACV,CAAC;QAED,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YACjB,MAAM,GAAG,IAAI,CAAC;YACd,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAChB,SAAS;QACV,CAAC;QAED,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YAChB,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACf,QAAQ,GAAG,IAAI,CAAC;gBAChB,KAAK,GAAG,CAAC,SAAS,CAAC,CAAC,gCAAgC;gBACpD,SAAS,GAAG,KAAK,CAAC;gBAClB,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACjB,CAAC;iBAAM,CAAC;gBACP,iEAAiE;gBACjE,mEAAmE;gBACnE,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;gBAC/B,MAAM,WAAW,GAAG,IAAI,CAAC,SAAS,EAAE,CAAC;gBAErC,uEAAuE;gBACvE,+DAA+D;gBAC/D,uDAAuD;gBACvD,IACC,CAAC,KAAK;oBACN,WAAW,CAAC,MAAM,GAAG,CAAC;oBACtB,CAAC,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC;oBAC5B,CAAC,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC;oBAC5B,CAAC,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC;oBAC5B,CAAC,WAAW,CAAC,UAAU,CAAC,GAAG,CAAC;oBAC5B,CAAC,WAAW,CAAC,UAAU,CAAC,IAAI,CAAC;oBAC7B,CAAC,WAAW,CAAC,UAAU,CAAC,IAAI,CAAC,EAC5B,CAAC;oBACF,4CAA4C;oBAC5C,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACnB,CAAC;gBACD,QAAQ,GAAG,KAAK,CAAC;gBACjB,KAAK,GAAG,KAAK,CAAC;gBACd,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACjB,CAAC;YACD,SAAS;QACV,CAAC;QAED,IAAI,CAAC,QAAQ,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YAC7B,SAAS,GAAG,IAAI,CAAC;QAClB,CAAC;aAAM,IAAI,CAAC,QAAQ,IAAI,CAAC,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,GAAG,CAAC,EAAE,CAAC;YAClE,SAAS,GAAG,KAAK,CAAC;QACnB,CAAC;aAAM,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC;YACvC,SAAS,GAAG,KAAK,CAAC;QACnB,CAAC;QAED,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACjB,CAAC;IAED,OAAO,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxB,CAAC;AAED;;;GAGG;AACH,SAAS,aAAa,CAAC,IAAY;IAClC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,IAAI,MAAM,GAAG,KAAK,CAAC;IAEnB,KAAK,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC;QACvB,IAAI,MAAM,EAAE,CAAC;YACZ,MAAM,GAAG,KAAK,CAAC;YACf,SAAS;QACV,CAAC;QACD,IAAI,EAAE,KAAK,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC7B,MAAM,GAAG,IAAI,CAAC;YACd,SAAS;QACV,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YAChB,QAAQ,GAAG,CAAC,QAAQ,CAAC;YACrB,SAAS;QACV,CAAC;QACD,IAAI,QAAQ;YAAE,SAAS;QAEvB,IAAI,EAAE,KAAK,GAAG;YAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;aAC3B,IAAI,EAAE,KAAK,GAAG;YAAE,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;aAChC,IAAI,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACnC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC;gBACxD,KAAK,CAAC,GAAG,EAAE,CAAC;YACb,CAAC;YACD,gDAAgD;QACjD,CAAC;IACF,CAAC;IAED,6CAA6C;IAC7C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IACpC,IAAI,KAAK,CAAC,MAAM,GAAG,EAAE;QAAE,OAAO,IAAI,CAAC,CAAC,aAAa;IACjD,OAAO,IAAI,GAAG,KAAK,CAAC,OAAO,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACxC,CAAC;AAED;;;;GAIG;AACH,SAAS,kBAAkB,CAAC,IAAY;IACvC,6DAA6D;IAC7D,MAAM,QAAQ,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACrC,IAAI,QAAQ,KAAK,IAAI;QAAE,OAAO,QAAQ,CAAC;IAEvC,0DAA0D;IAC1D,qDAAqD;IACrD,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QACxB,qCAAqC;QACrC,MAAM,OAAO,GAAG,GAAG,IAAI,GAAG,CAAC;QAC3B,IAAI,WAAW,CAAC,OAAO,CAAC;YAAE,OAAO,OAAO,CAAC;QAEzC,0CAA0C;QAC1C,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;QACpC,IAAI,WAAW,CAAC,GAAG,CAAC;YAAE,OAAO,GAAG,CAAC;IAClC,CAAC;IAED,kCAAkC;IAClC,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IACxD,MAAM,MAAM,GAAG,aAAa,CAAC,SAAS,CAAC,CAAC;IACxC,IAAI,MAAM,GAAG,CAAC,IAAI,MAAM,IAAI,CAAC,EAAE,CAAC;QAC/B,MAAM,MAAM,GAAG,SAAS,GAAG,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;QAC9C,IAAI,WAAW,CAAC,MAAM,CAAC;YAAE,OAAO,MAAM,CAAC;IACxC,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC;AAED,SAAS,aAAa,CAAC,IAAY;IAClC,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,IAAI,MAAM,GAAG,KAAK,CAAC;IAEnB,KAAK,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC;QACvB,IAAI,MAAM,EAAE,CAAC;YACZ,MAAM,GAAG,KAAK,CAAC;YACf,SAAS;QACV,CAAC;QACD,IAAI,EAAE,KAAK,IAAI,IAAI,QAAQ,EAAE,CAAC;YAC7B,MAAM,GAAG,IAAI,CAAC;YACd,SAAS;QACV,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YAChB,QAAQ,GAAG,CAAC,QAAQ,CAAC;YACrB,SAAS;QACV,CAAC;QACD,IAAI,QAAQ;YAAE,SAAS;QACvB,IAAI,EAAE,KAAK,GAAG;YAAE,KAAK,EAAE,CAAC;aACnB,IAAI,EAAE,KAAK,GAAG;YAAE,KAAK,EAAE,CAAC;IAC9B,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;AAC3B,CAAC;AAED,SAAS,WAAW,CAAC,IAAY;IAChC,IAAI,CAAC;QACJ,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACjB,OAAO,IAAI,CAAC;IACb,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAC;IACd,CAAC;AACF,CAAC","sourcesContent":["/**\n * JSON repair — bounded auto-recovery for local-model tool-call JSON.\n *\n * Local models frequently emit malformed JSON with:\n * - Trailing commas: {\"name\": \"read\", \"args\": {\"path\": \"x\",}}\n * - Unescaped inner quotes: {\"query\": \"find \"foo\" in bar\"}\n * - Truncated objects: {\"name\": \"bash\", \"args\": {\"command\": \"npm te\n * - Missing closing braces: {\"name\": \"edit\", \"args\": {\"path\": \"x\", \"oldStr\": \"y\"\n * - Mixed format: prose + partial JSON blocks\n *\n * Each repair is recorded in `fixes[]` for debugging. Returns `null` when\n * the input is unrepairable (garbage, empty, or too broken to guess).\n */\n\nexport interface RepairResult {\n\trepaired: string;\n\tfixes: string[];\n}\n\n/**\n * Attempt to repair malformed JSON produced by a local model.\n *\n * Repairs are applied in increasing order of invasiveness:\n * 1. Trim whitespace and surrounding noise\n * 2. Fix trailing commas\n * 3. Balance braces/brackets\n * 4. Heuristic inner-quote repair\n *\n * Returns `null` if the string is unrepairable.\n */\nexport function repairJson(raw: string): RepairResult | null {\n\tconst fixes: string[] = [];\n\tlet working = raw.trim();\n\n\tif (!working) return null;\n\n\t// Step 0: Extract the probable JSON region from surrounding text\n\tconst extracted = extractJsonRegion(working);\n\tif (extracted !== working) {\n\t\tfixes.push(\"extracted JSON from surrounding text\");\n\t\tworking = extracted;\n\t}\n\n\tif (!working) return null;\n\n\t// Step 1: Fix trailing commas\n\tconst trailingFixed = fixTrailingCommas(working);\n\tif (trailingFixed !== working) {\n\t\tfixes.push(\"removed trailing commas\");\n\t\tworking = trailingFixed;\n\t}\n\n\t// Step 2: Fix unescaped inner quotes (heuristic, conservative)\n\tconst quotesFixed = fixInnerQuotes(working);\n\tif (quotesFixed !== working) {\n\t\tfixes.push(\"escaped inner quotes\");\n\t\tworking = quotesFixed;\n\t}\n\n\t// Step 3: Balance braces and brackets\n\tconst balanced = balanceBraces(working);\n\tif (balanced !== working) {\n\t\tfixes.push(\"balanced braces/brackets\");\n\t\tworking = balanced;\n\t}\n\n\t// Step 4: Try to parse. If it works, return the repaired text.\n\tif (isValidJson(working)) {\n\t\treturn { repaired: working, fixes };\n\t}\n\n\t// Step 5: As a last resort, try trimming from the end (for truncated output)\n\tconst truncated = fixTruncatedObject(working);\n\tif (truncated && truncated !== working) {\n\t\tfixes.push(\"recovered truncated object\");\n\t\tworking = truncated;\n\t\tif (isValidJson(working)) {\n\t\t\treturn { repaired: working, fixes };\n\t\t}\n\t}\n\n\t// If nothing works, return null\n\treturn null;\n}\n\n// ─── Internal helpers ─────────────────────────────────────\n\n/**\n * Extract a JSON object or array from surrounding text.\n * Looks for the most likely JSON region.\n */\nfunction extractJsonRegion(text: string): string {\n\t// If text starts with { or [, it's likely already focused\n\tif (text.startsWith(\"{\") || text.startsWith(\"[\")) return text;\n\n\t// Try to find JSON object\n\tconst objStart = text.indexOf(\"{\");\n\tconst arrStart = text.indexOf(\"[\");\n\n\tconst start = objStart === -1 ? arrStart : arrStart === -1 ? objStart : Math.min(objStart, arrStart);\n\tif (start === -1) return text; // no JSON found, return as-is\n\n\t// Try to find matching closer\n\tconst closer = text[start] === \"{\" ? \"}\" : \"]\";\n\tconst opener = text[start];\n\tlet depth = 0;\n\tlet inString = false;\n\tlet escape = false;\n\n\tfor (let i = start; i < text.length; i++) {\n\t\tconst ch = text[i];\n\t\tif (escape) {\n\t\t\tescape = false;\n\t\t\tcontinue;\n\t\t}\n\t\tif (ch === \"\\\\\" && inString) {\n\t\t\tescape = true;\n\t\t\tcontinue;\n\t\t}\n\t\tif (ch === '\"') {\n\t\t\tinString = !inString;\n\t\t\tcontinue;\n\t\t}\n\t\tif (inString) continue;\n\t\tif (ch === opener) depth++;\n\t\telse if (ch === closer) {\n\t\t\tdepth--;\n\t\t\tif (depth === 0) return text.slice(start, i + 1);\n\t\t}\n\t}\n\n\t// If no matching closer found, return from start to end\n\treturn text.slice(start);\n}\n\n/**\n * Fix trailing commas before } or ].\n */\nfunction fixTrailingCommas(json: string): string {\n\t// Replace comma before } or ] with just the brace\n\treturn json.replace(/,\\s*([}\\]])/g, \"$1\");\n}\n\n/**\n * Heuristic: fix unescaped inner double-quotes inside string values.\n *\n * Pattern: string contains \" like \"this\" which should be escaped.\n * This is conservative — only targets patterns where JSON parsing fails\n * due to quote collision within string values.\n */\nfunction fixInnerQuotes(json: string): string {\n\t// This is inherently heuristic. We target the common case:\n\t// \"key\": \"value with \"unescaped\" quotes inside\"\n\t//\n\t// Strategy: find string values that contain un-escaped quotes.\n\t// We'll use the fact that a well-formed JSON string:\n\t// \" starts a string, \" ends it. Within a string, \" is escaped as \\\".\n\t//\n\t// When we see a \" that's inside what should be a string value\n\t// (preceded by :\" or ,\" or {\" or [\"), and the next unescaped \" would\n\t// make the JSON invalid, we escape it.\n\n\tconst result: string[] = [];\n\tlet inString = false;\n\tlet inKey = false; // we're inside a key (before the colon)\n\tlet escape = false;\n\tlet colonSeen = false; // just saw colon, next value is a string likely\n\n\tfor (let i = 0; i < json.length; i++) {\n\t\tconst ch = json[i];\n\n\t\tif (escape) {\n\t\t\tescape = false;\n\t\t\tresult.push(ch);\n\t\t\tcontinue;\n\t\t}\n\n\t\tif (ch === \"\\\\\") {\n\t\t\tescape = true;\n\t\t\tresult.push(ch);\n\t\t\tcontinue;\n\t\t}\n\n\t\tif (ch === '\"') {\n\t\t\tif (!inString) {\n\t\t\t\tinString = true;\n\t\t\t\tinKey = !colonSeen; // after colon, we're in a value\n\t\t\t\tcolonSeen = false;\n\t\t\t\tresult.push(ch);\n\t\t\t} else {\n\t\t\t\t// Closing a string — peek ahead to see if this is really the end\n\t\t\t\t// or if there's more string content that should have been escaped.\n\t\t\t\tconst peek = json.slice(i + 1);\n\t\t\t\tconst restTrimmed = peek.trimStart();\n\n\t\t\t\t// If the next non-whitespace is NOT one of: , } ] : (or end of input),\n\t\t\t\t// this is likely an inner quote that should have been escaped.\n\t\t\t\t// But only fix if we're in a string value (not a key).\n\t\t\t\tif (\n\t\t\t\t\t!inKey &&\n\t\t\t\t\trestTrimmed.length > 0 &&\n\t\t\t\t\t!restTrimmed.startsWith(\",\") &&\n\t\t\t\t\t!restTrimmed.startsWith(\"}\") &&\n\t\t\t\t\t!restTrimmed.startsWith(\"]\") &&\n\t\t\t\t\t!restTrimmed.startsWith(\":\") &&\n\t\t\t\t\t!restTrimmed.startsWith(\"\\n\") &&\n\t\t\t\t\t!restTrimmed.startsWith(\"\\r\")\n\t\t\t\t) {\n\t\t\t\t\t// This is likely an inner quote — escape it\n\t\t\t\t\tresult.push(\"\\\\\");\n\t\t\t\t}\n\t\t\t\tinString = false;\n\t\t\t\tinKey = false;\n\t\t\t\tresult.push(ch);\n\t\t\t}\n\t\t\tcontinue;\n\t\t}\n\n\t\tif (!inString && ch === \":\") {\n\t\t\tcolonSeen = true;\n\t\t} else if (!inString && (ch === \",\" || ch === \"{\" || ch === \"[\")) {\n\t\t\tcolonSeen = false;\n\t\t} else if (!inString && /\\S/.test(ch)) {\n\t\t\tcolonSeen = false;\n\t\t}\n\n\t\tresult.push(ch);\n\t}\n\n\treturn result.join(\"\");\n}\n\n/**\n * Balance unclosed braces and brackets using a stack.\n * Adds missing closing chars at the end.\n */\nfunction balanceBraces(json: string): string {\n\tconst stack: string[] = [];\n\tlet inString = false;\n\tlet escape = false;\n\n\tfor (const ch of json) {\n\t\tif (escape) {\n\t\t\tescape = false;\n\t\t\tcontinue;\n\t\t}\n\t\tif (ch === \"\\\\\" && inString) {\n\t\t\tescape = true;\n\t\t\tcontinue;\n\t\t}\n\t\tif (ch === '\"') {\n\t\t\tinString = !inString;\n\t\t\tcontinue;\n\t\t}\n\t\tif (inString) continue;\n\n\t\tif (ch === \"{\") stack.push(\"}\");\n\t\telse if (ch === \"[\") stack.push(\"]\");\n\t\telse if (ch === \"}\" || ch === \"]\") {\n\t\t\tif (stack.length > 0 && stack[stack.length - 1] === ch) {\n\t\t\t\tstack.pop();\n\t\t\t}\n\t\t\t// else: extra closer — we don't try to fix that\n\t\t}\n\t}\n\n\t// Only repair if the imbalance is reasonable\n\tif (stack.length === 0) return json;\n\tif (stack.length > 20) return json; // too broken\n\treturn json + stack.reverse().join(\"\");\n}\n\n/**\n * Attempt to recover a truncated JSON object.\n * Tries progressively removing the last character, then last comma-prefixed\n * segment, then adds a synthetic closing brace.\n */\nfunction fixTruncatedObject(json: string): string | null {\n\t// Strategy: try to close the object by adding missing braces\n\tconst balanced = balanceBraces(json);\n\tif (balanced !== json) return balanced;\n\n\t// If balancing didn't help and the string ends mid-value,\n\t// try truncating to the last complete key-value pair\n\tif (json.endsWith(\",\")) {\n\t\t// Already clean — just needs closing\n\t\tconst cleaned = `${json}}`;\n\t\tif (isValidJson(cleaned)) return cleaned;\n\n\t\t// Try removing trailing comma and closing\n\t\tconst alt = `${json.slice(0, -1)}}`;\n\t\tif (isValidJson(alt)) return alt;\n\t}\n\n\t// Try appending synthetic closing\n\tconst synthetic = json.replace(/[^}\\]]$/, \"\").trimEnd();\n\tconst braces = countUnclosed(synthetic);\n\tif (braces > 0 && braces <= 5) {\n\t\tconst closed = synthetic + \"}\".repeat(braces);\n\t\tif (isValidJson(closed)) return closed;\n\t}\n\n\treturn null;\n}\n\nfunction countUnclosed(json: string): number {\n\tlet depth = 0;\n\tlet inString = false;\n\tlet escape = false;\n\n\tfor (const ch of json) {\n\t\tif (escape) {\n\t\t\tescape = false;\n\t\t\tcontinue;\n\t\t}\n\t\tif (ch === \"\\\\\" && inString) {\n\t\t\tescape = true;\n\t\t\tcontinue;\n\t\t}\n\t\tif (ch === '\"') {\n\t\t\tinString = !inString;\n\t\t\tcontinue;\n\t\t}\n\t\tif (inString) continue;\n\t\tif (ch === \"{\") depth++;\n\t\telse if (ch === \"}\") depth--;\n\t}\n\treturn Math.max(0, depth);\n}\n\nfunction isValidJson(text: string): boolean {\n\ttry {\n\t\tJSON.parse(text);\n\t\treturn true;\n\t} catch {\n\t\treturn false;\n\t}\n}\n"]}

package/dist/hamr/providers/repair/reasoning-sanitizer.d.ts

@@ -0,0 +1,28 @@
+/**
+ * Reasoning sanitizer — strips model thinking/reasoning text from output.
+ *
+ * Local models frequently leak reasoning text into the visible output stream.
+ * This module provides sanitization for:
+ * - `<think>...</think>` blocks (Qwen, DeepSeek)
+ * - `<thinking>...</thinking>` blocks (Anthropic extended-thinking, various)
+ * - ` response/` fenced code blocks (DeepSeek)
+ * - DeepSeek reasoning_content leakage into content field
+ *
+ * Returns sanitized output with a flag indicating whether any reasoning was removed.
+ */
+export interface SanitizeResult {
+    content: string;
+    removedReasoning: boolean;
+}
+/**
+ * Sanitize model output by removing reasoning/thinking text.
+ *
+ * Handles three patterns:
+ * 1. `<think>...</think>` — Qwen-style reasoning blocks (with possible attributes)
+ * 2. `<thinking>...</thinking>` — More explicit reasoning markers
+ * 3. ```response / ```text / ``` reasoning blocks — DeepSeek-style
+ *
+ * Returns sanitized content and a flag indicating whether anything was removed.
+ */
+export declare function sanitizeReasoning(content: string): SanitizeResult;
+//# sourceMappingURL=reasoning-sanitizer.d.ts.map

package/dist/hamr/providers/repair/reasoning-sanitizer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"reasoning-sanitizer.d.ts","sourceRoot":"","sources":["../../../../src/hamr/providers/repair/reasoning-sanitizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,MAAM,WAAW,cAAc;IAC9B,OAAO,EAAE,MAAM,CAAC;IAChB,gBAAgB,EAAE,OAAO,CAAC;CAC1B;AAED;;;;;;;;;GASG;AACH,wBAAgB,iBAAiB,CAAC,OAAO,EAAE,MAAM,GAAG,cAAc,CAoGjE"}

package/dist/hamr/providers/repair/reasoning-sanitizer.js

@@ -0,0 +1,115 @@
+/**
+ * Reasoning sanitizer — strips model thinking/reasoning text from output.
+ *
+ * Local models frequently leak reasoning text into the visible output stream.
+ * This module provides sanitization for:
+ * - `<think>...</think>` blocks (Qwen, DeepSeek)
+ * - `<thinking>...</thinking>` blocks (Anthropic extended-thinking, various)
+ * - ` response/` fenced code blocks (DeepSeek)
+ * - DeepSeek reasoning_content leakage into content field
+ *
+ * Returns sanitized output with a flag indicating whether any reasoning was removed.
+ */
+/**
+ * Sanitize model output by removing reasoning/thinking text.
+ *
+ * Handles three patterns:
+ * 1. `<think>...</think>` — Qwen-style reasoning blocks (with possible attributes)
+ * 2. `<thinking>...</thinking>` — More explicit reasoning markers
+ * 3. ```response / ```text / ``` reasoning blocks — DeepSeek-style
+ *
+ * Returns sanitized content and a flag indicating whether anything was removed.
+ */
+export function sanitizeReasoning(content) {
+    let sanitized = content;
+    let removedReasoning = false;
+    // Pattern 1: <think>...</think> blocks (handle with attributes, multiline, case-insensitive)
+    // Replace with space to prevent word-joining across the removed block.
+    const thinkRegex = /<think\b[^>]*>[\s\S]*?<\/think>/gi;
+    if (thinkRegex.test(sanitized)) {
+        sanitized = sanitized.replace(thinkRegex, " ");
+        removedReasoning = true;
+    }
+    // Pattern 2: <thinking>...</thinking> blocks
+    const thinkingRegex = /<thinking\b[^>]*>[\s\S]*?<\/thinking>/gi;
+    if (thinkingRegex.test(sanitized)) {
+        sanitized = sanitized.replace(thinkingRegex, " ");
+        removedReasoning = true;
+    }
+    // Pattern 3: DeepSeek-style ```response / ```text / ```thinking / ```assistant blocks
+    // These are fenced code blocks that contain reasoning or redundant output
+    const fencedResponseRegex = /```\s*(?:response|assistant_text|thinking|reasoning)[\s\S]*?```/gi;
+    if (fencedResponseRegex.test(sanitized)) {
+        sanitized = sanitized.replace(fencedResponseRegex, " ");
+        removedReasoning = true;
+    }
+    // Pattern 3b: Bare DeepSeek / Qwen boundary markers that leak through
+    // when the inference engine doesn't properly separate reasoning from output.
+    // These appear as literal text like "<｜end▁of▁thinking｜>" or " answer" before
+    // the actual tool-call content. Strip the marker and everything before it
+    const leakedResponseRegex = /^[\s\S]*?<\/(?:think|thinking)>/i;
+    if (leakedResponseRegex.test(sanitized) && /<tool_call>/i.test(sanitized)) {
+        // Strip everything up to and including the closing think tag,
+        // keeping only the content after it (which should include tool calls).
+        // Only do this when tool calls ARE present — don't destroy legitimate
+        // prose responses that happen to contain  from a prior turn.
+        const stripped = sanitized.replace(leakedResponseRegex, "").trim();
+        if (stripped.length > 0 && /<tool_call>/i.test(stripped)) {
+            sanitized = stripped;
+            removedReasoning = true;
+        }
+    }
+    // Pattern 3c: Some providers emit raw " response" / " answer" / " final"
+    // as text markers after  tags (common with DeepSeek + vLLM combos).
+    const rawResponseMarkers = /(?:^|\n)\s*(?:response|answer|final|output)\s*(?:\n|$)/gi;
+    if (rawResponseMarkers.test(sanitized) && /<tool_call>/i.test(sanitized)) {
+        const stripped = sanitized.replace(rawResponseMarkers, "\n").trim();
+        if (stripped.length > 0) {
+            sanitized = stripped;
+            removedReasoning = true;
+        }
+    }
+    // Pattern 4: Self-closing <think/> or <thinking/> tags
+    const selfClosingRegex = /<think(?:ing)?\s*\/>/gi;
+    if (selfClosingRegex.test(sanitized)) {
+        sanitized = sanitized.replace(selfClosingRegex, " ");
+        removedReasoning = true;
+    }
+    // Pattern 5: Opening <think> without closing tag (truncated reasoning)
+    // If we see <think or <thinking without a matching close, trim from there
+    const openThinkIdx = sanitized.search(/<think(?:ing)?\b/i);
+    if (openThinkIdx !== -1) {
+        // Check if there's a matching close tag after it
+        const afterOpen = sanitized.slice(openThinkIdx);
+        const hasClose = /<\/think(?:ing)?>/i.test(afterOpen);
+        if (!hasClose) {
+            // Truncated thinking tag — remove from here to end
+            sanitized = sanitized.slice(0, openThinkIdx);
+            removedReasoning = true;
+        }
+        else {
+            // Has close tag, but the regex above should have caught it.
+            // Try a more aggressive cleanup
+            const thinkPattern = /<think(?:ing)?\b[^>]*>[\s\S]*?<\/think(?:ing)?>/gi;
+            if (thinkPattern.test(sanitized)) {
+                sanitized = sanitized.replace(thinkPattern, " ");
+                removedReasoning = true;
+            }
+        }
+    }
+    // Pattern 6: Stray closing tags from truncated/streamed reasoning blocks.
+    // Some local models emit only the visible answer plus a dangling </think>.
+    const strayClosingRegex = /<\/think(?:ing)?>/gi;
+    if (strayClosingRegex.test(sanitized)) {
+        sanitized = sanitized.replace(strayClosingRegex, " ");
+        removedReasoning = true;
+    }
+    // Pattern 7: Collapse horizontal whitespace introduced by removals,
+    // normalize excessive newlines, and trim.
+    sanitized = sanitized
+        .replace(/[ \t]+/g, " ")
+        .replace(/\n{3,}/g, "\n\n")
+        .trim();
+    return { content: sanitized, removedReasoning };
+}
+//# sourceMappingURL=reasoning-sanitizer.js.map

package/dist/hamr/providers/repair/reasoning-sanitizer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"reasoning-sanitizer.js","sourceRoot":"","sources":["../../../../src/hamr/providers/repair/reasoning-sanitizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAOH;;;;;;;;;GASG;AACH,MAAM,UAAU,iBAAiB,CAAC,OAAe;IAChD,IAAI,SAAS,GAAG,OAAO,CAAC;IACxB,IAAI,gBAAgB,GAAG,KAAK,CAAC;IAE7B,6FAA6F;IAC7F,uEAAuE;IACvE,MAAM,UAAU,GAAG,mCAAmC,CAAC;IACvD,IAAI,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QAChC,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC;QAC/C,gBAAgB,GAAG,IAAI,CAAC;IACzB,CAAC;IAED,6CAA6C;IAC7C,MAAM,aAAa,GAAG,yCAAyC,CAAC;IAChE,IAAI,aAAa,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QACnC,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC;QAClD,gBAAgB,GAAG,IAAI,CAAC;IACzB,CAAC;IAED,sFAAsF;IACtF,0EAA0E;IAC1E,MAAM,mBAAmB,GAAG,mEAAmE,CAAC;IAChG,IAAI,mBAAmB,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QACzC,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,mBAAmB,EAAE,GAAG,CAAC,CAAC;QACxD,gBAAgB,GAAG,IAAI,CAAC;IACzB,CAAC;IAED,sEAAsE;IACtE,6EAA6E;IAC7E,8EAA8E;IAC9E,0EAA0E;IAC1E,MAAM,mBAAmB,GAAG,kCAAkC,CAAC;IAC/D,IAAI,mBAAmB,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QAC3E,8DAA8D;QAC9D,uEAAuE;QACvE,sEAAsE;QACtE,6DAA6D;QAC7D,MAAM,QAAQ,GAAG,SAAS,CAAC,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QACnE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;YAC1D,SAAS,GAAG,QAAQ,CAAC;YACrB,gBAAgB,GAAG,IAAI,CAAC;QACzB,CAAC;IACF,CAAC;IAED,yEAAyE;IACzE,oEAAoE;IACpE,MAAM,kBAAkB,GAAG,0DAA0D,CAAC;IACtF,IAAI,kBAAkB,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QAC1E,MAAM,QAAQ,GAAG,SAAS,CAAC,OAAO,CAAC,kBAAkB,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;QACpE,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACzB,SAAS,GAAG,QAAQ,CAAC;YACrB,gBAAgB,GAAG,IAAI,CAAC;QACzB,CAAC;IACF,CAAC;IAED,uDAAuD;IACvD,MAAM,gBAAgB,GAAG,wBAAwB,CAAC;IAClD,IAAI,gBAAgB,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QACtC,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,gBAAgB,EAAE,GAAG,CAAC,CAAC;QACrD,gBAAgB,GAAG,IAAI,CAAC;IACzB,CAAC;IAED,uEAAuE;IACvE,0EAA0E;IAC1E,MAAM,YAAY,GAAG,SAAS,CAAC,MAAM,CAAC,mBAAmB,CAAC,CAAC;IAC3D,IAAI,YAAY,KAAK,CAAC,CAAC,EAAE,CAAC;QACzB,iDAAiD;QACjD,MAAM,SAAS,GAAG,SAAS,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC;QAChD,MAAM,QAAQ,GAAG,oBAAoB,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACtD,IAAI,CAAC,QAAQ,EAAE,CAAC;YACf,mDAAmD;YACnD,SAAS,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;YAC7C,gBAAgB,GAAG,IAAI,CAAC;QACzB,CAAC;aAAM,CAAC;YACP,4DAA4D;YAC5D,gCAAgC;YAChC,MAAM,YAAY,GAAG,mDAAmD,CAAC;YACzE,IAAI,YAAY,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;gBAClC,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,YAAY,EAAE,GAAG,CAAC,CAAC;gBACjD,gBAAgB,GAAG,IAAI,CAAC;YACzB,CAAC;QACF,CAAC;IACF,CAAC;IAED,0EAA0E;IAC1E,2EAA2E;IAC3E,MAAM,iBAAiB,GAAG,qBAAqB,CAAC;IAChD,IAAI,iBAAiB,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC;QACvC,SAAS,GAAG,SAAS,CAAC,OAAO,CAAC,iBAAiB,EAAE,GAAG,CAAC,CAAC;QACtD,gBAAgB,GAAG,IAAI,CAAC;IACzB,CAAC;IAED,oEAAoE;IACpE,0CAA0C;IAC1C,SAAS,GAAG,SAAS;SACnB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;SAC1B,IAAI,EAAE,CAAC;IAET,OAAO,EAAE,OAAO,EAAE,SAAS,EAAE,gBAAgB,EAAE,CAAC;AACjD,CAAC","sourcesContent":["/**\n * Reasoning sanitizer — strips model thinking/reasoning text from output.\n *\n * Local models frequently leak reasoning text into the visible output stream.\n * This module provides sanitization for:\n * - `<think>...</think>` blocks (Qwen, DeepSeek)\n * - `<thinking>...</thinking>` blocks (Anthropic extended-thinking, various)\n * - ` response/` fenced code blocks (DeepSeek)\n * - DeepSeek reasoning_content leakage into content field\n *\n * Returns sanitized output with a flag indicating whether any reasoning was removed.\n */\n\nexport interface SanitizeResult {\n\tcontent: string;\n\tremovedReasoning: boolean;\n}\n\n/**\n * Sanitize model output by removing reasoning/thinking text.\n *\n * Handles three patterns:\n * 1. `<think>...</think>` — Qwen-style reasoning blocks (with possible attributes)\n * 2. `<thinking>...</thinking>` — More explicit reasoning markers\n * 3. ```response / ```text / ``` reasoning blocks — DeepSeek-style\n *\n * Returns sanitized content and a flag indicating whether anything was removed.\n */\nexport function sanitizeReasoning(content: string): SanitizeResult {\n\tlet sanitized = content;\n\tlet removedReasoning = false;\n\n\t// Pattern 1: <think>...</think> blocks (handle with attributes, multiline, case-insensitive)\n\t// Replace with space to prevent word-joining across the removed block.\n\tconst thinkRegex = /<think\\b[^>]*>[\\s\\S]*?<\\/think>/gi;\n\tif (thinkRegex.test(sanitized)) {\n\t\tsanitized = sanitized.replace(thinkRegex, \" \");\n\t\tremovedReasoning = true;\n\t}\n\n\t// Pattern 2: <thinking>...</thinking> blocks\n\tconst thinkingRegex = /<thinking\\b[^>]*>[\\s\\S]*?<\\/thinking>/gi;\n\tif (thinkingRegex.test(sanitized)) {\n\t\tsanitized = sanitized.replace(thinkingRegex, \" \");\n\t\tremovedReasoning = true;\n\t}\n\n\t// Pattern 3: DeepSeek-style ```response / ```text / ```thinking / ```assistant blocks\n\t// These are fenced code blocks that contain reasoning or redundant output\n\tconst fencedResponseRegex = /```\\s*(?:response|assistant_text|thinking|reasoning)[\\s\\S]*?```/gi;\n\tif (fencedResponseRegex.test(sanitized)) {\n\t\tsanitized = sanitized.replace(fencedResponseRegex, \" \");\n\t\tremovedReasoning = true;\n\t}\n\n\t// Pattern 3b: Bare DeepSeek / Qwen boundary markers that leak through\n\t// when the inference engine doesn't properly separate reasoning from output.\n\t// These appear as literal text like \"<｜end▁of▁thinking｜>\" or \" answer\" before\n\t// the actual tool-call content. Strip the marker and everything before it\n\tconst leakedResponseRegex = /^[\\s\\S]*?<\\/(?:think|thinking)>/i;\n\tif (leakedResponseRegex.test(sanitized) && /<tool_call>/i.test(sanitized)) {\n\t\t// Strip everything up to and including the closing think tag,\n\t\t// keeping only the content after it (which should include tool calls).\n\t\t// Only do this when tool calls ARE present — don't destroy legitimate\n\t\t// prose responses that happen to contain  from a prior turn.\n\t\tconst stripped = sanitized.replace(leakedResponseRegex, \"\").trim();\n\t\tif (stripped.length > 0 && /<tool_call>/i.test(stripped)) {\n\t\t\tsanitized = stripped;\n\t\t\tremovedReasoning = true;\n\t\t}\n\t}\n\n\t// Pattern 3c: Some providers emit raw \" response\" / \" answer\" / \" final\"\n\t// as text markers after  tags (common with DeepSeek + vLLM combos).\n\tconst rawResponseMarkers = /(?:^|\\n)\\s*(?:response|answer|final|output)\\s*(?:\\n|$)/gi;\n\tif (rawResponseMarkers.test(sanitized) && /<tool_call>/i.test(sanitized)) {\n\t\tconst stripped = sanitized.replace(rawResponseMarkers, \"\\n\").trim();\n\t\tif (stripped.length > 0) {\n\t\t\tsanitized = stripped;\n\t\t\tremovedReasoning = true;\n\t\t}\n\t}\n\n\t// Pattern 4: Self-closing <think/> or <thinking/> tags\n\tconst selfClosingRegex = /<think(?:ing)?\\s*\\/>/gi;\n\tif (selfClosingRegex.test(sanitized)) {\n\t\tsanitized = sanitized.replace(selfClosingRegex, \" \");\n\t\tremovedReasoning = true;\n\t}\n\n\t// Pattern 5: Opening <think> without closing tag (truncated reasoning)\n\t// If we see <think or <thinking without a matching close, trim from there\n\tconst openThinkIdx = sanitized.search(/<think(?:ing)?\\b/i);\n\tif (openThinkIdx !== -1) {\n\t\t// Check if there's a matching close tag after it\n\t\tconst afterOpen = sanitized.slice(openThinkIdx);\n\t\tconst hasClose = /<\\/think(?:ing)?>/i.test(afterOpen);\n\t\tif (!hasClose) {\n\t\t\t// Truncated thinking tag — remove from here to end\n\t\t\tsanitized = sanitized.slice(0, openThinkIdx);\n\t\t\tremovedReasoning = true;\n\t\t} else {\n\t\t\t// Has close tag, but the regex above should have caught it.\n\t\t\t// Try a more aggressive cleanup\n\t\t\tconst thinkPattern = /<think(?:ing)?\\b[^>]*>[\\s\\S]*?<\\/think(?:ing)?>/gi;\n\t\t\tif (thinkPattern.test(sanitized)) {\n\t\t\t\tsanitized = sanitized.replace(thinkPattern, \" \");\n\t\t\t\tremovedReasoning = true;\n\t\t\t}\n\t\t}\n\t}\n\n\t// Pattern 6: Stray closing tags from truncated/streamed reasoning blocks.\n\t// Some local models emit only the visible answer plus a dangling </think>.\n\tconst strayClosingRegex = /<\\/think(?:ing)?>/gi;\n\tif (strayClosingRegex.test(sanitized)) {\n\t\tsanitized = sanitized.replace(strayClosingRegex, \" \");\n\t\tremovedReasoning = true;\n\t}\n\n\t// Pattern 7: Collapse horizontal whitespace introduced by removals,\n\t// normalize excessive newlines, and trim.\n\tsanitized = sanitized\n\t\t.replace(/[ \\t]+/g, \" \")\n\t\t.replace(/\\n{3,}/g, \"\\n\\n\")\n\t\t.trim();\n\n\treturn { content: sanitized, removedReasoning };\n}\n"]}

package/dist/hamr/providers/repair/xml-repair.d.ts

@@ -0,0 +1,31 @@
+/**
+ * XML repair — bounded auto-recovery for Qwen-style XML tool calls.
+ *
+ * Qwen-family models emit `<tool_call>` blocks in XML format. Local
+ * models frequently produce:
+ * - Unclosed `<tool_call>` tags
+ * - Leaked `<thinking>` / `<think>` tags inside tool calls
+ * - Mixed XML + text content where thinking bleeds into tool blocks
+ * - Nested `<tool_call>` with missing closing function tags
+ *
+ * Each repair is recorded in `fixes[]` for debugging. Returns `null` when
+ * the input is unrepairable.
+ */
+export interface RepairResult {
+    repaired: string;
+    fixes: string[];
+}
+/**
+ * Attempt to repair malformed XML tool-call text from a local model.
+ *
+ * Repairs applied:
+ * 1. Strip leaked reasoning tags inside tool-call blocks
+ * 2. Close unclosed `<tool_call>` tags
+ * 3. Close unclosed `<function=...>` tags
+ * 4. Close unclosed `<parameter=...>` tags
+ * 5. Strip stray text between tool-call segments
+ *
+ * Returns `null` if the string is unrepairable (empty or garbage).
+ */
+export declare function repairXml(raw: string): RepairResult | null;
+//# sourceMappingURL=xml-repair.d.ts.map

package/dist/hamr/providers/repair/xml-repair.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"xml-repair.d.ts","sourceRoot":"","sources":["../../../../src/hamr/providers/repair/xml-repair.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,MAAM,WAAW,YAAY;IAC5B,QAAQ,EAAE,MAAM,CAAC;IACjB,KAAK,EAAE,MAAM,EAAE,CAAC;CAChB;AAED;;;;;;;;;;;GAWG;AACH,wBAAgB,SAAS,CAAC,GAAG,EAAE,MAAM,GAAG,YAAY,GAAG,IAAI,CAsD1D"}