rag-search-engine / cli /lib /search_utils.py
Utkarsh736's picture
[Add] Gradio demo + runtime download of movies dataset
9ef70e8 unverified
Raw History Blame Contribute Delete
526 Bytes
import json
import os
DEFAULT_SEARCH_LIMIT = 5
PROJECT_ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..'))
DATA_PATH = os.path.join(PROJECT_ROOT, "data", "movies.json")
STOPWORDS_PATH = os.path.join(PROJECT_ROOT, "data", "stopwords.txt")
def load_movies() -> list[dict]:
with open(DATA_PATH, "r") as f:
data = json.load(f)
return data["movies"]
def load_stopwords() -> list[str]:
with open(STOPWORDS_PATH, "r") as f:
content = f.read()
return content.splitlines()