ably.do/finding.py

import weaviate
from weaviate.connect import ConnectionParams
import re

# Konfiguracja klienta Weaviate
client = weaviate.WeaviateClient(
    connection_params=ConnectionParams.from_params(
        http_host="weaviate",
        http_port=8080,
        http_secure=False,
        grpc_host="weaviate",
        grpc_port=50051,
        grpc_secure=False,
    )
)
client.connect()

# Pobierz kolekcję
collection = client.collections.get("Document")

def extract_full_article(content, article_number):
    pattern = rf"Art\.\s*{article_number}\..*?(?=Art\.\s*\d+\.|\Z)"
    match = re.search(pattern, content, re.DOTALL)
    if match:
        return match.group(0).strip()
    return None

def extract_relevant_fragment(content, query, context_size=100):
    article_match = re.match(r"Art\.\s*(\d+)", query)
    if article_match:
        article_number = article_match.group(1)
        full_article = extract_full_article(content, article_number)
        if full_article:
            return full_article

    index = content.lower().find(query.lower())
    if index != -1:
        start = max(0, index - context_size)
        end = min(len(content), index + len(query) + context_size)
        return f"...{content[start:end]}..."
    return content[:200] + "..."

def vector_search(query, limit=5):
    print(f"\nWyszukiwanie wektorowe dla zapytania: '{query}'")
    response = collection.query.near_text(
        query=query,
        limit=limit
    )
    for obj in response.objects:
        print(f"UUID: {obj.uuid}")
        relevant_fragment = extract_relevant_fragment(obj.properties['content'], query)
        print(f"Relewantny fragment:\n{relevant_fragment}")
        print(f"Nazwa pliku: {obj.properties['fileName']}")
        print("---")

def hybrid_search(query, limit=5, alpha=0.5):
    print(f"\nWyszukiwanie hybrydowe dla zapytania: '{query}'")
    response = collection.query.hybrid(
        query=query,
        alpha=alpha,
        limit=limit
    )
    for obj in response.objects:
        print(f"UUID: {obj.uuid}")
        relevant_fragment = extract_relevant_fragment(obj.properties['content'], query)
        print(f"Relewantny fragment:\n{relevant_fragment}")
        print(f"Nazwa pliku: {obj.properties['fileName']}")
        print("---")


#exists = client.collections.exists("Document")
#print(f"Czy kolekcja 'Document' istnieje: {exists}")

#schema = collection.config.get()
#print(f"Nazwa kolekcji: {schema.name}")
#print("Właściwości:")
#for prop in schema.properties:
#    print(f"- {prop.name}: {prop.data_type}")

#collection = client.collections.get("Document")
#count = collection.aggregate.over_all(total_count=True).total_count
#print(f"Liczba obiektów w kolekcji: {count}")

#results = collection.query.fetch_objects(limit=5)
#for obj in results.objects:
#    print(f"UUID: {obj.uuid}")
#    print(f"Nazwa pliku: {obj.properties['fileName']}")
#    print(f"Zawartość: {obj.properties['content'][:100]}...")  # Pierwsze 100 znaków
#    print("---")


# Przykładowe użycie
queries = ["Art. 154", "urlop wypoczynkowy", "Państwowa Inspekcja Pracy", "Art. 154 Kodeks pracy"]

for query in queries:
    vector_search(query)
    hybrid_search(query)

# Zamknij połączenie
client.close()