-
Notifications
You must be signed in to change notification settings - Fork 1
/
app.py
144 lines (115 loc) · 4.98 KB
/
app.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
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
import base64
import random
from io import BytesIO
import streamlit as st
import weaviate
import weaviate.classes as wvc
from PIL import Image
from streamlit_image_select import image_select
COLLECTION_NAME = 'Dishes'
NUM_RANDOM_PICS = 10
search_params = {
'return_properties': ['image', 'cuisine'],
'return_metadata': wvc.query.MetadataQuery(distance=True)
}
@st.cache_resource
def get_collection():
# setting up client
client = weaviate.connect_to_local()
# Fetch CRUD collection object
return client.collections.get(COLLECTION_NAME)
# Function to get label for a certain text
def search_by_description(description):
response = dishes.query.near_text(
query=description,
limit=NUM_RANDOM_PICS,
**search_params
)
return response.objects
def search_by_object(dish):
response = dishes.query.near_object(near_object=dish.uuid,
limit=NUM_RANDOM_PICS + 1,
**search_params)
return response.objects
def search_by_image():
response = dishes.query.near_image(near_image=base64_image,
limit=NUM_RANDOM_PICS,
**search_params)
return response.objects
def visualize_response(response_iterable, no_show_first=False):
cols = st.columns(5)
for _ in range((NUM_RANDOM_PICS - 5) // 5):
cols.extend(st.columns(5))
if no_show_first:
response_iterable.pop(0)
for obj, col in zip(response_iterable, cols):
with col:
st.image(decode_image_to_bytes(obj))
st.write(f"{obj.properties['cuisine']} cuisine \n"
f'distance {round(obj.metadata.distance, 3)}')
def get_random_objects(num_of_objects: int):
all_uuids = [dish.uuid for dish in dishes.iterator()]
return [dishes.query.fetch_object_by_id(uuid=uuid, return_properties=['image', 'cuisine']) for uuid in
random.sample(all_uuids, num_of_objects)]
def get_selected_image_index() -> int:
return image_select(label='Random images',
images=[Image.open(decode_image_to_bytes(obj)) for obj in st.session_state.random_images],
return_value='index',
index=-1,
use_container_width=False)
def decode_image_to_bytes(obj):
return BytesIO(base64.b64decode(obj.properties['image']))
st.set_page_config(
page_title="Weaviate Dishes Search",
page_icon=":stuffed_flatbread:",
layout="wide",
)
dishes = get_collection()
col1, col2 = st.columns([1, 7])
col1.image(
'https://avatars.githubusercontent.com/u/37794290?s=200&v=4')
col2.title("Weaviate Dishes Search")
col2.header("Multimodal :camera_with_flash: and Multilingual :globe_with_meridians:")
col2.subheader("Discover a dish that appeals to you through its description, "
"a picture, or by finding one similar to one you already enjoy.")
st.subheader('How do you want to search?')
selection = st.radio("How do you want to search?",
["By description", "By eye", "By picture"],
horizontal=True,
label_visibility='collapsed')
# random_images = get_random_objects(10)
if selection == "By description":
with st.form('description_form'):
st.subheader('Describe the dish you want to find :')
dish_description = st.text_input(label='Describe the dish you want to find :',
label_visibility='collapsed')
submitted = st.form_submit_button('Search')
if submitted and dish_description != '':
search_result = search_by_description(dish_description)
visualize_response(search_result)
if selection == "By eye":
if 'random_images' not in st.session_state:
st.session_state.random_images = get_random_objects(NUM_RANDOM_PICS)
st.subheader("Choose food that you like to search for similar ones")
if st.button("Show me more of what you have"):
st.session_state.random_images = get_random_objects(NUM_RANDOM_PICS)
selected_image = get_selected_image_index()
st.subheader("Selected image:")
if selected_image != -1:
selected_image_object = st.session_state.random_images[selected_image]
st.image(decode_image_to_bytes(selected_image_object))
st.subheader("Dishes that you would like:")
result = search_by_object(selected_image_object)
visualize_response(result, no_show_first=True)
if selection == "By picture":
col1, col2 = st.columns([1, 2])
image_file = col1.file_uploader(label="upload a picture of your dish to find a similar one",
type=['png', 'jpg'])
if image_file is not None:
image_bytes = image_file.getvalue()
col2.image(image=image_bytes,
width=256)
if st.button("Search"):
base64_image = base64.b64encode(image_bytes).decode()
result = search_by_image()
visualize_response(result)