clawed 2.3.3__tar.gz → 2.3.4__tar.gz

{clawed-2.3.3 → clawed-2.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: clawed
-Version: 2.3.3
+Version: 2.3.4
 Summary: Claw-ED — personal AI teaching agent. Learns your voice, works while you sleep.
 Project-URL: Homepage, https://github.com/SirhanMacx/Claw-ED
 Project-URL: Documentation, https://github.com/SirhanMacx/Claw-ED#readme

{clawed-2.3.3 → clawed-2.3.4}/clawed/__init__.py

@@ -17,7 +17,7 @@ if hasattr(sys.stderr, "reconfigure"):
     except Exception:
         pass
 
-__version__ = "2.3.3"
+__version__ = "2.3.4"
 __author__ = "Jon Maccarello & Claw-ED contributors"
 __description__ = "Personal AI teaching agent. Learns your voice, works while you sleep."
 

{clawed-2.3.3 → clawed-2.3.4}/clawed/agent_core/prompt.py

@@ -91,7 +91,11 @@ def build_system_prompt(
         "  'Building your lesson package now — plan, handout, and slides coming up.'\n"
         "The teacher should always know you're working, not stuck.\n"
         "1. Read SOUL.md to know your voice and values\n"
-        "2. Search the teacher's curriculum files for relevant prior work\n"
+        "2. **MANDATORY: Before calling generate_lesson_bundle, ALWAYS call "
+        "search_my_materials first** with the lesson topic. This is non-negotiable. "
+        "The teacher has uploaded materials — if you skip this step, you will "
+        "generate generic content instead of building on their prior work. "
+        "Tell the teacher what you found before generating.\n"
         "3. Generate complete packages (lesson plan + student handout + slideshow) "
         "using generate_lesson_bundle\n"
         "4. Never ask 'want me to create materials?' -- just create them\n"

{clawed-2.3.3 → clawed-2.3.4}/clawed/agent_core/tools/generate_lesson_bundle.py

@@ -22,7 +22,10 @@ class GenerateLessonBundleTool:
                 "description": (
                     "Generate a COMPLETE teaching package for a topic: "
                     "a lesson plan (DOCX), a student handout (DOCX), and "
-                    "a slideshow (PPTX). All three files are created at once."
+                    "a slideshow (PPTX). All three files are created at once. "
+                    "IMPORTANT: Always call search_my_materials FIRST to find "
+                    "the teacher's existing materials on this topic before "
+                    "calling this tool."
                 ),
                 "parameters": {
                     "type": "object",

{clawed-2.3.3 → clawed-2.3.4}/clawed/agent_core/tools/ingest_materials.py

@@ -65,7 +65,31 @@ class IngestMaterialsTool:
                 from clawed.persona import extract_persona, save_persona
 
                 persona = await extract_persona(docs, context.config)
+                # Override LLM-inferred name with configured teacher name
+                try:
+                    if context.config and context.config.teacher_profile and context.config.teacher_profile.name:
+                        persona.name = f"{context.config.teacher_profile.name} Teaching Persona"
+                except Exception:
+                    pass
+                try:
+                    _id_path = Path.home() / ".eduagent" / "workspace" / "identity.md"
+                    if _id_path.exists():
+                        import re as _re
+                        _id_content = _id_path.read_text(encoding="utf-8")
+                        _name_match = _re.match(r"^#\s+(.+)", _id_content)
+                        if _name_match:
+                            _tname = _name_match.group(1).strip()
+                            if _tname and _tname != "Teacher":
+                                persona.name = f"{_tname} Teaching Persona"
+                except Exception:
+                    pass
                 save_persona(persona, Path.home() / ".eduagent")
+                # Track persona changes for evolution
+                try:
+                    from clawed.persona_evolution import record_ingestion_changes
+                    record_ingestion_changes(old_persona=None, new_persona=persona)
+                except Exception:
+                    pass
             except Exception:
                 pass
 
@@ -129,6 +153,39 @@ class IngestMaterialsTool:
             except Exception as e:
                 logger.debug("KB indexing failed: %s", e)
 
+            # Register assets for file-level search
+            try:
+                from clawed.asset_registry import AssetRegistry
+                registry = AssetRegistry()
+                asset_count = 0
+                for doc in docs:
+                    doc_type_val = (
+                        doc.doc_type.value
+                        if hasattr(doc.doc_type, "value")
+                        else str(doc.doc_type)
+                    )
+                    extraction = None
+                    if doc.source_path:
+                        try:
+                            from clawed.ingestor import extract_rich
+                            extraction = extract_rich(Path(doc.source_path))
+                        except Exception:
+                            pass
+                    aid = registry.register_asset(
+                        teacher_id=context.teacher_id,
+                        source_path=doc.source_path or "",
+                        title=doc.title,
+                        doc_type=doc_type_val,
+                        text=doc.content,
+                        extraction=extraction,
+                    )
+                    if aid:
+                        asset_count += 1
+                if asset_count:
+                    summary += f" ({asset_count} files catalogued for search)"
+            except Exception as e:
+                logger.debug("Asset registration failed: %s", e)
+
             # Update SOUL.md with what we learned
             try:
                 soul_path = Path.home() / ".eduagent" / "workspace" / "SOUL.md"

{clawed-2.3.3 → clawed-2.3.4}/clawed/commands/generate.py

@@ -130,8 +130,36 @@ def ingest(
         persona = _run_async(extract_persona(documents))
         progress.update(task, description="Persona extracted!")
 
+    # Override LLM-inferred name with configured teacher name
+    try:
+        cfg = AppConfig.load()
+        if cfg.teacher_profile and cfg.teacher_profile.name:
+            persona.name = f"{cfg.teacher_profile.name} Teaching Persona"
+    except Exception:
+        pass
+    # Also check identity.md
+    try:
+        identity_path = Path.home() / ".eduagent" / "workspace" / "identity.md"
+        if identity_path.exists():
+            import re
+            content = identity_path.read_text(encoding="utf-8")
+            name_match = re.match(r"^#\s+(.+)", content)
+            if name_match:
+                teacher_name = name_match.group(1).strip()
+                if teacher_name and teacher_name != "Teacher":
+                    persona.name = f"{teacher_name} Teaching Persona"
+    except Exception:
+        pass
+
     out = save_persona(persona, _output_dir())
 
+    # Track persona changes for evolution
+    try:
+        from clawed.persona_evolution import record_ingestion_changes
+        record_ingestion_changes(old_persona=None, new_persona=persona)
+    except Exception:
+        pass
+
     # Index documents into curriculum knowledge base for KB search
     kb_msg = ""
     try:

{clawed-2.3.3 → clawed-2.3.4}/clawed/handlers/ingest.py

@@ -60,6 +60,25 @@ class IngestHandler:
             try:
                 from clawed.models import AppConfig
                 persona = await extract_persona(documents, AppConfig.load())
+                # Override LLM-inferred name with configured teacher name
+                try:
+                    config = AppConfig.load()
+                    if config.teacher_profile and config.teacher_profile.name:
+                        persona.name = f"{config.teacher_profile.name} Teaching Persona"
+                except Exception:
+                    pass
+                try:
+                    _id_path = Path.home() / ".eduagent" / "workspace" / "identity.md"
+                    if _id_path.exists():
+                        import re as _re
+                        _id_content = _id_path.read_text(encoding="utf-8")
+                        _name_match = _re.match(r"^#\s+(.+)", _id_content)
+                        if _name_match:
+                            _tname = _name_match.group(1).strip()
+                            if _tname and _tname != "Teacher":
+                                persona.name = f"{_tname} Teaching Persona"
+                except Exception:
+                    pass
                 from clawed.persona import save_persona
                 save_persona(persona, Path.home() / ".eduagent")
                 style_info = f"\nLearned teaching style: {persona.teaching_style}"
@@ -97,6 +116,35 @@ class IngestHandler:
             except Exception as e:
                 logger.debug("KB indexing skipped: %s", e)
 
+            # Register assets (files, images, YouTube links) for search
+            try:
+                from clawed.asset_registry import AssetRegistry
+                registry = AssetRegistry()
+                asset_count = 0
+                for doc in documents:
+                    doc_type_val = doc.doc_type.value if hasattr(doc.doc_type, "value") else str(doc.doc_type)
+                    extraction = None
+                    if doc.source_path:
+                        try:
+                            from clawed.ingestor import extract_rich
+                            extraction = extract_rich(Path(doc.source_path))
+                        except Exception:
+                            pass
+                    aid = registry.register_asset(
+                        teacher_id=teacher_id,
+                        source_path=doc.source_path or "",
+                        title=doc.title,
+                        doc_type=doc_type_val,
+                        text=doc.content,
+                        extraction=extraction,
+                    )
+                    if aid:
+                        asset_count += 1
+                if asset_count:
+                    kb_info += f" ({asset_count} files catalogued with images and links)"
+            except Exception as e:
+                logger.debug("Asset registration skipped: %s", e)
+
             return GatewayResponse(
                 text=f"Ingested {len(documents)} document(s).{style_info}{kb_info}"
             )

{clawed-2.3.3 → clawed-2.3.4}/clawed/models.py

@@ -488,6 +488,15 @@ class WorksheetItem(BaseModel):
     answer_key: str = ""
     point_value: int = 1
 
+    @field_validator("point_value", mode="before")
+    @classmethod
+    def _coerce_point_value(cls, v):
+        if isinstance(v, str):
+            import re
+            match = re.match(r"(\d+)", v.strip())
+            return int(match.group(1)) if match else 1
+        return v
+
 
 class AssessmentQuestion(BaseModel):
     """A single assessment question."""
@@ -499,6 +508,15 @@ class AssessmentQuestion(BaseModel):
     correct_answer: str = ""
     point_value: int = 1
 
+    @field_validator("point_value", mode="before")
+    @classmethod
+    def _coerce_point_value(cls, v):
+        if isinstance(v, str):
+            import re
+            match = re.match(r"(\d+)", v.strip())
+            return int(match.group(1)) if match else 1
+        return v
+
 
 class RubricCriterion(BaseModel):
     """A single rubric criterion with levels."""
@@ -674,6 +692,15 @@ class SummativeQuestion(BaseModel):
     point_value: int = 1
     standard_aligned: str = ""
 
+    @field_validator("point_value", mode="before")
+    @classmethod
+    def _coerce_point_value(cls, v):
+        if isinstance(v, str):
+            import re
+            match = re.match(r"(\d+)", v.strip())
+            return int(match.group(1)) if match else 1
+        return v
+
 
 class SummativeAssessment(BaseModel):
     """Unit test — mix of MC, short answer, DBQ/essay aligned to all unit objectives."""

{clawed-2.3.3 → clawed-2.3.4}/clawed/slide_images.py

@@ -253,16 +253,17 @@ def _select_sources(subject: str, topic: str = "") -> list[str]:
 
     Returns an ordered list of source identifiers to try.
     Teacher's own images are always checked first.
+    Only uses academic sources (LOC, Wikimedia). No stock photos.
     """
     subject_lower = subject.strip().lower()
     if any(s in subject_lower for s in ("history", "social", "civics", "government")):
-        return ["teacher_files", "loc", "wikimedia", "unsplash"]
+        return ["teacher_files", "loc", "wikimedia"]
     elif any(s in subject_lower for s in ("science", "biology", "chemistry", "physics")):
-        return ["teacher_files", "wikimedia", "loc", "unsplash"]
+        return ["teacher_files", "wikimedia", "loc"]
     elif any(s in subject_lower for s in ("art", "music")):
-        return ["teacher_files", "wikimedia", "loc", "unsplash"]
+        return ["teacher_files", "wikimedia", "loc"]
     else:
-        return ["teacher_files", "loc", "wikimedia", "unsplash"]
+        return ["teacher_files", "loc", "wikimedia"]
 
 
 # ── Cache helpers ────────────────────────────────────────────────────
@@ -497,14 +498,25 @@ async def _fetch_teacher_image(
 
         with sqlite3.connect(db_path) as conn:
             conn.row_factory = sqlite3.Row
-            # Join asset_images to assets to search by context + title
-            rows = conn.execute(
-                "SELECT ai.image_path, ai.context_text, a.title "
-                "FROM asset_images ai "
-                "JOIN assets a ON ai.asset_id = a.id "
-                "WHERE ai.image_path != '' "
-                "LIMIT 500",
-            ).fetchall()
+            # Pre-filter by topic keywords in SQL for efficiency
+            if keywords:
+                like_clauses = " OR ".join(
+                    "(lower(ai.context_text) LIKE '%' || ? || '%' OR lower(a.title) LIKE '%' || ? || '%')"
+                    for _ in keywords[:5]
+                )
+                params = []
+                for kw in keywords[:5]:
+                    params.extend([kw, kw])
+                rows = conn.execute(
+                    f"SELECT ai.image_path, ai.context_text, a.title "
+                    f"FROM asset_images ai "
+                    f"JOIN assets a ON ai.asset_id = a.id "
+                    f"WHERE ai.image_path != '' AND ({like_clauses}) "
+                    f"LIMIT 100",
+                    params,
+                ).fetchall()
+            else:
+                rows = []
 
         best_path: Optional[str] = None
         best_score = 0
@@ -540,8 +552,8 @@ async def _fetch_teacher_image(
                 best_score = score
                 best_path = row["image_path"]
 
-        # Require minimum score of 2 (at least one keyword match)
-        if best_path and best_score >= 2 and Path(best_path).exists():
+        # Require minimum score of 6 (exact phrase match or 3+ keyword hits)
+        if best_path and best_score >= 6 and Path(best_path).exists():
             logger.info("Using teacher's own image for '%s' -> %s", query, best_path)
             return Path(best_path)
     except Exception as e:

{clawed-2.3.3 → clawed-2.3.4}/clawed/sub_packet.py

@@ -43,6 +43,13 @@ class SubPacket(BaseModel):
     closing_notes: str = ""
     generated_at: datetime = Field(default_factory=datetime.now)
 
+    @field_validator("student_notes", mode="before")
+    @classmethod
+    def _coerce_student_notes(cls, v):
+        if isinstance(v, list):
+            return "\n".join(str(item) for item in v)
+        return v
+
     @field_validator("emergency_info", mode="before")
     @classmethod
     def coerce_emergency_info(cls, v):

{clawed-2.3.3 → clawed-2.3.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "clawed"
-version = "2.3.3"
+version = "2.3.4"
 description = "Claw-ED — personal AI teaching agent. Learns your voice, works while you sleep."
 readme = "README.md"
 license = "MIT"