Spaces:
Sleeping
Sleeping
Update src/streamlit_app.py
Browse files- src/streamlit_app.py +29 -20
src/streamlit_app.py
CHANGED
@@ -10,20 +10,25 @@ from skimage import data as skdata
|
|
10 |
from skimage.io import imsave
|
11 |
import uuid
|
12 |
|
13 |
-
#
|
14 |
-
|
15 |
-
IMAGES_DIR =
|
|
|
16 |
os.makedirs(IMAGES_DIR, exist_ok=True)
|
17 |
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
|
|
|
|
|
|
|
|
25 |
|
26 |
-
# === Image
|
27 |
def extract_images_from_pdf(pdf_bytes):
|
28 |
pdf = fitz.open(stream=pdf_bytes, filetype="pdf")
|
29 |
saved_images = []
|
@@ -47,26 +52,31 @@ def extract_images_from_pdf(pdf_bytes):
|
|
47 |
|
48 |
return saved_images
|
49 |
|
|
|
50 |
def index_images(image_paths):
|
51 |
ids = []
|
52 |
uris = []
|
53 |
-
for
|
54 |
-
if path.endswith((".png", ".jpeg", ".jpg")):
|
55 |
ids.append(str(uuid.uuid4()))
|
56 |
uris.append(path)
|
57 |
|
58 |
if ids:
|
59 |
image_collection.add(ids=ids, uris=uris)
|
60 |
|
|
|
61 |
def query_similar_images(image_file, top_k=5):
|
62 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as tmp:
|
63 |
tmp.write(image_file.read())
|
64 |
tmp_path = tmp.name
|
65 |
|
66 |
-
|
67 |
-
|
68 |
-
|
|
|
|
|
69 |
|
|
|
70 |
def load_skimage_demo_images():
|
71 |
demo_images = {
|
72 |
"astronaut": skdata.astronaut(),
|
@@ -87,7 +97,6 @@ def load_skimage_demo_images():
|
|
87 |
# === Streamlit UI ===
|
88 |
st.title("π Image Similarity Search from PDF or Custom Dataset")
|
89 |
|
90 |
-
# Source Selector
|
91 |
source = st.radio(
|
92 |
"Select Image Source",
|
93 |
["Upload PDF", "Upload Images", "Load Demo Dataset"],
|
@@ -104,7 +113,9 @@ if source == "Upload PDF":
|
|
104 |
st.image(images, width=150)
|
105 |
|
106 |
elif source == "Upload Images":
|
107 |
-
uploaded_imgs = st.file_uploader(
|
|
|
|
|
108 |
if uploaded_imgs:
|
109 |
saved_paths = []
|
110 |
for img in uploaded_imgs:
|
@@ -124,10 +135,8 @@ elif source == "Load Demo Dataset":
|
|
124 |
st.success("Demo images loaded and indexed.")
|
125 |
st.image(demo_paths, width=150)
|
126 |
|
127 |
-
# Divider
|
128 |
st.divider()
|
129 |
|
130 |
-
# Query Interface
|
131 |
st.subheader("π Search for Similar Images")
|
132 |
query_img = st.file_uploader("Upload a query image", type=["jpg", "jpeg", "png"])
|
133 |
if query_img:
|
|
|
10 |
from skimage.io import imsave
|
11 |
import uuid
|
12 |
|
13 |
+
# Use safe temp directories for Streamlit or restricted environments
|
14 |
+
TEMP_DIR = tempfile.gettempdir()
|
15 |
+
IMAGES_DIR = os.path.join(TEMP_DIR, "extracted_images")
|
16 |
+
DB_PATH = os.path.join(TEMP_DIR, "image_vdb")
|
17 |
os.makedirs(IMAGES_DIR, exist_ok=True)
|
18 |
|
19 |
+
@st.cache_resource
|
20 |
+
def get_chroma_collection():
|
21 |
+
chroma_client = PersistentClient(path=DB_PATH)
|
22 |
+
image_loader = ImageLoader()
|
23 |
+
embedding_fn = OpenCLIPEmbeddingFunction()
|
24 |
+
collection = chroma_client.get_or_create_collection(
|
25 |
+
name="image", embedding_function=embedding_fn, data_loader=image_loader
|
26 |
+
)
|
27 |
+
return collection
|
28 |
+
|
29 |
+
image_collection = get_chroma_collection()
|
30 |
|
31 |
+
# === Image Extraction ===
|
32 |
def extract_images_from_pdf(pdf_bytes):
|
33 |
pdf = fitz.open(stream=pdf_bytes, filetype="pdf")
|
34 |
saved_images = []
|
|
|
52 |
|
53 |
return saved_images
|
54 |
|
55 |
+
# === Indexing ===
|
56 |
def index_images(image_paths):
|
57 |
ids = []
|
58 |
uris = []
|
59 |
+
for path in sorted(image_paths):
|
60 |
+
if path.lower().endswith((".png", ".jpeg", ".jpg")):
|
61 |
ids.append(str(uuid.uuid4()))
|
62 |
uris.append(path)
|
63 |
|
64 |
if ids:
|
65 |
image_collection.add(ids=ids, uris=uris)
|
66 |
|
67 |
+
# === Querying ===
|
68 |
def query_similar_images(image_file, top_k=5):
|
69 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as tmp:
|
70 |
tmp.write(image_file.read())
|
71 |
tmp_path = tmp.name
|
72 |
|
73 |
+
try:
|
74 |
+
results = image_collection.query(query_uris=[tmp_path], n_results=top_k)
|
75 |
+
return results['uris'][0]
|
76 |
+
finally:
|
77 |
+
os.remove(tmp_path)
|
78 |
|
79 |
+
# === Demo images ===
|
80 |
def load_skimage_demo_images():
|
81 |
demo_images = {
|
82 |
"astronaut": skdata.astronaut(),
|
|
|
97 |
# === Streamlit UI ===
|
98 |
st.title("π Image Similarity Search from PDF or Custom Dataset")
|
99 |
|
|
|
100 |
source = st.radio(
|
101 |
"Select Image Source",
|
102 |
["Upload PDF", "Upload Images", "Load Demo Dataset"],
|
|
|
113 |
st.image(images, width=150)
|
114 |
|
115 |
elif source == "Upload Images":
|
116 |
+
uploaded_imgs = st.file_uploader(
|
117 |
+
"π€ Upload one or more images", type=["jpg", "jpeg", "png"], accept_multiple_files=True
|
118 |
+
)
|
119 |
if uploaded_imgs:
|
120 |
saved_paths = []
|
121 |
for img in uploaded_imgs:
|
|
|
135 |
st.success("Demo images loaded and indexed.")
|
136 |
st.image(demo_paths, width=150)
|
137 |
|
|
|
138 |
st.divider()
|
139 |
|
|
|
140 |
st.subheader("π Search for Similar Images")
|
141 |
query_img = st.file_uploader("Upload a query image", type=["jpg", "jpeg", "png"])
|
142 |
if query_img:
|