File size: 13,444 Bytes
ab5071c 6c8b2ac ab5071c 6c8b2ac ab5071c 0a4b653 ab5071c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 |
import torch
import torch.nn as nn
from torchvision import transforms
from PIL import Image
import gradio as gr
import networkx as nx
import os
from transformers import AutoModel, AutoConfig
from openai import OpenAI
import json
import gdown
# .....................................
MODEL_PATH = "best.pt"
if not os.path.exists(MODEL_PATH):
# Replace FILE_ID with Google Drive file ID
file_id = "1bGRLEC2_5GB53E-zEVH1Z4EQdKGA-YGI"
url = f"https://drive.google.com/uc?id={file_id}"
print("Downloading model from Google Drive...")
gdown.download(url, MODEL_PATH, quiet=False)
# -------------------------------
# Load Knowledge Graph + Symptom Map
# -------------------------------
# 1. Create Extended Knowledge Graph (KG)
# -------------------------------
G = nx.DiGraph()
# -------------------------------
# Symptoms β Diseases
# -------------------------------
# Apple
G.add_edge("Olive-brown velvety spots on leaves/fruits", "Apple___Apple_scab")
G.add_edge("Dark sunken lesions with concentric rings", "Apple___Black_rot")
G.add_edge("Orange/yellow spots with black centers", "Apple___Cedar_apple_rust")
G.add_edge("No visible disease", "Apple___healthy")
# Blueberry
G.add_edge("No visible disease", "Blueberry___healthy")
# Cherry
G.add_edge("White powdery coating on leaves", "Cherry_(including_sour)___Powdery_mildew")
G.add_edge("No visible disease", "Cherry_(including_sour)___healthy")
# Corn
G.add_edge("Gray/tan lesions with dark borders", "Corn_(maize)___Cercospora_leaf_spot Gray_leaf_spot")
G.add_edge("Small reddish-brown pustules on leaves", "Corn_(maize)___Common_rust_")
G.add_edge("Cigar-shaped gray-green lesions on leaves", "Corn_(maize)___Northern_Leaf_Blight")
G.add_edge("No visible disease", "Corn_(maize)___healthy")
# Grape
G.add_edge("Circular black spots on leaves/fruits", "Grape___Black_rot")
G.add_edge("Interveinal chlorosis, necrosis (black measles)", "Grape___Esca_(Black_Measles)")
G.add_edge("Irregular brown spots with yellow halo", "Grape___Leaf_blight_(Isariopsis_Leaf_Spot)")
G.add_edge("No visible disease", "Grape___healthy")
# Orange
G.add_edge("Yellow shoots, mottled leaves, misshapen fruits", "Orange___Haunglongbing_(Citrus_greening)")
# Peach
G.add_edge("Small dark water-soaked spots on leaves", "Peach___Bacterial_spot")
G.add_edge("No visible disease", "Peach___healthy")
# Pepper (typo fix: Pepper instead of 'pper,_bell')
G.add_edge("Brown lesions with yellow halo", "Pepper,_bell___Bacterial_spot")
G.add_edge("No visible disease", "Pepper,_bell___healthy")
# Potato
G.add_edge("Dark concentric spots on leaves", "Potato___Early_blight")
G.add_edge("Large irregular brown/black lesions", "Potato___Late_blight")
G.add_edge("No visible disease", "Potato___healthy")
# Raspberry
G.add_edge("No visible disease", "Raspberry___healthy")
# Soybean
G.add_edge("No visible disease", "Soybean___healthy")
# Squash
G.add_edge("White powdery patches on leaves", "Squash___Powdery_mildew")
# Strawberry
G.add_edge("Irregular brown leaf margins, scorching", "Strawberry___Leaf_scorch")
G.add_edge("No visible disease", "Strawberry___healthy")
# Tomato
G.add_edge("Water-soaked brown spots on leaves", "Tomato___Bacterial_spot")
G.add_edge("Concentric rings, target-like spots", "Tomato___Early_blight")
G.add_edge("Large dark blotches with fuzzy growth", "Tomato___Late_blight")
G.add_edge("Yellow patches on upper leaf, fuzzy underside", "Tomato___Leaf_Mold")
G.add_edge("Small circular dark spots with yellow halo", "Tomato___Septoria_leaf_spot")
G.add_edge("White/yellow stippling + webbing", "Tomato___Spider_mites Two-spotted_spider_mite")
G.add_edge("Brown/black target-like spots", "Tomato___Target_Spot")
G.add_edge("Leaf curling + yellow mosaic", "Tomato___Tomato_Yellow_Leaf_Curl_Virus")
G.add_edge("Mosaic mottling on leaves", "Tomato___Tomato_mosaic_virus")
G.add_edge("No visible disease", "Tomato___healthy")
# -------------------------------
# Diseases β Treatments
# -------------------------------
# Apple
G.add_edge("Apple___Apple_scab", "Fungicides (captan, myclobutanil), prune infected leaves")
G.add_edge("Apple___Black_rot", "Remove mummified fruits, use fungicides")
G.add_edge("Apple___Cedar_apple_rust", "Remove nearby junipers, apply fungicides")
G.add_edge("Apple___healthy", "No treatment needed")
# Blueberry
G.add_edge("Blueberry___healthy", "No treatment needed")
# Cherry
G.add_edge("Cherry_(including_sour)___Powdery_mildew", "Fungicides: sulfur, myclobutanil")
G.add_edge("Cherry_(including_sour)___healthy", "No treatment needed")
# Corn
G.add_edge("Corn_(maize)___Cercospora_leaf_spot Gray_leaf_spot", "Use resistant hybrids, fungicides")
G.add_edge("Corn_(maize)___Common_rust_", "Resistant hybrids, fungicide if severe")
G.add_edge("Corn_(maize)___Northern_Leaf_Blight", "Crop rotation, resistant varieties, fungicides")
G.add_edge("Corn_(maize)___healthy", "No treatment needed")
# Grape
G.add_edge("Grape___Black_rot", "Fungicides (mancozeb, myclobutanil), prune infected vines")
G.add_edge("Grape___Esca_(Black_Measles)", "Remove infected wood, fungicides not very effective")
G.add_edge("Grape___Leaf_blight_(Isariopsis_Leaf_Spot)", "Remove infected leaves, fungicides")
G.add_edge("Grape___healthy", "No treatment needed")
# Orange
G.add_edge("Orange___Haunglongbing_(Citrus_greening)", "No cure, control psyllid vector, use resistant rootstocks")
# Peach
G.add_edge("Peach___Bacterial_spot", "Copper fungicides, resistant varieties")
G.add_edge("Peach___healthy", "No treatment needed")
# Pepper
G.add_edge("Pepper,_bell___Bacterial_spot", "Use copper-based bactericides, resistant cultivars")
G.add_edge("Pepper,_bell___healthy", "No treatment needed")
# Potato
G.add_edge("Potato___Early_blight", "Use fungicides: Mancozeb, Chlorothalonil")
G.add_edge("Potato___Late_blight", "Copper-based fungicides, resistant varieties")
G.add_edge("Potato___healthy", "No treatment needed")
# Raspberry
G.add_edge("Raspberry___healthy", "No treatment needed")
# Soybean
G.add_edge("Soybean___healthy", "No treatment needed")
# Squash
G.add_edge("Squash___Powdery_mildew", "Sulfur-based fungicides, resistant varieties")
# Strawberry
G.add_edge("Strawberry___Leaf_scorch", "Remove infected leaves, apply fungicides")
G.add_edge("Strawberry___healthy", "No treatment needed")
# Tomato
G.add_edge("Tomato___Bacterial_spot", "Copper sprays, avoid overhead irrigation")
G.add_edge("Tomato___Early_blight", "Fungicides: Chlorothalonil, crop rotation")
G.add_edge("Tomato___Late_blight", "Copper fungicides, remove infected plants")
G.add_edge("Tomato___Leaf_Mold", "Fungicides: Chlorothalonil, improve ventilation")
G.add_edge("Tomato___Septoria_leaf_spot", "Apply fungicides, remove infected leaves")
G.add_edge("Tomato___Spider_mites Two-spotted_spider_mite", "Insecticidal soap, neem oil, predatory mites")
G.add_edge("Tomato___Target_Spot", "Fungicides: Chlorothalonil, Mancozeb")
G.add_edge("Tomato___Tomato_Yellow_Leaf_Curl_Virus", "No cure: use resistant varieties, control whiteflies")
G.add_edge("Tomato___Tomato_mosaic_virus", "Remove infected plants, disinfect tools")
G.add_edge("Tomato___healthy", "No treatment needed")
# -------------------------------
# 2. Map Model Predictions β Symptoms
# -------------------------------
symptom_map = {
"Apple___Apple_scab": "Olive-brown velvety spots on leaves/fruits",
"Apple___Black_rot": "Dark sunken lesions with concentric rings",
"Apple___Cedar_apple_rust": "Orange/yellow spots with black centers",
"Apple___healthy": "No visible disease",
"Blueberry___healthy": "No visible disease",
"Cherry_(including_sour)___Powdery_mildew": "White powdery coating on leaves",
"Cherry_(including_sour)___healthy": "No visible disease",
"Corn_(maize)___Cercospora_leaf_spot Gray_leaf_spot": "Gray/tan lesions with dark borders",
"Corn_(maize)___Common_rust_": "Small reddish-brown pustules on leaves",
"Corn_(maize)___Northern_Leaf_Blight": "Cigar-shaped gray-green lesions on leaves",
"Corn_(maize)___healthy": "No visible disease",
"Grape___Black_rot": "Circular black spots on leaves/fruits",
"Grape___Esca_(Black_Measles)": "Interveinal chlorosis, necrosis (black measles)",
"Grape___Leaf_blight_(Isariopsis_Leaf_Spot)": "Irregular brown spots with yellow halo",
"Grape___healthy": "No visible disease",
"Orange___Haunglongbing_(Citrus_greening)": "Yellow shoots, mottled leaves, misshapen fruits",
"Peach___Bacterial_spot": "Small dark water-soaked spots on leaves",
"Peach___healthy": "No visible disease",
"Pepper,_bell___Bacterial_spot": "Brown lesions with yellow halo",
"Pepper,_bell___healthy": "No visible disease",
"Potato___Early_blight": "Dark concentric spots on leaves",
"Potato___Late_blight": "Large irregular brown/black lesions",
"Potato___healthy": "No visible disease",
"Raspberry___healthy": "No visible disease",
"Soybean___healthy": "No visible disease",
"Squash___Powdery_mildew": "White powdery patches on leaves",
"Strawberry___Leaf_scorch": "Irregular brown leaf margins, scorching",
"Strawberry___healthy": "No visible disease",
"Tomato___Bacterial_spot": "Water-soaked brown spots on leaves",
"Tomato___Early_blight": "Concentric rings, target-like spots",
"Tomato___Late_blight": "Large dark blotches with fuzzy growth",
"Tomato___Leaf_Mold": "Yellow patches on upper leaf, fuzzy underside",
"Tomato___Septoria_leaf_spot": "Small circular dark spots with yellow halo",
"Tomato___Spider_mites Two-spotted_spider_mite": "White/yellow stippling + webbing",
"Tomato___Target_Spot": "Brown/black target-like spots",
"Tomato___Tomato_Yellow_Leaf_Curl_Virus": "Leaf curling + yellow mosaic",
"Tomato___Tomato_mosaic_virus": "Mosaic mottling on leaves",
"Tomato___healthy": "No visible disease"
}
# -------------------------------
# Model Setup
# -------------------------------
class DinoClassifier(nn.Module):
def __init__(self, base_model, num_classes, hidden_size):
super().__init__()
self.base = base_model
self.classifier = nn.Linear(hidden_size, num_classes)
def forward(self, x):
outputs = self.base(x)
pooled = outputs.last_hidden_state.mean(dim=1)
return self.classifier(pooled)
model_name = "facebook/dinov3-vits16-pretrain-lvd1689m"
hf_token = os.environ.get("HF_TOKEN") # Token from environment variable
config = AutoConfig.from_pretrained(model_name, use_auth_token=hf_token)
base_model = AutoModel.from_pretrained(model_name, config=config, use_auth_token=hf_token)
num_classes = 38
model = DinoClassifier(base_model, num_classes, config.hidden_size)
checkpoint = torch.load(MODEL_PATH, map_location="cpu")
model.load_state_dict(checkpoint["model_state_dict"])
class_names = checkpoint["classes"]
model.eval()
# -------------------------------
# Image Preprocessing
# -------------------------------
val_test_tfms = transforms.Compose([
transforms.Resize((224, 224)),
transforms.ToTensor(),
transforms.Normalize(mean=(0.5, 0.5, 0.5),
std=(0.5, 0.5, 0.5)),
])
# -------------------------------
# OpenAI API Setup
# -------------------------------
client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))
# -------------------------------
# Prediction Pipeline (fast)
# -------------------------------
def predict_pipeline(image: Image.Image, get_explanation: bool = True):
# Preprocess
x = val_test_tfms(image).unsqueeze(0)
# Prediction
with torch.no_grad():
logits = model(x)
pred_idx = logits.argmax(1).item()
pred_class = class_names[pred_idx]
# KG Lookup
symptom = symptom_map.get(pred_class, "Unknown")
treatment = list(G.neighbors(pred_class)) if pred_class in G else ["No treatment found"]
# JSON summary (black text)
result_json = {
"Predicted Disease": pred_class,
"Symptom": symptom,
"Treatment": ", ".join(treatment),
}
# GPT explanation (optional, can be skipped to speed up)
if get_explanation:
prompt = f"""
You are an agriculture expert.
Disease: {pred_class}
Symptom: {symptom}
Treatment: {', '.join(treatment)}
Provide a **detailed explanation** in Markdown with sections:
1. What causes it
2. How to control it
3. Prevention methods
"""
try:
response = client.chat.completions.create(
model="gpt-4o-mini",
messages=[{"role": "user", "content": prompt}],
temperature=0.7
)
llm_text = response.choices[0].message.content
except Exception as e:
llm_text = f"**LLM Error:** {str(e)}"
else:
llm_text = "GPT explanation skipped for speed."
return json.dumps(result_json, indent=2), llm_text
# -------------------------------
# Gradio App
# -------------------------------
demo = gr.Interface(
fn=predict_pipeline,
inputs=[gr.Image(type="pil")],
outputs=[
gr.Textbox(label="π Prediction Summary", lines=10, interactive=False),
gr.Markdown(label="π Detailed Explanation")
],
title="π± Plant Disease Detection App",
description="Upload a plant leaf image to detect disease, get symptoms, treatment, and optionally an expert explanation."
)
if __name__ == "__main__":
demo.launch()
|