Update app.py
Browse files
app.py
CHANGED
@@ -18,6 +18,32 @@ import matplotlib.pyplot as plt
|
|
18 |
logging.basicConfig(level=logging.INFO)
|
19 |
logger = logging.getLogger(__name__)
|
20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
class RAGSystem:
|
22 |
def __init__(self):
|
23 |
self.embedding_model = SentenceTransformer('all-MiniLM-L6-v2')
|
@@ -27,7 +53,6 @@ class RAGSystem:
|
|
27 |
|
28 |
def load_knowledge_base(self) -> List[Dict]:
|
29 |
"""Load and preprocess knowledge base"""
|
30 |
-
# Using a simplified version of your knowledge base
|
31 |
kb = {
|
32 |
"spalling": [
|
33 |
{
|
@@ -75,7 +100,6 @@ class RAGSystem:
|
|
75 |
D, I = self.vector_store.search(np.array(query_embedding).astype('float32'), k)
|
76 |
context = "\n\n".join([self.knowledge_base[i]["text"] for i in I[0]])
|
77 |
|
78 |
-
# Log query
|
79 |
self.query_history.append({
|
80 |
"timestamp": datetime.now().isoformat(),
|
81 |
"query": query
|
@@ -88,26 +112,37 @@ class RAGSystem:
|
|
88 |
|
89 |
class ImageAnalyzer:
|
90 |
def __init__(self):
|
91 |
-
self.device =
|
92 |
self.defect_classes = ["spalling", "structural_cracks", "surface_deterioration"]
|
93 |
-
self.model =
|
94 |
-
self.processor =
|
95 |
self.history = []
|
96 |
|
97 |
-
def
|
98 |
-
model
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
104 |
|
105 |
def analyze_image(self, image: Image.Image) -> Dict:
|
106 |
"""Analyze image for defects"""
|
107 |
try:
|
|
|
|
|
|
|
108 |
# Preprocess image
|
109 |
-
inputs = self.
|
110 |
-
|
|
|
|
|
111 |
# Get model predictions
|
112 |
with torch.no_grad():
|
113 |
outputs = self.model(**inputs)
|
@@ -123,7 +158,7 @@ class ImageAnalyzer:
|
|
123 |
img_array = np.array(image)
|
124 |
stats = {
|
125 |
"mean_brightness": float(np.mean(img_array)),
|
126 |
-
"image_size": image.size
|
127 |
}
|
128 |
|
129 |
result = {
|
@@ -182,29 +217,51 @@ def main():
|
|
182 |
st.session_state.rag_system = RAGSystem()
|
183 |
if 'image_analyzer' not in st.session_state:
|
184 |
st.session_state.image_analyzer = ImageAnalyzer()
|
|
|
|
|
185 |
|
186 |
# Create two columns
|
187 |
col1, col2 = st.columns([1, 1])
|
188 |
|
189 |
with col1:
|
|
|
190 |
uploaded_file = st.file_uploader(
|
191 |
"Upload a construction image",
|
192 |
-
type=['jpg', 'jpeg', 'png']
|
|
|
193 |
)
|
194 |
|
195 |
-
if uploaded_file:
|
196 |
-
|
197 |
-
|
198 |
-
|
199 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
200 |
results = st.session_state.image_analyzer.analyze_image(image)
|
|
|
201 |
|
202 |
if results:
|
|
|
|
|
|
|
203 |
st.subheader("Detected Defects")
|
204 |
|
205 |
# Create bar chart
|
206 |
defect_probs = results["defect_probabilities"]
|
207 |
-
fig, ax = plt.subplots()
|
208 |
defects = list(defect_probs.keys())
|
209 |
probs = list(defect_probs.values())
|
210 |
ax.barh(defects, probs)
|
@@ -213,11 +270,18 @@ def main():
|
|
213 |
st.pyplot(fig)
|
214 |
|
215 |
# Show image statistics
|
216 |
-
|
217 |
st.json(results["image_statistics"])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
218 |
|
219 |
with col2:
|
220 |
-
st.
|
221 |
user_query = st.text_input(
|
222 |
"Enter your question about construction defects:",
|
223 |
help="Example: What are the repair methods for severe spalling?"
|
@@ -228,11 +292,10 @@ def main():
|
|
228 |
context = st.session_state.rag_system.get_relevant_context(user_query)
|
229 |
response = get_groq_response(user_query, context)
|
230 |
|
231 |
-
st.
|
232 |
st.write(response)
|
233 |
|
234 |
-
|
235 |
-
st.write("Context Used:")
|
236 |
st.text(context)
|
237 |
|
238 |
# Sidebar for history
|
|
|
18 |
logging.basicConfig(level=logging.INFO)
|
19 |
logger = logging.getLogger(__name__)
|
20 |
|
21 |
+
# Ensure CUDA is available or use CPU
|
22 |
+
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
|
23 |
+
|
24 |
+
@st.cache_resource
|
25 |
+
def load_vit_model():
|
26 |
+
"""Load and cache the ViT model"""
|
27 |
+
try:
|
28 |
+
model = ViTForImageClassification.from_pretrained(
|
29 |
+
"google/vit-base-patch16-224",
|
30 |
+
num_labels=3, # Number of defect classes
|
31 |
+
ignore_mismatched_sizes=True
|
32 |
+
)
|
33 |
+
return model.to(DEVICE)
|
34 |
+
except Exception as e:
|
35 |
+
logger.error(f"Error loading ViT model: {e}")
|
36 |
+
return None
|
37 |
+
|
38 |
+
@st.cache_resource
|
39 |
+
def load_vit_processor():
|
40 |
+
"""Load and cache the ViT processor"""
|
41 |
+
try:
|
42 |
+
return ViTImageProcessor.from_pretrained("google/vit-base-patch16-224")
|
43 |
+
except Exception as e:
|
44 |
+
logger.error(f"Error loading ViT processor: {e}")
|
45 |
+
return None
|
46 |
+
|
47 |
class RAGSystem:
|
48 |
def __init__(self):
|
49 |
self.embedding_model = SentenceTransformer('all-MiniLM-L6-v2')
|
|
|
53 |
|
54 |
def load_knowledge_base(self) -> List[Dict]:
|
55 |
"""Load and preprocess knowledge base"""
|
|
|
56 |
kb = {
|
57 |
"spalling": [
|
58 |
{
|
|
|
100 |
D, I = self.vector_store.search(np.array(query_embedding).astype('float32'), k)
|
101 |
context = "\n\n".join([self.knowledge_base[i]["text"] for i in I[0]])
|
102 |
|
|
|
103 |
self.query_history.append({
|
104 |
"timestamp": datetime.now().isoformat(),
|
105 |
"query": query
|
|
|
112 |
|
113 |
class ImageAnalyzer:
|
114 |
def __init__(self):
|
115 |
+
self.device = DEVICE
|
116 |
self.defect_classes = ["spalling", "structural_cracks", "surface_deterioration"]
|
117 |
+
self.model = load_vit_model()
|
118 |
+
self.processor = load_vit_processor()
|
119 |
self.history = []
|
120 |
|
121 |
+
def preprocess_image(self, image: Image.Image) -> torch.Tensor:
|
122 |
+
"""Preprocess image for model input"""
|
123 |
+
try:
|
124 |
+
# Ensure image is RGB
|
125 |
+
if image.mode != 'RGB':
|
126 |
+
image = image.convert('RGB')
|
127 |
+
|
128 |
+
# Process image using ViT processor
|
129 |
+
inputs = self.processor(images=image, return_tensors="pt")
|
130 |
+
return inputs.to(self.device)
|
131 |
+
except Exception as e:
|
132 |
+
logger.error(f"Image preprocessing error: {e}")
|
133 |
+
return None
|
134 |
|
135 |
def analyze_image(self, image: Image.Image) -> Dict:
|
136 |
"""Analyze image for defects"""
|
137 |
try:
|
138 |
+
if self.model is None or self.processor is None:
|
139 |
+
raise ValueError("Model or processor not properly initialized")
|
140 |
+
|
141 |
# Preprocess image
|
142 |
+
inputs = self.preprocess_image(image)
|
143 |
+
if inputs is None:
|
144 |
+
raise ValueError("Image preprocessing failed")
|
145 |
+
|
146 |
# Get model predictions
|
147 |
with torch.no_grad():
|
148 |
outputs = self.model(**inputs)
|
|
|
158 |
img_array = np.array(image)
|
159 |
stats = {
|
160 |
"mean_brightness": float(np.mean(img_array)),
|
161 |
+
"image_size": f"{image.size[0]}x{image.size[1]}"
|
162 |
}
|
163 |
|
164 |
result = {
|
|
|
217 |
st.session_state.rag_system = RAGSystem()
|
218 |
if 'image_analyzer' not in st.session_state:
|
219 |
st.session_state.image_analyzer = ImageAnalyzer()
|
220 |
+
if 'processed_images' not in st.session_state:
|
221 |
+
st.session_state.processed_images = {}
|
222 |
|
223 |
# Create two columns
|
224 |
col1, col2 = st.columns([1, 1])
|
225 |
|
226 |
with col1:
|
227 |
+
st.header("Image Analysis")
|
228 |
uploaded_file = st.file_uploader(
|
229 |
"Upload a construction image",
|
230 |
+
type=['jpg', 'jpeg', 'png'],
|
231 |
+
key="image_uploader"
|
232 |
)
|
233 |
|
234 |
+
if uploaded_file is not None:
|
235 |
+
try:
|
236 |
+
# Display upload progress
|
237 |
+
progress_bar = st.progress(0)
|
238 |
+
status_text = st.empty()
|
239 |
+
|
240 |
+
# Update progress
|
241 |
+
status_text.text("Loading image...")
|
242 |
+
progress_bar.progress(25)
|
243 |
+
|
244 |
+
# Load and display image
|
245 |
+
image = Image.open(uploaded_file)
|
246 |
+
st.image(image, caption="Uploaded Image", use_column_width=True)
|
247 |
+
|
248 |
+
# Update progress
|
249 |
+
status_text.text("Analyzing image...")
|
250 |
+
progress_bar.progress(50)
|
251 |
+
|
252 |
+
# Analyze image
|
253 |
results = st.session_state.image_analyzer.analyze_image(image)
|
254 |
+
progress_bar.progress(75)
|
255 |
|
256 |
if results:
|
257 |
+
status_text.text("Analysis complete!")
|
258 |
+
progress_bar.progress(100)
|
259 |
+
|
260 |
st.subheader("Detected Defects")
|
261 |
|
262 |
# Create bar chart
|
263 |
defect_probs = results["defect_probabilities"]
|
264 |
+
fig, ax = plt.subplots(figsize=(8, 4))
|
265 |
defects = list(defect_probs.keys())
|
266 |
probs = list(defect_probs.values())
|
267 |
ax.barh(defects, probs)
|
|
|
270 |
st.pyplot(fig)
|
271 |
|
272 |
# Show image statistics
|
273 |
+
with st.expander("Image Details"):
|
274 |
st.json(results["image_statistics"])
|
275 |
+
else:
|
276 |
+
status_text.text("Analysis failed. Please try again.")
|
277 |
+
progress_bar.empty()
|
278 |
+
|
279 |
+
except Exception as e:
|
280 |
+
st.error(f"Error processing image: {str(e)}")
|
281 |
+
logger.error(f"Image processing error: {e}")
|
282 |
|
283 |
with col2:
|
284 |
+
st.header("Ask About Defects")
|
285 |
user_query = st.text_input(
|
286 |
"Enter your question about construction defects:",
|
287 |
help="Example: What are the repair methods for severe spalling?"
|
|
|
292 |
context = st.session_state.rag_system.get_relevant_context(user_query)
|
293 |
response = get_groq_response(user_query, context)
|
294 |
|
295 |
+
st.subheader("AI Response")
|
296 |
st.write(response)
|
297 |
|
298 |
+
with st.expander("View Context Used"):
|
|
|
299 |
st.text(context)
|
300 |
|
301 |
# Sidebar for history
|