Update main.py

main.py

@@ -1,6 +1,5 @@
 import os
 import zipfile
-import tempfile
 from fastapi import FastAPI, HTTPException
 from pydantic import BaseModel
 
@@ -19,49 +18,14 @@ vectorstore = None
 retriever = None
 chain = None
 
-
 class QueryRequest(BaseModel):
     question: str
 
-
-def _unpack_faiss(src_path: str) -> str:
-    """
-    If src_path is a ZIP, unzip it into a temp dir and return the folder
-    containing the .faiss files; if it’s already a folder, return it.
-    """
-    if zipfile.is_zipfile(src_path):
-        tmp = tempfile.TemporaryDirectory()
-        with zipfile.ZipFile(src_path, "r") as zf:
-            zf.extractall(tmp.name)
-        for root, _, files in os.walk(tmp.name):
-            if any(f.endswith(".faiss") for f in files):
-                return root
-        raise RuntimeError(f"No .faiss index found inside ZIP: {src_path}")
-    elif os.path.isdir(src_path):
-        return src_path
-    else:
-        raise RuntimeError(f"Path is neither a valid ZIP nor a directory: {src_path}")
-
-
-def load_and_merge_faiss(path1: str, path2: str, embeddings: HuggingFaceEmbeddings) -> FAISS:
-    """
-    Load two FAISS indexes (either zip files or folders), merge them,
-    and return the combined FAISS vectorstore.
-    """
-    dir1 = _unpack_faiss(path1)
-    dir2 = _unpack_faiss(path2)
-
-    vs1 = FAISS.load_local(dir1, embeddings, allow_dangerous_deserialization=True)
-    vs2 = FAISS.load_local(dir2, embeddings, allow_dangerous_deserialization=True)
-    vs1.merge_from(vs2)
-    return vs1
-
-
 @app.on_event("startup")
 def load_components():
     global llm, embeddings, vectorstore, retriever, chain
 
-    # --- 1) Init LLM & Embeddings ---
+    # 1) Init LLM & Embeddings
     llm = ChatGroq(
         model="meta-llama/llama-4-scout-17b-16e-instruct",
         temperature=0,
@@ -74,13 +38,42 @@ def load_components():
         encode_kwargs={"normalize_embeddings": True},
     )
 
-    # --- 2) Load & merge two FAISS indexes ---
-    src1 = os.getenv("FAISS_INDEX_PATH_1", "faiss_index.zip")
-    src2 = os.getenv("FAISS_INDEX_PATH_2", "faiss_index_extra.zip")
-    vectorstore = load_and_merge_faiss(src1, src2, embeddings)
+    # 2) Unzip & Load both FAISS vectorstores
+    # — First index
+    zip1 = "faiss_index.zip"
+    dir1 = "faiss_index"
+    if not os.path.exists(dir1):
+        with zipfile.ZipFile(zip1, 'r') as z:
+            z.extractall(dir1)
+        print("✅ Unzipped FAISS index 1.")
+    vs1 = FAISS.load_local(
+        dir1,
+        embeddings,
+        allow_dangerous_deserialization=True
+    )
+    print("✅ FAISS index 1 loaded.")
+
+    # — Second index
+    zip2 = "faiss_index_extra.zip"
+    dir2 = "faiss_index_extra"
+    if not os.path.exists(dir2):
+        with zipfile.ZipFile(zip2, 'r') as z:
+            z.extractall(dir2)
+        print("✅ Unzipped FAISS index 2.")
+    vs2 = FAISS.load_local(
+        dir2,
+        embeddings,
+        allow_dangerous_deserialization=True
+    )
+    print("✅ FAISS index 2 loaded.")
+
+    # 3) Merge them
+    vs1.merge_from(vs2)
+    vectorstore = vs1
+    print("✅ Merged FAISS indexes into a single vectorstore.")
 
-    # --- 3) Build retriever & QA chain ---
-    retriever = vectorstore.as_retriever(search_kwargs={"k": 5})
+    # 4) Create retriever & QA chain
+    retriever = vectorstore.as_retriever(search_kwargs={"k": 3})
     prompt = PromptTemplate(
         template="""
 You are an expert assistant on Islamic knowledge.
@@ -105,15 +98,12 @@ Your response:
         return_source_documents=False,
         chain_type_kwargs={"prompt": prompt},
     )
-
-    print("✅ Loaded & merged both FAISS indexes, QA chain ready.")
-
+    print("✅ QA chain ready.")
 
 @app.get("/")
 def root():
     return {"message": "Arabic Hadith Finder API is up and running!"}
 
-
 @app.post("/query")
 def query(request: QueryRequest):
     try: