Pythonpinecone_search.py
Copy
1
import os
2
import random
3
import time
4
from pinecone import Pinecone, ServerlessSpec
5
 
6
DIM = 384
7
 
8
pc = Pinecone(api_key=os.environ["PINECONE_API_KEY"])
9
 
10
INDEX_NAME = "learn-index"
11
 
12
# Create index if it doesn't exist
13
if INDEX_NAME not in [idx.name for idx in pc.list_indexes()]:
14
    pc.create_index(
15
        name=INDEX_NAME,
16
        dimension=DIM,
17
        metric="cosine",
18
        spec=ServerlessSpec(cloud="aws", region="us-east-1"),
19
    )
20
    # Wait for index to be ready
21
    while not pc.describe_index(INDEX_NAME).status["ready"]:
22
        time.sleep(1)
23
 
24
index = pc.Index(INDEX_NAME)
25
 
26
def random_vector(dim: int) -> list[float]:
27
    v = [random.gauss(0, 1) for _ in range(dim)]
28
    norm = sum(x * x for x in v) ** 0.5
29
    return [x / norm for x in v]
30
 
31
categories = ["news", "blog", "research"]
32
 
33
# Upsert in batches (max 100 per call)
34
vectors = [
35
    {
36
        "id": f"doc-{i}",
37
        "values": random_vector(DIM),
38
        "metadata": {
39
            "category": random.choice(categories),
40
            "source": f"site-{random.randint(1, 5)}.com",
41
            "score": round(random.uniform(0, 1), 3),
42
        },
43
    }
44
    for i in range(100)
45
]
46
index.upsert(vectors=vectors, namespace="demo")
47
print("Upserted 100 vectors")
48
 
49
# Wait for freshness
50
time.sleep(5)
51
 
52
# Filtered similarity query
53
query_vec = random_vector(DIM)
54
results = index.query(
55
    vector=query_vec,
56
    top_k=5,
57
    filter={"category": {"$eq": "news"}},
58
    include_metadata=True,
59
    namespace="demo",
60
)
61
print("Top 5 'news' matches:")
62
for match in results.matches:
63
    print(f"  id={match.id}, score={match.score:.4f}, meta={match.metadata}")
64
 
65
# Fetch by ID
66
fetch_result = index.fetch(ids=["doc-0"], namespace="demo")
67
print(f"Fetched doc-0 metadata: {fetch_result.vectors['doc-0'].metadata}")
68
 
69
# Delete
70
index.delete(ids=["doc-0", "doc-1"], namespace="demo")
71
print("Deleted doc-0 and doc-1")
72