api folder

api/READMEPlantAPI.md

@@ -0,0 +1,170 @@
+
+🌱 Plant Health & Identification API
+
+An intelligent backend service built for our hackathon project. This API uses a dual-model AI system to analyze an image of a plant, first identifying its species and then assessing its health.
+
+🚀 The Problem
+
+Have you ever wondered what kind of plant you have or why its leaves are suddenly turning yellow? Our project aims to provide a simple, accessible answer. This repository contains the backend API that powers our application, capable of providing a comprehensive plant analysis from a single image.
+
+🧠 The AI Pipeline
+
+This API uses a two-stage process to analyze an image:
+
+Species Identification: We use a custom-trained TensorFlow/Keras model, built on a MobileNetV2 architecture. This model was fine-tuned on a dataset of six specific plant types to achieve high accuracy in identifying which plant is in the image.
+
+Health Assessment: Once the plant's species is known (e.g., "tomato"), we use OpenAI's powerful, open-source CLIP model. We dynamically generate text prompts like "a photo of a healthy tomato plant" or "a photo of a sick tomato plant with yellow spots" and ask CLIP which description best matches the image. This zero-shot approach allows for a flexible and nuanced understanding of the plant's health.
+
+✨ Features
+
+Identify 6 Plant Species: Accurately distinguishes between tomato, basil, mint, lettuce, rosemary, and strawberry.
+
+Assess 4 Health States: Classifies plants as Healthy, Diseased, Dehydrated, or Dead.
+
+Confidence Scores: Provides confidence levels for both the species identification and the health assessment.
+
+Simple JSON API: Easy to integrate with any frontend or mobile application.
+
+🛠️ Tech Stack
+
+Backend Framework: Flask
+
+AI / Machine Learning: TensorFlow (Keras), PyTorch, OpenAI CLIP
+
+Image Processing: Pillow, OpenCV
+
+Core Language: Python 3.10
+
+🔌 API Documentation
+
+This is the documentation for the main analysis endpoint.
+
+Analyze a Plant Image
+
+Endpoint: /analyze
+
+Method: POST
+
+Body: multipart/form-data
+
+The request must contain a file field named image.
+
+Successful Response (Status 200 OK)
+
+The API will return a JSON object with the analysis.
+
+Example Response:
+
+Generated json
+{
+  "plant_species": "tomato",
+  "identification_confidence": "97.45%",
+  "health_status": "Healthy",
+  "health_confidence": "89.12%",
+  "health_breakdown": {
+    "Healthy": 0.8912,
+    "Diseased": 0.0562,
+    "Dehydrated": 0.0421,
+    "Dead": 0.0105
+  }
+}
+
+
+Field Descriptions:
+
+Key	Type	Description
+plant_species	String	The identified species of the plant.
+identification_confidence	String	The model's confidence in the species identification.
+health_status	String	The most likely health status of the plant.
+health_confidence	String	The model's confidence in the health assessment.
+health_breakdown	Object	A dictionary of raw probability scores for each health state.
+Error Responses
+
+If the request is missing an image, the API will return:
+
+Status: 400 Bad Request
+
+Body: {"error": "No image file provided"}
+
+For any other server-side issues, the API will return:
+
+Status: 500 Internal Server Error
+
+Body: {"error": "An internal server error occurred."}
+
+🖥️ How to Run Locally
+
+To run this backend server on your own machine, follow these steps.
+
+Clone the Repository:
+
+Generated bash
+git clone <repository-url>
+cd <repository-name>/backend
+IGNORE_WHEN_COPYING_START
+content_copy
+download
+Use code with caution.
+Bash
+IGNORE_WHEN_COPYING_END
+
+Create a Virtual Environment:
+
+Generated bash
+python3 -m venv venv
+source venv/bin/activate
+IGNORE_WHEN_COPYING_START
+content_copy
+download
+Use code with caution.
+Bash
+IGNORE_WHEN_COPYING_END
+
+Install Dependencies:
+This can take a while as it will download TensorFlow and PyTorch.
+
+Generated bash
+pip install -r requirements.txt
+IGNORE_WHEN_COPYING_START
+content_copy
+download
+Use code with caution.
+Bash
+IGNORE_WHEN_COPYING_END
+
+Place the Model:
+Make sure you have the trained Keras model (BestModel.keras) inside the models/ directory.
+
+Run the Server:
+
+Generated bash
+python app.py
+IGNORE_WHEN_COPYING_START
+content_copy
+download
+Use code with caution.
+Bash
+IGNORE_WHEN_COPYING_END
+
+The API will now be running on your local machine at http://127.0.0.1:5000.
+
+📁 Project Structure
+Generated code
+/backend
+|-- app.py             # The main Flask server and API logic.
+|-- models/            # Folder for the trained Keras model.
+|   |-- BestModel.keras
+|-- requirements.txt   # Python dependencies.
+|-- .gitignore         # Files to be ignored by Git (like the venv).
+IGNORE_WHEN_COPYING_START
+content_copy
+download
+Use code with caution.
+IGNORE_WHEN_COPYING_END
+👥 Authors
+
+[Your Name]
+
+[Teammate's Name]
+
+[Teammate's Name]

api/abb.py

@@ -0,0 +1,115 @@
+import io
+import traceback
+from flask import Flask, request, jsonify
+from PIL import Image
+import numpy as np
+import torch
+import clip
+import tensorflow as tf
+
+# --- Configuration ---
+class Config:
+    """Groups all required configuration variables in one place."""
+    PLANT_MODEL_PATH = 'models/BestModel.keras'
+
+    PLANT_CLASSES = ['tomato', 'basil', 'mint', 'lettuce', 'rosemary', 'strawberry']
+    IMG_SIZE = (384, 384)
+    
+    # Health analysis prompt
+    HEALTH_PROMPTS = [
+        "a photo of a healthy {plant} plant with vibrant green leaves",
+        "a photo of a sick {plant} plant with yellow spots or discoloration",
+        "a photo of a dehydrated {plant} plant with wilted or drooping leaves",
+        "a photo of a dead {plant} plant with brown, dry, or crispy leaves"
+    ]
+    HEALTH_LABELS = ["Healthy", "Diseased", "Dehydrated", "Dead"]
+
+# --- Application Setup ---
+app = Flask(__name__)
+
+# --- Model Loading ---
+def load_models():
+    """Loads and initializes all required machine learning models."""
+    # Load plant identification model
+    print("1. Loading plant identification model...")
+    plant_model = tf.keras.models.load_model(Config.PLANT_MODEL_PATH)
+    
+    # Load model for health analysis
+    print("2. Loading CLIP model for health analysis...")
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    clip_model, clip_preprocess = clip.load("ViT-B/32", device=device)
+    
+    print("\nAll models loaded successfully.")
+    return plant_model, clip_model, clip_preprocess, device
+
+# Load models
+plant_model, clip_model, clip_preprocess, device = load_models()
+
+# --- Core ML Functions ---
+def identify_plant(image):
+    """Plant identification"""
+    img = image.resize(Config.IMG_SIZE)
+    img_array = tf.keras.preprocessing.image.img_to_array(img)
+    img_array = tf.expand_dims(img_array, 0)
+    img_array = tf.keras.applications.mobilenet_v2.preprocess_input(img_array)
+
+    preds = plant_model.predict(img_array, verbose=0)
+    best_idx = np.argmax(preds[0])
+    
+    plant_name = Config.PLANT_CLASSES[best_idx]
+    confidence = float(np.max(preds))
+    
+    return plant_name, confidence
+
+def assess_health(plant_name, image):
+    """Plant health"""
+    prompts = [p.format(plant=plant_name) for p in Config.HEALTH_PROMPTS]
+    image_input = clip_preprocess(image).unsqueeze(0).to(device)
+    text_tokens = clip.tokenize(prompts).to(device)
+    
+    with torch.no_grad():
+        image_features = clip_model.encode_image(image_input)
+        text_features = clip_model.encode_text(text_tokens)
+        
+        image_features /= image_features.norm(dim=-1, keepdim=True)
+        text_features /= text_features.norm(dim=-1, keepdim=True)
+        
+        similarity = (100.0 * image_features @ text_features.T).softmax(dim=-1)
+        probs = similarity.cpu().numpy()[0]
+
+    status = Config.HEALTH_LABELS[np.argmax(probs)]
+    confidence = float(np.max(probs))
+    probabilities = {label: float(p) for label, p in zip(Config.HEALTH_LABELS, probs)}
+    
+    return status, confidence, probabilities
+
+# --- API Endpoint ---
+@app.route('/analyze', methods=['POST'])
+def analyze_plant_image():
+    """Image analysis."""
+    if 'image' not in request.files:
+        return jsonify({'error': 'No image file provided'}), 400
+
+    try:
+        file = request.files['image']
+        image = Image.open(io.BytesIO(file.read())).convert("RGB")
+        
+        plant_name, plant_conf = identify_plant(image.copy())
+        health_status, health_conf, health_probs = assess_health(plant_name, image)
+        
+        return jsonify({
+            'plant_species': plant_name,
+            'identification_confidence': f"{plant_conf:.2%}",
+            'health_status': health_status,
+            'health_confidence': f"{health_conf:.2%}",
+            'health_breakdown': health_probs
+        })
+        
+    except Exception as e:
+        print("An error occurred:", str(e))
+        traceback.print_exc()
+        return jsonify({'error': 'An internal server error occurred.'}), 500
+
+# --- Main Execution ---
+if __name__ == '__main__':
+    app.run(host='0.0.0.0', port=5000, debug=True)<EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>

api/requirements.txt

@@ -0,0 +1,55 @@
+absl-py==2.3.1
+astunparse==1.6.3
+blinker==1.9.0
+certifi==2025.7.14
+charset-normalizer==3.4.2
+click==8.2.1
+-e git+https://github.com/openai/CLIP.git@dcba3cb2e2827b402d2701e7e1c7d9fed8a20ef1#egg=clip
+filelock==3.18.0
+Flask==3.1.1
+flatbuffers==25.2.10
+fsspec==2025.7.0
+ftfy==6.3.1
+gast==0.6.0
+google-pasta==0.2.0
+grpcio==1.74.0
+gunicorn==21.2.0
+h5py==3.14.0
+idna==3.10
+itsdangerous==2.2.0
+Jinja2==3.1.6
+keras==3.11.1
+libclang==18.1.1
+Markdown==3.8.2
+markdown-it-py==3.0.0
+MarkupSafe==3.0.2
+mdurl==0.1.2
+ml_dtypes==0.5.3
+mpmath==1.3.0
+namex==0.1.0
+networkx==3.4.2
+numpy==2.1.3
+opt_einsum==3.4.0
+optree==0.17.0
+packaging==25.0
+pillow==11.3.0
+protobuf==5.29.5
+Pygments==2.19.2
+regex==2025.7.34
+requests==2.32.4
+rich==14.1.0
+six==1.17.0
+sympy==1.14.0
+tensorboard==2.19.0
+tensorboard-data-server==0.7.2
+tensorflow==2.19.0
+tensorflow-io-gcs-filesystem==0.37.1
+termcolor==3.1.0
+torch==2.7.1
+torchvision==0.22.1
+tqdm==4.67.1
+typing_extensions==4.14.1
+urllib3==2.5.0
+wcwidth==0.2.13
+Werkzeug==3.1.3
+wrapt==1.17.2