import os
import streamlit as st
import tensorflow as tf
from tensorflow import image
from keras import models
import numpy as np
from PIL import Image
import pandas as pd
import logging
import logger


log = logging.getLogger(__name__)

st.title("Rice Disease Classifier")

log.info("Looking for Model...: ")

versions = os.listdir("models")
latest = os.listdir('models/'+ versions[-1])

log.info(f"models/{versions[-1]}/{latest[0]}")

if latest:
    log.info(f"Model found : {latest}")

else:
    versions = ['0.3']
    latest = ['model.h5']

desc = pd.read_csv("files/description.csv")
model  = models.load_model(f"models/{versions[-1]}/{latest[0]}", compile=False)

log.info("Compiling Model!")

model.compile(
        optimizer='adam',
        loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
        metrics=['accuracy']
)

log.info("Model Loaded Succesfully.")

dis = list(desc.disease.values)

def image_classifier(inp):
    inp = image.resize(inp, (256,256))
    inp = np.expand_dims(inp,0)
    pred= model.predict(inp)
    return dis[np.argmax(pred)] , f"Confidence - {round(max(pred[0])*100,2)}%"

def detail(pro):
    x = desc[desc["disease"]==pro]
    return list(x["hindi"])[0], list(x["desc"])[0], list(x["hndesc"])[0], list(x["pre"])[0], list(x["hnpre"])[0]


cho = st.file_uploader("Upload Image From Gallery", type=['png','jpg','jpeg','webp'])


if cho is not None:
    img = Image.open(cho)

st.write("or")
if st.button("Open Camera"):
    cam = st.camera_input("Take image")
    if cam is not None:
        img = Image.open(cam)


if st.button("Detect"):
    col1,col2,col3 = st.columns(3)
    pro, conf = image_classifier(img)
    hin, des, hnd, pre, hnp = detail(pro)

    with col2:
        st.image(img)
        st.write("\n\n")
        st.header(pro)
        st.subheader(f"({hin})")
        st.subheader(conf)
        st.write("\n\n\n\n")

        st.subheader(f"Description :")
        st.write(des)
        st.write("\n\n")
        st.write(hnd)
        st.write("\n\n\n")

        st.subheader(f"Precautions :")
        st.write(pre)
        st.write("\n\n")
        st.write(hnp)