Initial commit

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+data/IMG_0050.jpg filter=lfs diff=lfs merge=lfs -text
+data/IMG_0051.jpg filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -0,0 +1,27 @@
+FROM python:3.9
+
+WORKDIR /content
+
+RUN mkdir /content/cache/
+
+RUN export TRANSFORMERS_CACHE=/content/cache/
+
+COPY ./requirements.txt /content/requirements.txt
+
+RUN pip install --no-cache-dir --upgrade -r /content/requirements.txt
+RUN apt-get update
+RUN apt-get install -y ffmpeg
+
+COPY . .
+
+RUN adduser --disabled-password --gecos '' admin
+RUN adduser admin sudo
+RUN echo '%sudo ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers
+
+RUN chown -R admin:admin /content
+RUN chmod -R 777 /content
+USER admin
+
+EXPOSE 7860
+
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

main.py

@@ -0,0 +1,84 @@
+import cv2
+import numpy as np
+from fastapi import FastAPI, File, UploadFile
+from fastapi.responses import JSONResponse, Response
+import uvicorn
+import logging
+import time
+import supervision as sv
+from ultralytics import YOLO
+
+app = FastAPI()
+model = YOLO("models/best_v2.pt", task="detect")
+
+
+def parse_detection(detections):
+    parsed_rows = []
+    for i in range(len(detections.xyxy)):
+        x_min = float(detections.xyxy[i][0])
+        y_min = float(detections.xyxy[i][1])
+        x_max = float(detections.xyxy[i][2])
+        y_max = float(detections.xyxy[i][3])
+
+        width = int(x_max - x_min)
+        height = int(y_max - y_min)
+
+        row = {
+            "x": int(y_min),
+            "y": int(x_min),
+            "width": width,
+            "height": height,
+            "class_id": ""
+            if detections.class_id is None
+            else int(detections.class_id[i]),
+            "confidence": ""
+            if detections.confidence is None
+            else float(detections.confidence[i]),
+            "tracker_id": ""
+            if detections.tracker_id is None
+            else int(detections.tracker_id[i]),
+        }
+
+        if hasattr(detections, "data"):
+            for key, value in detections.data.items():
+                if key == "class_name":
+                    key = "class"
+                row[key] = (
+                    str(value[i])
+                    if hasattr(value, "__getitem__") and value.ndim != 0
+                    else str(value)
+                )
+        parsed_rows.append(row)
+    return parsed_rows
+
+
+def infer(image):
+    image_arr = np.frombuffer(image, np.uint8)
+    image = cv2.imdecode(image_arr, cv2.IMREAD_COLOR)
+    image = cv2.resize(image, (1920, 1920))
+    results = model(image)[0]
+    width, height = results.orig_shape[1], results.orig_shape[0]
+    print(results.speed)
+    detections = sv.Detections.from_ultralytics(results)
+    parsed_rows = parse_detection(detections)
+    parsed_result = {'predictions': parsed_rows, 'image': {'width': width, 'height': height}}
+    return parsed_result
+
+
+@app.post("/process-image/")
+async def process_image(image: UploadFile = File(...)):
+    filename = image.filename
+    logging.info(f"Received process-image request for file: {filename}")
+    image_data = await image.read()
+    results = infer(image_data)
+    logging.info("Returning JSON results")
+    return JSONResponse(content=results)
+
+
+@app.get("/")
+def hello_world():
+    return 'Hello World from Detomo AI!'
+
+
+if __name__ == "__main__":
+    uvicorn.run("main:app", port=8001, reload=True)

models/best_v1.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccea111609078c2f7b05161233c3050f67166f82c577d310241cd37bf8ecf4bc
+size 36686178

models/best_v2.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28b05675545530542aa4a3a6a57ce9d48c4af46921b094a034f1d4ac2f7d8101
+size 6259417

requirements.txt

@@ -0,0 +1,15 @@
+fastapi
+uvicorn
+flup
+a2wsgi
+python-multipart
+numpy
+imutils
+opencv-python
+pytesseract
+scikit-learn
+roboflow
+supervision
+ultralytics
+onnx
+onnxruntime

test.py

@@ -0,0 +1,52 @@
+from ultralytics import YOLO
+import supervision as sv
+
+
+def parse_detection(detections):
+    parsed_rows = []
+    for i in range(len(detections.xyxy)):
+        x_min = float(detections.xyxy[i][0])
+        y_min = float(detections.xyxy[i][1])
+        x_max = float(detections.xyxy[i][2])
+        y_max = float(detections.xyxy[i][3])
+
+        width = int(x_max - x_min)
+        height = int(y_max - y_min)
+
+        row = {
+            "x": int(y_min),
+            "y": int(x_min),
+            "width": width,
+            "height": height,
+            "class_id": ""
+            if detections.class_id is None
+            else int(detections.class_id[i]),
+            "confidence": ""
+            if detections.confidence is None
+            else float(detections.confidence[i]),
+            "tracker_id": ""
+            if detections.tracker_id is None
+            else int(detections.tracker_id[i]),
+        }
+
+        if hasattr(detections, "data"):
+            for key, value in detections.data.items():
+                if key == "class_name":
+                    key = "class"
+                row[key] = (
+                    str(value[i])
+                    if hasattr(value, "__getitem__") and value.ndim != 0
+                    else str(value)
+                )
+        parsed_rows.append(row)
+    return parsed_rows
+
+model = YOLO("models/best_v2.pt", task="detect")
+results = model(["data/IMG_0050.jpg"])[0]
+width, height = results.orig_shape[1], results.orig_shape[0]
+print(results.orig_shape)
+print(results.speed)
+output = sv.Detections.from_ultralytics(results)
+output = parse_detection(output)
+parse_result = {'predictions': output, 'image': {'width': width, 'height': height}}
+print(parse_result)