weaviate_quickstart_collection_import_and_near_text_search.py

1import weaviate
2import weaviate.classes as wvc
3import os
4import requests
5import json
6
7# Best practice: store your credentials in environment variables
8# For this example, we'll use a local Weaviate instance or a cloud URL
9client = weaviate.connect_to_local(
10    headers={
11        "X-OpenAI-Api-Key": os.getenv("OPENAI_APIKEY")  # Replace with your actual key
12    }
13)
14
15try:
16    # 1. Create a collection (with a vectorizer)
17    # This defines the data structure and how it should be vectorized
18    questions = client.collections.create(
19        name="Question",
20        vectorizer_config=wvc.config.Configure.Vectorizer.text2vec_openai(),  # Use OpenAI to vectorize text
21        generative_config=wvc.config.Configure.Generative.openai()            # Optional: for RAG
22    )
23
24    # 2. Import data
25    # Load data from a JSON source
26    resp = requests.get('https://raw.githubusercontent.com/weaviate-tutorials/quickstart/main/data/jeopardy_tiny.json')
27    data = json.loads(resp.text)
28
29    # Use a batch process to import data efficiently
30    with questions.batch.dynamic() as batch:
31        for item in data:
32            batch.add_object({
33                "question": item["Question"],
34                "answer": item["Answer"],
35                "category": item["Category"],
36            })
37    
38    # 3. Perform a Search
39    # "Near Text" search finds items with similar meaning to the query
40    response = questions.query.near_text(
41        query="biology",
42        limit=2
43    )
44
45    for obj in response.objects:
46        print(f"ID: {obj.uuid}")
47        print(f"Data: {json.dumps(obj.properties, indent=2)}")
48
49finally:
50    client.close()  # Always close the connection