mango-brain 3.3.3__tar.gz → 3.3.4__tar.gz

{mango_brain-3.3.3 → mango_brain-3.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mango-brain
-Version: 3.3.3
+Version: 3.3.4
 Summary: The learning layer for Claude Code — persistent associative memory + development workflow. Claude Code gets smarter the longer you use it.
 Author-email: Federico Anastasi <federico.anastasi@outlook.com>
 License: MIT
@@ -44,7 +44,7 @@ Dynamic: license-file
 
 <p align="center">
   <a href="https://pypi.org/project/mango-brain/"><img src="https://img.shields.io/pypi/v/mangobrain?style=flat-square&color=7c3aed&label=PyPI" /></a>
-  <img src="https://img.shields.io/badge/v3.3.2-7c3aed?style=flat-square" />
+  <img src="https://img.shields.io/badge/v3.3.4-7c3aed?style=flat-square" />
   <img src="https://img.shields.io/badge/python-3.11+-blue?style=flat-square&logo=python&logoColor=white" />
   <img src="https://img.shields.io/badge/MCP-compatible-8A2BE2?style=flat-square" />
   <img src="https://img.shields.io/badge/license-MIT-green?style=flat-square" />

{mango_brain-3.3.3 → mango_brain-3.3.4}/README.md

@@ -15,7 +15,7 @@
 
 <p align="center">
   <a href="https://pypi.org/project/mango-brain/"><img src="https://img.shields.io/pypi/v/mangobrain?style=flat-square&color=7c3aed&label=PyPI" /></a>
-  <img src="https://img.shields.io/badge/v3.3.2-7c3aed?style=flat-square" />
+  <img src="https://img.shields.io/badge/v3.3.4-7c3aed?style=flat-square" />
   <img src="https://img.shields.io/badge/python-3.11+-blue?style=flat-square&logo=python&logoColor=white" />
   <img src="https://img.shields.io/badge/MCP-compatible-8A2BE2?style=flat-square" />
   <img src="https://img.shields.io/badge/license-MIT-green?style=flat-square" />

{mango_brain-3.3.3 → mango_brain-3.3.4}/mango_brain.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mango-brain
-Version: 3.3.3
+Version: 3.3.4
 Summary: The learning layer for Claude Code — persistent associative memory + development workflow. Claude Code gets smarter the longer you use it.
 Author-email: Federico Anastasi <federico.anastasi@outlook.com>
 License: MIT
@@ -44,7 +44,7 @@ Dynamic: license-file
 
 <p align="center">
   <a href="https://pypi.org/project/mango-brain/"><img src="https://img.shields.io/pypi/v/mangobrain?style=flat-square&color=7c3aed&label=PyPI" /></a>
-  <img src="https://img.shields.io/badge/v3.3.2-7c3aed?style=flat-square" />
+  <img src="https://img.shields.io/badge/v3.3.4-7c3aed?style=flat-square" />
   <img src="https://img.shields.io/badge/python-3.11+-blue?style=flat-square&logo=python&logoColor=white" />
   <img src="https://img.shields.io/badge/MCP-compatible-8A2BE2?style=flat-square" />
   <img src="https://img.shields.io/badge/license-MIT-green?style=flat-square" />

{mango_brain-3.3.3 → mango_brain-3.3.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mango-brain"
-version = "3.3.3"
+version = "3.3.4"
 description = "The learning layer for Claude Code — persistent associative memory + development workflow. Claude Code gets smarter the longer you use it."
 readme = "README.md"
 license = {text = "MIT"}

{mango_brain-3.3.3 → mango_brain-3.3.4}/server/config.py

@@ -164,6 +164,7 @@ QUICK_BUDGET = int(_get("retrieval", "quick_budget", "2000"))
 SESSION_QUICK_BUDGET = int(_get("retrieval", "session_quick_budget", "4000"))
 DEEP_MAX_RESULTS = int(_get("retrieval", "deep_max_results", "20"))
 QUICK_MAX_RESULTS = int(_get("retrieval", "quick_max_results", "6"))
+RECENT_MAX_RESULTS = int(_get("retrieval", "recent_max_results", "20"))
 RELEVANCE_THRESHOLD_RATIO = float(_get("retrieval", "deep_threshold", "0.78"))
 QUICK_RELEVANCE_THRESHOLD_RATIO = float(_get("retrieval", "quick_threshold", "0.85"))
 

{mango_brain-3.3.3 → mango_brain-3.3.4}/server/mcp_tools.py

@@ -103,26 +103,9 @@ def register_tools(
                 recent_count = sum(1 for m in memories if m.id in recent_ids)
                 neighbor_count = len(memories) - recent_count
 
-                result = {
-                    "memories": [
-                        {
-                            "id": m.id,
-                            "content": m.content,
-                            "type": m.type.value if hasattr(m.type, 'value') else m.type,
-                            "project": m.project,
-                            "tags": m.tags,
-                            "relevance": round(scores[i], 4),
-                            "file_path": m.file_path,
-                            "code_signature": m.code_signature,
-                            "is_recent": m.id in recent_ids,
-                        }
-                        for i, m in enumerate(memories)
-                    ],
-                    "total_tokens": total_tokens,
-                    "count": len(memories),
-                    "recent_count": recent_count,
-                    "neighbor_count": neighbor_count,
-                }
+                count = len(memories)
+                recent_count_val = recent_count
+                neighbor_count_val = neighbor_count
             else:
                 if not query:
                     return json.dumps({"error": "query is required for deep/quick mode"})
@@ -130,29 +113,37 @@ def register_tools(
                     query=query, mode=mode, project=project,
                     budget=budget, session_id=session_id,
                 )
-                result = {
-                    "memories": [
-                        {
-                            "id": m.id,
-                            "content": m.content,
-                            "type": m.type.value if hasattr(m.type, 'value') else m.type,
-                            "project": m.project,
-                            "tags": m.tags,
-                            "relevance": round(scores[i], 4),
-                            "file_path": m.file_path,
-                            "code_signature": m.code_signature,
-                        }
-                        for i, m in enumerate(memories)
-                    ],
-                    "total_tokens": total_tokens,
-                    "count": len(memories),
-                }
+                count = len(memories)
+                recent_count_val = None
+                neighbor_count_val = None
+
+            # Format as readable text (not raw JSON)
             _dur = int((_time.monotonic_ns() - _t0) / 1_000_000)
             await _log_op(db, "remember", project=project,
                           params={"query": query, "mode": mode, "limit": limit},
-                          result={"count": result["count"], "total_tokens": result["total_tokens"]},
+                          result={"count": count, "total_tokens": total_tokens},
                           duration_ms=_dur)
-            return json.dumps(result, indent=2)
+
+            lines = []
+            header = f"{count} memories · {mode} mode · {_dur}ms"
+            if recent_count_val is not None:
+                header += f" · {recent_count_val} recent + {neighbor_count_val} neighbors"
+            lines.append(header)
+            lines.append("")
+
+            for i, m in enumerate(memories):
+                mtype = m.type.value if hasattr(m.type, 'value') else m.type
+                rel = round(scores[i], 4)
+                lines.append(f"[{mtype}] {m.content}")
+                meta_parts = []
+                if m.file_path:
+                    meta_parts.append(f"file: {m.file_path}")
+                meta_parts.append(f"relevance: {rel}")
+                meta_parts.append(f"id: {m.id}")
+                lines.append("  " + " · ".join(meta_parts))
+                lines.append("")
+
+            return "\n".join(lines)
         except ValueError as e:
             _dur = int((_time.monotonic_ns() - _t0) / 1_000_000)
             await _log_op(db, "remember", project=project,

{mango_brain-3.3.3 → mango_brain-3.3.4}/server/retrieval.py

@@ -185,10 +185,13 @@ class RetrievalEngine:
                 candidates.append((score / max(m.token_count, 1), score, m))
         candidates.sort(key=lambda x: x[0], reverse=True)
 
+        from server.config import RECENT_MAX_RESULTS
         selected: list[Memory] = []
         scores: list[float] = []
         total_tokens = 0
         for _, score, m in candidates:
+            if len(selected) >= RECENT_MAX_RESULTS:
+                break
             if total_tokens + m.token_count > budget:
                 continue
             selected.append(m)