added api setup

Dockerfile

@@ -0,0 +1,36 @@
+FROM python:3.9-slim
+
+# Install system dependencies for OpenCV and TFOD Protobuf
+RUN apt-get update && apt-get install -y \
+    libgl1-mesa-glx \
+    libglib2.0-0 \
+    protobuf-compiler \
+    git \
+    wget \
+    && rm -rf /var/lib/apt/lists/*
+
+# Set up a new user 
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:$PATH \
+    PYTHONPATH=$HOME/app:$HOME/app/models/research:$HOME/app/models/research/slim
+
+WORKDIR $HOME/app
+
+# Install Python dependencies
+COPY --chown=user requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Clone and install TensorFlow Object Detection API
+RUN git clone --depth 1 https://github.com/tensorflow/models.git
+WORKDIR $HOME/app/models/research
+RUN protoc object_detection/protos/*.proto --python_out=.
+RUN cp object_detection/packages/tf2/setup.py . && pip install .
+
+WORKDIR $HOME/app
+
+# Copy the rest of the application
+COPY --chown=user . .
+
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -0,0 +1,92 @@
+import os
+import io
+import cv2
+import numpy as np
+import tensorflow as tf
+from fastapi import FastAPI, UploadFile, File
+from PIL import Image
+from huggingface_hub import snapshot_download
+
+from object_detection.utils import label_map_util, config_util
+from object_detection.builders import model_builder
+
+app = FastAPI()
+
+# 1. Download Private Models
+HF_TOKEN = os.getenv("HF_Token")
+REPO_ID = "SaniaE/Car_Damage_Detection"
+
+print("Downloading models from Hugging Face...")
+model_dir = snapshot_download(
+    repo_id=REPO_ID, 
+    token=HF_TOKEN, 
+    local_dir="./models_data"
+)
+
+PIPELINE_CONFIG = os.path.join(model_dir, "object_detection_model/pipeline.config")
+CHECKPOINT_PATH = os.path.join(model_dir, "object_detection_model/ckpt-37")
+LABEL_MAP_PATH = os.path.join(model_dir, "object_detection_model/label_map.pbtxt")
+CNN_MODEL_PATH = os.path.join(model_dir, "cnn_filter.h5")
+
+# 3. Load Models
+# Load CNN Filter
+cnn_filter = tf.keras.models.load_model(CNN_MODEL_PATH)
+
+# Load Object Detection Model
+configs = config_util.get_configs_from_pipeline_file(PIPELINE_CONFIG)
+detection_model = model_builder.build(model_config=configs['model'], is_training=False)
+ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
+ckpt.restore(CHECKPOINT_PATH).expect_partial()
+
+category_index = label_map_util.create_category_index_from_labelmap(LABEL_MAP_PATH)
+
+@tf.function
+def detect_fn(image):
+    image, shapes = detection_model.preprocess(image)
+    prediction_dict = detection_model.predict(image, shapes)
+    detections = detection_model.postprocess(prediction_dict, shapes)
+    return detections
+
+@app.get("/")
+def read_root():
+    return {"status": "Model is Online", "model_repo": REPO_ID}
+
+@app.post("/predict")
+async def predict(file: UploadFile = File(...)):
+    # Read Image
+    contents = await file.read()
+    image_pil = Image.open(io.BytesIO(contents)).convert("RGB")
+    image_np = np.array(image_pil)
+    
+    # Step 1: CNN Filter (Check for damage)
+    img_cnn = image_pil.resize((64, 64))
+    x = tf.keras.preprocessing.image.img_to_array(img_cnn)
+    x = np.expand_dims(x, axis=0)
+    
+    cnn_pred = cnn_filter.predict(x)
+    is_damage_labels = ['Clear', 'Damaged']
+    status = is_damage_labels[np.argmax(cnn_pred)]
+    
+    results = {"status": status}
+    
+    # Step 2: Object Detection (If damaged)
+    if status == 'Damaged':
+        input_tensor = tf.convert_to_tensor(np.expand_dims(image_np, 0), dtype=tf.float32)
+        detections = detect_fn(input_tensor)
+        
+        # Format detections for JSON response
+        scores = detections['detection_scores'][0].numpy()
+        classes = detections['detection_classes'][0].numpy().astype(int)
+        boxes = detections['detection_boxes'][0].numpy()
+        
+        detected_objects = []
+        for i in range(len(scores)):
+            if scores[i] > 0.4: # min_score_thresh
+                detected_objects.append({
+                    "label": category_index.get(classes[i] + 1, {}).get('name', 'unknown'),
+                    "score": float(scores[i]),
+                    "box": boxes[i].tolist()
+                })
+        results["detections"] = detected_objects
+        
+    return results

requirements.txt

@@ -0,0 +1,21 @@
+tensorflow==2.13.0
+tf-keras==2.15.0
+keras==2.13.1
+numpy==1.24.3
+pandas==2.0.3
+pillow==10.4.0
+opencv-python-headless==4.10.0.84
+scipy==1.10.1
+protobuf==3.20.3
+pycocotools==2.0.7
+lvis==0.5.3
+gin-config==0.5.0
+tf-slim==1.1.0
+pyyaml==6.0.2
+matplotlib==3.2.0
+lxml==5.3.0
+fastapi
+uvicorn
+python-multipart
+huggingface_hub
+wget