-
Notifications
You must be signed in to change notification settings - Fork 2
/
query_image.py
44 lines (34 loc) · 965 Bytes
/
query_image.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
import io
import json
import os
import base64
import argparse
import weaviate
from PIL import Image
from models.resenet50 import ResNet50Vectorizer
from models.clipmodel import ClipImageEmbed
if __name__ == '__main__':
ap = argparse.ArgumentParser()
ap.add_argument('--image', required=True, help='Image to query for')
args, _ = ap.parse_known_args()
WEAVIATE_URL = os.getenv('WEAVIATE_URL')
if not WEAVIATE_URL:
WEAVIATE_URL = 'http://localhost:8080'
client = weaviate.Client(WEAVIATE_URL)
max_distance = 0.18
# model = ResNet50Vectorizer()
model = ClipImageEmbed()
query_vector = {
'vector': model.embed(args.image),
'distance': max_distance
}
res = (
client.query.get(
'Image', ['filepath']
)
.with_near_vector(query_vector)
.with_limit(5)
.with_additional(['distance'])
.do()
)
print(json.dumps(res, indent=2))