feat: implement graph data management

feat: add knowledge category search
refactor: remove duplicate knowledge results
maintenance: update version to 1.47.0

CHANGELOG.md

@@ -42,6 +42,14 @@
 
 
 
+
+
+## Version 1.46.0 - 2025-11-08
+
+Users can now create, delete, and search for nodes and relationships within a graph. The system can also load graph data from text and manage its database schema.
+
+**Changes:** 3 files, 298 lines
+**Languages:** Markdown (8 lines), Python (288 lines), TOML (2 lines)
 
 ## Version 1.45.0 - 2025-11-08
 

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rp"
-version = "1.45.0"
+version = "1.46.0"
 description = "R python edition. The ultimate autonomous AI CLI."
 readme = "README.md"
 requires-python = ">=3.10"

rp/core/knowledge_context.py

@@ -16,7 +16,22 @@ def inject_knowledge_context(assistant, user_message):
             logger.debug(f"Removed existing knowledge base message at index {i}")
             break
     try:
-        knowledge_results = assistant.enhanced.knowledge_store.search_entries(user_message, top_k=5)
+        # Run all search methods
+        knowledge_results = assistant.enhanced.knowledge_store.search_entries(user_message, top_k=5)  # Hybrid semantic + keyword + category
+        # Additional keyword search if needed (but already in hybrid)
+        # Category-specific: preferences and general
+        pref_results = assistant.enhanced.knowledge_store.get_by_category("preferences", limit=5)
+        general_results = assistant.enhanced.knowledge_store.get_by_category("general", limit=5)
+        category_results = []
+        for entry in pref_results + general_results:
+            if any(word in entry.content.lower() for word in user_message.lower().split()):
+                category_results.append({
+                    "content": entry.content,
+                    "score": 0.6,
+                    "source": f"Knowledge Base ({entry.category})",
+                    "type": "knowledge_category",
+                })
+
         conversation_results = []
         if hasattr(assistant.enhanced, "conversation_memory"):
             history_results = assistant.enhanced.conversation_memory.search_conversations(
@@ -48,6 +63,8 @@ def inject_knowledge_context(assistant, user_message):
                     "type": "knowledge",
                 }
             )
+        for res in category_results:
+            all_results.append(res)
         for conv in conversation_results:
             all_results.append(
                 {
@@ -57,8 +74,15 @@ def inject_knowledge_context(assistant, user_message):
                     "type": "conversation",
                 }
             )
-        all_results.sort(key=lambda x: x["score"], reverse=True)
+        # Remove duplicates by content
-        top_results = all_results[:5]
+        seen = set()
+        unique_results = []
+        for res in all_results:
+            if res["content"] not in seen:
+                seen.add(res["content"])
+                unique_results.append(res)
+        unique_results.sort(key=lambda x: x["score"], reverse=True)
+        top_results = unique_results[:5]
         if not top_results:
             logger.debug("No relevant knowledge or conversation matches found")
             return