healthcheck_bm25.py

from pathlib import Path

from llama_index.retrievers.bm25 import BM25Retriever
from main import extract_court_decision_text


PERSIST_PATH = Path("Save_Index_Local")

INDEX_NAME = "bm25_retriever"
# INDEX_NAME = "bm25_retriever_meta"

TEST_CD_URL = "https://reyestr.court.gov.ua/Review/118766467"
# TEST_CD_URL = "https://reyestr.court.gov.ua/Review/118763429"

PRINT_CD = False


retriever = BM25Retriever.from_persist_dir(str(PERSIST_PATH / INDEX_NAME))

court_decision_text = extract_court_decision_text(TEST_CD_URL)

if PRINT_CD:
    print(court_decision_text, "\n\n\n\n\n")

nodes_with_score = retriever.retrieve(court_decision_text)
for index, node_with_score in enumerate(nodes_with_score, start=1):
    source_title = node_with_score.node.metadata.get("title", "Невідомий заголовок")
    print(index, f"{node_with_score.score:.4f}", source_title, "\n", sep="\t")