Delete qa_markdown_faiss_hnsw_externalized.py

qa_markdown_faiss_hnsw_externalized.py

@@ -1,49 +0,0 @@
-#!/usr/bin/env python3
-from qa_common import parse_args, build_prompt, generate_answer, save_result
-import faiss
-import pickle
-from sentence_transformers import SentenceTransformer
-from llama_cpp import Llama
-
-# ====================== FAISS HNSW SPECIFIC ======================
-INDEX_PATH = "faiss_hnsw/vector_search.index"
-METADATA_PATH = "faiss_hnsw/metadata.pkl"
-MODEL_PATH = "articles-Q4_K_M.gguf"
-
-print("Loading embedding model...")
-embed_model = SentenceTransformer("BAAI/bge-small-en-v1.5")
-
-print("Loading FAISS HNSW index...")
-index = faiss.read_index(INDEX_PATH)
-
-print("Loading metadata...")
-with open(METADATA_PATH, "rb") as f:
-    metadata = pickle.load(f)
-
-print("Loading LLM...")
-llm = Llama(model_path=MODEL_PATH, n_ctx=25000, n_threads=8, verbose=False)
-
-
-def get_context(query: str, k=5) -> str:
-    query_vec = embed_model.encode([query], normalize_embeddings=True).astype('float32')
-    _, indices = index.search(query_vec, k)
-    
-    chunks = []
-    for idx in indices[0]:
-        row = metadata.iloc[idx]
-        chunk = f"[Article: {row['article_title']}] \n{row['article_content']}"
-        chunks.append(chunk)
-    return "\n\n".join(chunks)
-
-
-if __name__ == "__main__":
-    args = parse_args()
-    query = args.prompt if args.prompt else input("\nQuestion: ")
-
-    print("Retrieving context and generating answer...\n")
-    
-    context = get_context(query, k=5)
-    prompt = build_prompt(query, context)
-    answer = generate_answer(llm, prompt)
-    
-    save_result(query, answer, args.output)