memorymaster 3.4.0__tar.gz → 3.4.1__tar.gz

{memorymaster-3.4.0/memorymaster.egg-info → memorymaster-3.4.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: memorymaster
-Version: 3.4.0
+Version: 3.4.1
 Summary: Production-grade memory reliability system for AI coding agents. Lifecycle-managed claims with citations, conflict detection, steward governance, and MCP integration.
 Author: wolverin0
 License: MIT

{memorymaster-3.4.0 → memorymaster-3.4.1}/memorymaster/__init__.py

@@ -2,4 +2,4 @@
 
 __all__ = ["__version__"]
 
-__version__ = "3.4.0"
+__version__ = "3.4.1"

{memorymaster-3.4.0 → memorymaster-3.4.1}/memorymaster/config_templates/hooks/memorymaster-auto-ingest.py

@@ -41,9 +41,12 @@ def get_recent_assistant_messages(transcript_path, max_chars=3000):
                 entry = json.loads(line)
             except json.JSONDecodeError:
                 continue
-            if entry.get("role") != "assistant":
+            # Claude Code transcripts wrap role+content inside `message`.
+            # Fall back to top-level for older/alternate schemas.
+            msg = entry.get("message") if isinstance(entry.get("message"), dict) else entry
+            if msg.get("role") != "assistant":
                 continue
-            content = entry.get("content", "")
+            content = msg.get("content", "")
             text = ""
             if isinstance(content, str):
                 text = content

{memorymaster-3.4.0 → memorymaster-3.4.1/memorymaster.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: memorymaster
-Version: 3.4.0
+Version: 3.4.1
 Summary: Production-grade memory reliability system for AI coding agents. Lifecycle-managed claims with citations, conflict detection, steward governance, and MCP integration.
 Author: wolverin0
 License: MIT

{memorymaster-3.4.0 → memorymaster-3.4.1}/memorymaster.egg-info/SOURCES.txt

@@ -132,6 +132,7 @@ scripts/webhook_to_turns.py
 tests/conftest.py
 tests/test_access_control.py
 tests/test_auto_extractor.py
+tests/test_auto_ingest_hook_schema.py
 tests/test_auto_resolver.py
 tests/test_auto_validate.py
 tests/test_claim_links.py

{memorymaster-3.4.0 → memorymaster-3.4.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "memorymaster"
-version = "3.4.0"
+version = "3.4.1"
 description = "Production-grade memory reliability system for AI coding agents. Lifecycle-managed claims with citations, conflict detection, steward governance, and MCP integration."
 license = {text = "MIT"}
 authors = [{name = "wolverin0"}]

memorymaster-3.4.1/tests/test_auto_ingest_hook_schema.py

@@ -0,0 +1,145 @@
+"""Regression test for auto-ingest hook transcript schema adapter.
+
+Guards against the bug where the hook read `entry.get("role")` at the
+top level, but Claude Code transcripts wrap the role inside
+`entry["message"]["role"]`. The fix must accept BOTH shapes.
+
+Tests the extraction pattern used by:
+  - ~/.claude/hooks/memorymaster-auto-ingest.py (deployed)
+  - memorymaster/config_templates/hooks/memorymaster-auto-ingest.py (ships to new installs)
+  - scripts/llm_benchmark.py (benchmark harness)
+"""
+
+from __future__ import annotations
+
+import json
+from pathlib import Path
+
+
+def _extract(transcript_path: Path, max_chars: int = 3000) -> str:
+    """Mirror of the hook's extraction logic — if this test passes, the hook pattern is correct."""
+    messages: list[str] = []
+    lines = transcript_path.read_text(encoding="utf-8", errors="replace").splitlines()
+    for line in reversed(lines[-200:]):
+        try:
+            entry = json.loads(line)
+        except json.JSONDecodeError:
+            continue
+        msg = entry.get("message") if isinstance(entry.get("message"), dict) else entry
+        if msg.get("role") != "assistant":
+            continue
+        content = msg.get("content", "")
+        text = ""
+        if isinstance(content, str):
+            text = content
+        elif isinstance(content, list):
+            text = " ".join(
+                c.get("text", "")
+                for c in content
+                if isinstance(c, dict) and c.get("type") == "text"
+            )
+        text = text.strip()
+        if text and len(text) > 30:
+            messages.append(text[:500])
+            if sum(len(m) for m in messages) > max_chars:
+                break
+    return "\n---\n".join(reversed(messages))
+
+
+def _write_jsonl(path: Path, entries: list[dict]) -> None:
+    path.write_text(
+        "\n".join(json.dumps(e) for e in entries) + "\n",
+        encoding="utf-8",
+    )
+
+
+def test_wrapped_schema_is_parsed(tmp_path: Path) -> None:
+    """Claude Code current schema: role+content wrapped in entry['message']."""
+    p = tmp_path / "t.jsonl"
+    _write_jsonl(p, [
+        {"type": "assistant", "message": {"role": "assistant", "content": "This is a substantive assistant message with more than 30 characters."}},
+        {"type": "user", "message": {"role": "user", "content": "a user msg"}},
+        {"type": "assistant", "message": {"role": "assistant", "content": [{"type": "text", "text": "Another assistant reply that is long enough to include."}]}},
+    ])
+    out = _extract(p)
+    assert "substantive assistant message" in out
+    assert "Another assistant reply" in out
+    assert "a user msg" not in out
+
+
+def test_flat_schema_is_parsed(tmp_path: Path) -> None:
+    """Legacy/alternate schema: role at top level."""
+    p = tmp_path / "t.jsonl"
+    _write_jsonl(p, [
+        {"role": "assistant", "content": "Flat-schema assistant message with enough characters to pass the 30-char filter."},
+        {"role": "user", "content": "flat user msg"},
+    ])
+    out = _extract(p)
+    assert "Flat-schema assistant message" in out
+    assert "flat user msg" not in out
+
+
+def test_mixed_schema_is_parsed(tmp_path: Path) -> None:
+    """Both shapes in the same file — extractor must handle both without losing entries."""
+    p = tmp_path / "t.jsonl"
+    _write_jsonl(p, [
+        {"role": "assistant", "content": "First assistant reply in flat schema format with enough length."},
+        {"message": {"role": "assistant", "content": "Second reply in wrapped schema format with enough length."}},
+    ])
+    out = _extract(p)
+    assert "First assistant reply" in out
+    assert "Second reply" in out
+
+
+def test_short_messages_are_filtered(tmp_path: Path) -> None:
+    """The hook ignores assistant messages shorter than 30 chars (the 'ok' filter)."""
+    p = tmp_path / "t.jsonl"
+    _write_jsonl(p, [
+        {"message": {"role": "assistant", "content": "short"}},
+        {"message": {"role": "assistant", "content": "ok"}},
+    ])
+    out = _extract(p)
+    assert out == ""
+
+
+def test_empty_transcript(tmp_path: Path) -> None:
+    p = tmp_path / "t.jsonl"
+    p.write_text("", encoding="utf-8")
+    assert _extract(p) == ""
+
+
+def test_malformed_lines_are_skipped(tmp_path: Path) -> None:
+    p = tmp_path / "t.jsonl"
+    p.write_text(
+        "not json at all\n"
+        + json.dumps({"message": {"role": "assistant", "content": "Valid assistant message with enough length to pass."}})
+        + "\n{malformed\n",
+        encoding="utf-8",
+    )
+    out = _extract(p)
+    assert "Valid assistant message" in out
+
+
+def test_deployed_hook_has_adapter_pattern() -> None:
+    """Static check: the deployed hook must use the wrapped-schema adapter."""
+    import os
+
+    p = Path(os.path.expanduser("~/.claude/hooks/memorymaster-auto-ingest.py"))
+    if not p.exists():
+        return  # not installed in this environment — skip silently
+    src = p.read_text(encoding="utf-8", errors="replace")
+    assert 'entry.get("message")' in src, (
+        "auto-ingest hook missing wrapped-schema adapter — "
+        "must normalize entry['message'] before checking role"
+    )
+
+
+def test_template_hook_has_adapter_pattern() -> None:
+    """Static check: the shipped template must use the wrapped-schema adapter."""
+    p = Path(__file__).parent.parent / "memorymaster" / "config_templates" / "hooks" / "memorymaster-auto-ingest.py"
+    assert p.exists(), f"template not found at {p}"
+    src = p.read_text(encoding="utf-8", errors="replace")
+    assert 'entry.get("message")' in src, (
+        "template hook missing wrapped-schema adapter — "
+        "new installs will silently extract zero learnings"
+    )