Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -18,7 +18,7 @@ from PIL import Image
|
|
| 18 |
DEFAULT_MODEL_PATH = "models/yolov8_safety.pt"
|
| 19 |
FALLBACK_MODEL = "yolov8n.pt"
|
| 20 |
MODEL_PATH = os.getenv("SAFETY_MODEL_PATH", DEFAULT_MODEL_PATH)
|
| 21 |
-
OUTPUT_DIR = "output"
|
| 22 |
os.makedirs(OUTPUT_DIR, exist_ok=True)
|
| 23 |
|
| 24 |
VIOLATION_LABELS = {
|
|
@@ -37,18 +37,25 @@ print(f"✅ Using device: {device}")
|
|
| 37 |
# ==========================
|
| 38 |
# Load Model
|
| 39 |
# ==========================
|
| 40 |
-
|
| 41 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
|
| 43 |
# ==========================
|
| 44 |
# Video Processing
|
| 45 |
# ==========================
|
| 46 |
def process_video(video_data, frame_skip=5, max_frames=100):
|
| 47 |
try:
|
|
|
|
| 48 |
# Save uploaded video data to a temporary file
|
| 49 |
video_path = os.path.join(OUTPUT_DIR, f"temp_{int(time.time())}.mp4")
|
| 50 |
with open(video_path, "wb") as f:
|
| 51 |
f.write(video_data)
|
|
|
|
| 52 |
|
| 53 |
video = cv2.VideoCapture(video_path)
|
| 54 |
if not video.isOpened():
|
|
@@ -94,7 +101,7 @@ def process_video(video_data, frame_skip=5, max_frames=100):
|
|
| 94 |
snapshots.append({
|
| 95 |
"violation": label,
|
| 96 |
"frame": frame_count,
|
| 97 |
-
"snapshot_url": snapshot_path
|
| 98 |
})
|
| 99 |
|
| 100 |
frame_count += 1
|
|
@@ -108,7 +115,7 @@ def process_video(video_data, frame_skip=5, max_frames=100):
|
|
| 108 |
break
|
| 109 |
|
| 110 |
video.release()
|
| 111 |
-
os.remove(video_path)
|
| 112 |
|
| 113 |
score = calculate_safety_score(violations)
|
| 114 |
pdf_report_path = generate_pdf_report(violations, snapshots, score)
|
|
@@ -149,61 +156,74 @@ def calculate_safety_score(violations):
|
|
| 149 |
# PDF Report Generation
|
| 150 |
# ==========================
|
| 151 |
def generate_pdf_report(violations, snapshots, score):
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
for v in violations:
|
| 171 |
-
c.setFont("Helvetica", 10)
|
| 172 |
-
text = f"Violation: {v['violation']}, Timestamp: {v['timestamp']:.2f}s, Confidence: {v['confidence']}"
|
| 173 |
-
c.drawString(50, y, text)
|
| 174 |
y -= 20
|
| 175 |
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 189 |
|
| 190 |
# ==========================
|
| 191 |
# Gradio Interface
|
| 192 |
# ==========================
|
| 193 |
def gradio_interface(video_file):
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
| 201 |
-
|
| 202 |
-
|
| 203 |
-
|
| 204 |
-
|
| 205 |
-
|
| 206 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 207 |
|
| 208 |
interface = gr.Interface(
|
| 209 |
fn=gradio_interface,
|
|
|
|
| 18 |
DEFAULT_MODEL_PATH = "models/yolov8_safety.pt"
|
| 19 |
FALLBACK_MODEL = "yolov8n.pt"
|
| 20 |
MODEL_PATH = os.getenv("SAFETY_MODEL_PATH", DEFAULT_MODEL_PATH)
|
| 21 |
+
OUTPUT_DIR = "output"
|
| 22 |
os.makedirs(OUTPUT_DIR, exist_ok=True)
|
| 23 |
|
| 24 |
VIOLATION_LABELS = {
|
|
|
|
| 37 |
# ==========================
|
| 38 |
# Load Model
|
| 39 |
# ==========================
|
| 40 |
+
try:
|
| 41 |
+
selected_model = MODEL_PATH if os.path.isfile(MODEL_PATH) else FALLBACK_MODEL
|
| 42 |
+
model = YOLO(selected_model)
|
| 43 |
+
print(f"✅ Model loaded: {selected_model}")
|
| 44 |
+
except Exception as e:
|
| 45 |
+
print(f"❌ Failed to load model: {e}")
|
| 46 |
+
raise
|
| 47 |
|
| 48 |
# ==========================
|
| 49 |
# Video Processing
|
| 50 |
# ==========================
|
| 51 |
def process_video(video_data, frame_skip=5, max_frames=100):
|
| 52 |
try:
|
| 53 |
+
print("Processing video data...")
|
| 54 |
# Save uploaded video data to a temporary file
|
| 55 |
video_path = os.path.join(OUTPUT_DIR, f"temp_{int(time.time())}.mp4")
|
| 56 |
with open(video_path, "wb") as f:
|
| 57 |
f.write(video_data)
|
| 58 |
+
print(f"Video saved to {video_path}")
|
| 59 |
|
| 60 |
video = cv2.VideoCapture(video_path)
|
| 61 |
if not video.isOpened():
|
|
|
|
| 101 |
snapshots.append({
|
| 102 |
"violation": label,
|
| 103 |
"frame": frame_count,
|
| 104 |
+
"snapshot_url": f"https://huggingface.co/spaces/PrashanthB461/AI_Safety_Demo1/output/{os.path.basename(snapshot_path)}"
|
| 105 |
})
|
| 106 |
|
| 107 |
frame_count += 1
|
|
|
|
| 115 |
break
|
| 116 |
|
| 117 |
video.release()
|
| 118 |
+
os.remove(video_path)
|
| 119 |
|
| 120 |
score = calculate_safety_score(violations)
|
| 121 |
pdf_report_path = generate_pdf_report(violations, snapshots, score)
|
|
|
|
| 156 |
# PDF Report Generation
|
| 157 |
# ==========================
|
| 158 |
def generate_pdf_report(violations, snapshots, score):
|
| 159 |
+
try:
|
| 160 |
+
pdf_path = os.path.join(OUTPUT_DIR, f"report_{int(time.time())}.pdf")
|
| 161 |
+
c = canvas.Canvas(pdf_path, pagesize=letter)
|
| 162 |
+
width, height = letter
|
| 163 |
+
|
| 164 |
+
# Title
|
| 165 |
+
c.setFont("Helvetica-Bold", 16)
|
| 166 |
+
c.drawString(50, height - 50, "Worksite Safety Compliance Report")
|
| 167 |
+
|
| 168 |
+
# Compliance Score
|
| 169 |
+
c.setFont("Helvetica", 12)
|
| 170 |
+
c.drawString(50, height - 80, f"Compliance Score: {score}%")
|
| 171 |
+
|
| 172 |
+
# Violations Table
|
| 173 |
+
y = height - 120
|
| 174 |
+
c.setFont("Helvetica-Bold", 12)
|
| 175 |
+
c.drawString(50, y, "Detected Violations:")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 176 |
y -= 20
|
| 177 |
|
| 178 |
+
for v in violations:
|
| 179 |
+
c.setFont("Helvetica", 10)
|
| 180 |
+
text = f"Violation: {v['violation']}, Timestamp: {v['timestamp']:.2f}s, Confidence: {v['confidence']}"
|
| 181 |
+
c.drawString(50, y, text)
|
| 182 |
+
y -= 20
|
| 183 |
+
|
| 184 |
+
# Add snapshot if available
|
| 185 |
+
snapshot = next((s for s in snapshots if s["frame"] == v["frame"] and s["violation"] == v["violation"]), None)
|
| 186 |
+
if snapshot and os.path.exists(snapshot["snapshot_url"].split('/')[-1]):
|
| 187 |
+
img = ImageReader(snapshot["snapshot_url"].split('/')[-1])
|
| 188 |
+
c.drawImage(img, 50, y - 100, width=200, height=150)
|
| 189 |
+
y -= 170
|
| 190 |
+
|
| 191 |
+
if y < 50:
|
| 192 |
+
c.showPage()
|
| 193 |
+
y = height - 50
|
| 194 |
+
|
| 195 |
+
c.save()
|
| 196 |
+
print(f"PDF generated at {pdf_path}")
|
| 197 |
+
# Return a publicly accessible URL
|
| 198 |
+
base_url = "https://huggingface.co/spaces/PrashanthB461/AI_Safety_Demo1"
|
| 199 |
+
pdf_url = f"{base_url}/output/{os.path.basename(pdf_path)}"
|
| 200 |
+
print(f"PDF URL: {pdf_url}")
|
| 201 |
+
return pdf_url
|
| 202 |
+
except Exception as e:
|
| 203 |
+
print(f"❌ Error generating PDF: {e}")
|
| 204 |
+
return ""
|
| 205 |
|
| 206 |
# ==========================
|
| 207 |
# Gradio Interface
|
| 208 |
# ==========================
|
| 209 |
def gradio_interface(video_file):
|
| 210 |
+
try:
|
| 211 |
+
if not video_file:
|
| 212 |
+
return {"error": "Please upload a video file."}, "", "", []
|
| 213 |
+
|
| 214 |
+
with open(video_file, "rb") as f:
|
| 215 |
+
video_data = f.read()
|
| 216 |
+
|
| 217 |
+
result = process_video(video_data)
|
| 218 |
+
return (
|
| 219 |
+
result["violations"],
|
| 220 |
+
f"Safety Score: {result['score']}%",
|
| 221 |
+
result["pdf_report_url"],
|
| 222 |
+
result["snapshots"]
|
| 223 |
+
)
|
| 224 |
+
except Exception as e:
|
| 225 |
+
print(f"❌ Error in gradio_interface: {e}")
|
| 226 |
+
return {"error": str(e)}, "", "", []
|
| 227 |
|
| 228 |
interface = gr.Interface(
|
| 229 |
fn=gradio_interface,
|