Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -9,8 +9,7 @@ import cv2
|
|
9 |
import numpy as np
|
10 |
from PIL import Image
|
11 |
import torch
|
12 |
-
|
13 |
-
# Adjust import paths as needed
|
14 |
sys.path.append('Utils')
|
15 |
sys.path.append('model')
|
16 |
from model.CBAM.reunet_cbam import reunet_cbam
|
@@ -21,7 +20,7 @@ from split_merge import split, merge
|
|
21 |
from Utils.convert import read_pansharpened_rgb
|
22 |
|
23 |
# Define base directory for Hugging Face Spaces
|
24 |
-
BASE_DIR = "
|
25 |
|
26 |
# Define subdirectories
|
27 |
UPLOAD_DIR = os.path.join(BASE_DIR, "uploaded_images")
|
@@ -34,6 +33,18 @@ CSV_LOG_PATH = os.path.join(BASE_DIR, "image_log.csv")
|
|
34 |
for directory in [UPLOAD_DIR, MASK_DIR, PATCHES_DIR, PRED_PATCHES_DIR]:
|
35 |
os.makedirs(directory, exist_ok=True)
|
36 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
# Load model
|
38 |
@st.cache_resource
|
39 |
def load_model():
|
@@ -49,6 +60,18 @@ def predict(image):
|
|
49 |
output = model(image.unsqueeze(0))
|
50 |
return output.squeeze().cpu().numpy()
|
51 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
def log_image_details(image_id, image_filename, mask_filename):
|
53 |
file_exists = os.path.exists(CSV_LOG_PATH)
|
54 |
|
@@ -70,6 +93,9 @@ def log_image_details(image_id, image_filename, mask_filename):
|
|
70 |
sno = 1
|
71 |
|
72 |
writer.writerow([sno, date, time, image_id, image_filename, mask_filename])
|
|
|
|
|
|
|
73 |
|
74 |
def upload_page():
|
75 |
if 'file_uploaded' not in st.session_state:
|
@@ -106,6 +132,9 @@ def upload_page():
|
|
106 |
|
107 |
st.success(f"Image saved to {filepath}")
|
108 |
|
|
|
|
|
|
|
109 |
# Check if the uploaded file is a GeoTIFF
|
110 |
if file_extension in ['.tiff', '.tif']:
|
111 |
st.info('Processing GeoTIFF image...')
|
@@ -169,6 +198,13 @@ def upload_page():
|
|
169 |
Image.fromarray(mask).save(mask_filepath)
|
170 |
st.session_state.mask_filename = mask_filename
|
171 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
172 |
st.session_state.file_uploaded = True
|
173 |
|
174 |
except Exception as e:
|
@@ -222,7 +258,7 @@ def result_page():
|
|
222 |
original_np = cv2.imread(original_img_path)
|
223 |
mask_np = cv2.imread(mask_path, cv2.IMREAD_GRAYSCALE)
|
224 |
|
225 |
-
#
|
226 |
_, mask_np = cv2.threshold(mask_np, 127, 255, cv2.THRESH_BINARY)
|
227 |
|
228 |
# Resize mask to match original image size if necessary
|
|
|
9 |
import numpy as np
|
10 |
from PIL import Image
|
11 |
import torch
|
12 |
+
from huggingface_hub import HfApi
|
|
|
13 |
sys.path.append('Utils')
|
14 |
sys.path.append('model')
|
15 |
from model.CBAM.reunet_cbam import reunet_cbam
|
|
|
20 |
from Utils.convert import read_pansharpened_rgb
|
21 |
|
22 |
# Define base directory for Hugging Face Spaces
|
23 |
+
BASE_DIR = "data"
|
24 |
|
25 |
# Define subdirectories
|
26 |
UPLOAD_DIR = os.path.join(BASE_DIR, "uploaded_images")
|
|
|
33 |
for directory in [UPLOAD_DIR, MASK_DIR, PATCHES_DIR, PRED_PATCHES_DIR]:
|
34 |
os.makedirs(directory, exist_ok=True)
|
35 |
|
36 |
+
# Initialize Hugging Face API
|
37 |
+
hf_api = HfApi()
|
38 |
+
|
39 |
+
# Get the token from environment variable
|
40 |
+
HF_TOKEN = os.environ.get("HF_TOKEN")
|
41 |
+
if not HF_TOKEN:
|
42 |
+
st.error("HF_TOKEN not found in environment variables. Please set it in your Space settings.")
|
43 |
+
st.stop()
|
44 |
+
|
45 |
+
|
46 |
+
REPO_ID = "Pavan2k4/Building_area"
|
47 |
+
|
48 |
# Load model
|
49 |
@st.cache_resource
|
50 |
def load_model():
|
|
|
60 |
output = model(image.unsqueeze(0))
|
61 |
return output.squeeze().cpu().numpy()
|
62 |
|
63 |
+
def save_to_hf_repo(local_path, repo_path):
|
64 |
+
try:
|
65 |
+
hf_api.upload_file(
|
66 |
+
path_or_fileobj=local_path,
|
67 |
+
path_in_repo=repo_path,
|
68 |
+
repo_id=REPO_ID,
|
69 |
+
token=HF_TOKEN
|
70 |
+
)
|
71 |
+
st.success(f"File uploaded successfully to {repo_path}")
|
72 |
+
except Exception as e:
|
73 |
+
st.error(f"Error uploading file: {str(e)}")
|
74 |
+
|
75 |
def log_image_details(image_id, image_filename, mask_filename):
|
76 |
file_exists = os.path.exists(CSV_LOG_PATH)
|
77 |
|
|
|
93 |
sno = 1
|
94 |
|
95 |
writer.writerow([sno, date, time, image_id, image_filename, mask_filename])
|
96 |
+
|
97 |
+
# Save CSV to Hugging Face repo
|
98 |
+
save_to_hf_repo(CSV_LOG_PATH, 'image_log.csv')
|
99 |
|
100 |
def upload_page():
|
101 |
if 'file_uploaded' not in st.session_state:
|
|
|
132 |
|
133 |
st.success(f"Image saved to {filepath}")
|
134 |
|
135 |
+
# Save image to Hugging Face repo
|
136 |
+
save_to_hf_repo(filepath, f'uploaded_images/{filename}')
|
137 |
+
|
138 |
# Check if the uploaded file is a GeoTIFF
|
139 |
if file_extension in ['.tiff', '.tif']:
|
140 |
st.info('Processing GeoTIFF image...')
|
|
|
198 |
Image.fromarray(mask).save(mask_filepath)
|
199 |
st.session_state.mask_filename = mask_filename
|
200 |
|
201 |
+
# Save mask to Hugging Face repo
|
202 |
+
mask_filepath = os.path.join(MASK_DIR, st.session_state.mask_filename)
|
203 |
+
save_to_hf_repo(mask_filepath, f'generated_masks/{st.session_state.mask_filename}')
|
204 |
+
|
205 |
+
# Log image details
|
206 |
+
log_image_details(timestamp, converted_filename, st.session_state.mask_filename)
|
207 |
+
|
208 |
st.session_state.file_uploaded = True
|
209 |
|
210 |
except Exception as e:
|
|
|
258 |
original_np = cv2.imread(original_img_path)
|
259 |
mask_np = cv2.imread(mask_path, cv2.IMREAD_GRAYSCALE)
|
260 |
|
261 |
+
# mask is binary
|
262 |
_, mask_np = cv2.threshold(mask_np, 127, 255, cv2.THRESH_BINARY)
|
263 |
|
264 |
# Resize mask to match original image size if necessary
|