Initial implementation of FAQ bot with ChromaDB and MCP-style tool: update src/utils.py

src/utils.py

@@ -1,11 +1,14 @@
 import os
 from pathlib import Path
+import json
+import httpx
 from langchain_ollama import OllamaEmbeddings
 from langchain_chroma import Chroma
 from langchain_text_splitters import RecursiveCharacterTextSplitter
 
 CHROMA_PATH = Path("./chroma_faq")
 
+# Load FAQ markdown files into ChromaDB
 def load_faq_to_chroma(md_dir: str = "data"):
     Path(md_dir).mkdir(parents=True, exist_ok=True)
     md_files = list(Path(md_dir).glob("*.md"))
@@ -22,7 +25,23 @@ def load_faq_to_chroma(md_dir: str = "data"):
     chroma.persist()
     return chroma
 
+# Search the ChromaDB for relevant documents
 def search_course_docs(query: str, k: int = 3):
     chroma = Chroma(persist_directory=str(CHROMA_PATH), embedding_function=OllamaEmbeddings(model="nomic-embed-text"))
     retriever = chroma.as_retriever(search_kwargs={"k": k})
     return retriever.get_relevant_documents(query)
+
+# MCP‑style tool: fetch metadata from a mock HTTP endpoint
+# In production this would be a real MCP server. Here we use a local JSON file served by http.server.
+def fetch_course_meta(query: str):
+    url = f"http://localhost:8000/course_meta.json"
+    try:
+        response = httpx.get(url, timeout=5.0)
+        response.raise_for_status()
+        data = response.json()
+    except Exception:
+        # Fallback to local static file if server not running
+        data = json.loads(Path("data/course_meta.json").read_text(encoding="utf-8"))
+    # Simple filtering: return items where query is in title or description
+    results = [item for item in data if query.lower() in item.get("title", "").lower() or query.lower() in item.get("description", "").lower()]
+    return results