!pip install -q -U google-generativeai

import tensorflow as tf
import numpy as np
from PIL import Image
from numpy import asarray
import gradio as gr
import google.generativeai as genai
#import cv2

print("Dependency Imported Successfully!")


GOOGLE_API_KEY = "AIzaSyA4pL0voDE0py8q8iXtQNQRYYMx_UdFeLk"
genai.configure(api_key=GOOGLE_API_KEY)
model = genai.GenerativeModel('gemini-pro')


#kindly add model path here (Save model in your drive and copy path and paste it below)
model_path="./Models/"

sugarcane_model = tf.keras.models.load_model(f"{model_path}sugracane.h5")
tomato_model = tf.keras.models.load_model(f"{model_path}Tomato.h5")
corn_model = tf.keras.models.load_model(f"{model_path}Corn.h5")
potato_model = tf.keras.models.load_model(f"{model_path}Potato.h5")
rice_model = tf.keras.models.load_model(f"{model_path}Rice.h5")
wheat_model = tf.keras.models.load_model(f"{model_path}Wheat.h5")


print("Models Imported Successfully!")

pred_class = "Plant"


def predict(model,class_name,img):
  global pred_class
  img = tf.image.resize(img, (256, 256))
  img_array = tf.keras.preprocessing.image.img_to_array(img)
  img_array = tf.expand_dims(img_array, 0)
  predictions = model.predict(img_array)
  predicted_class = class_name[np.argmax(predictions[0])]
  confidence = round(100 * (np.max(predictions[0])), 2)
  pred_class = predicted_class
  return predicted_class,f"{confidence} %"




def generate_output(crop,Leaf_Image):
  if crop=="Sugarcane":
    sugarcane_classes = ['Sugarcane___Bacterial_Blight', 'Sugarcane___Healthy', 'Sugarcane___Red_Rot']
    #sugarcane_model = tf.keras.models.load_model(f"{model_path}sugracane.h5")
    return predict(sugarcane_model,sugarcane_classes,Leaf_Image)

  if crop=="Tomato":
    tomato_classes = ['Tomato___Bacterial_spot','Tomato___Early_blight','Tomato___Late_blight','Tomato___healthy']
    #tomato_model = tf.keras.models.load_model(f"{model_path}Tomato.h5")
    return predict(tomato_model,tomato_classes,Leaf_Image)

  if crop=="Corn":
    corn_classes = ['Corn_(maize)___Cercospora_leaf_spot Gray_leaf_spot','Corn_(maize)___Common_rust_','Corn_(maize)___Northern_Leaf_Blight','Corn_(maize)___healthy']
    #corn_model = tf.keras.models.load_model(f"{model_path}Corn.h5")
    return predict(corn_model,corn_classes,Leaf_Image)

  if crop=="Potato":
    potato_classes = ['Potato___Early_blight', 'Potato___Late_blight', 'Potato___healthy']
    #potato_model = tf.keras.models.load_model(f"{model_path}Potato.h5")
    return predict(potato_model,potato_classes,Leaf_Image)

  if crop=="Rice":
    rice_classes = ['Rice___Brown_Spot', 'Rice___Healthy', 'Rice___Hispa', 'Rice___Leaf_Blast']
    #rice_model = tf.keras.models.load_model(f"{model_path}Rice.h5")
    return predict(rice_model,rice_classes,Leaf_Image)

  if crop=="Wheat":
    wheat_classes = ['Wheat___Brown_Rust', 'Wheat___Healthy', 'Wheat___Yellow_Rust']
    #wheat_model = tf.keras.models.load_model(f"{model_path}Wheat.h5")
    return predict(wheat_model,wheat_classes,Leaf_Image)


def transform_history(history):
    new_history = []
    new_history.append({"parts": [{"text": "You are a Agriculture experts named PlantDoc, that specializes in Plant Diseases"}], "role": "user"})
    new_history.append({"parts": [{"text": "ok"}], "role": "model"})
    print(history)
    for chat in history:
        new_history.append({"parts": [{"text": chat[0]}], "role": "user"})
        new_history.append({"parts": [{"text": chat[1]}], "role": "model"})
    return new_history



def response(message, history):
    global chat
    # The history will be the same as in Gradio, the 'Undo' and 'Clear' buttons will work correctly.
    chat.history = transform_history(history)
    response = chat.send_message(message)
    response.resolve()

    # Each character of the answer is displayed
    for i in range(len(response.text)):
        time.sleep(0.01)
        yield response.text[: i+1]
css = """
#textbox {height: 700px;}
"""

with gr.Blocks(css=css) as demo:
  with gr.Row():
    with gr.Column():
      leaf_img = gr.Image(type="numpy",label="Upload Plant's Leaf Image")
    with gr.Column():
      crop = gr.Radio(["Potato", "Tomato", "Wheat","Rice","Corn","Sugarcane"], label="Crop", info="Please Select a Crop?")
      disease = gr.Textbox(label="Disease Predicted :- ")
      confi = gr.Textbox(label="Level of Condidence:- ")
  with gr.Row():
    with gr.Column(elem_id="textbox"):
      greet_btn = gr.Button("Predict")
      greet_btn.click(fn=generate_output, inputs=[crop,leaf_img], outputs=[disease,confi], api_name="predict")
      gr.ChatInterface(fn=response,examples=["Give me more Info about this disease!", "How to treat this Plant disease!"], title="Talk to PlantDoc(AI Expert).")
if __name__ == "__main__":
  chat = model.start_chat(history=[])
  demo.launch()