import gradio as gr

import keras
from keras.preprocessing.image import img_to_array
import imutils
import cv2
from keras.models import load_model
import numpy as np

# parameters for loading data and images
detection_model_path = 'haarcascade_files/haarcascade_frontalface_default.xml'
emotion_model_path = 'model2/model2_entire_model.h5'

# hyper-parameters for bounding boxes shape
# loading models
face_detection = cv2.CascadeClassifier(detection_model_path)
emotion_classifier = load_model(emotion_model_path, compile=False)
EMOTIONS = ['neutral','happiness','surprise','sadness','anger','disgust','fear','contempt','unknown']


# def predict(frame):

#     frame = imutils.resize(frame, width=300)
#     gray = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)
#     faces = face_detection.detectMultiScale(gray, scaleFactor=1.1,
#                                             minNeighbors=5, minSize=(30, 30),
#                                             flags=cv2.CASCADE_SCALE_IMAGE)

#     frameClone = frame.copy()
#     if len(faces) > 0:
#         faces = sorted(faces, reverse=True,
#                        key=lambda x: (x[2] - x[0]) * (x[3] - x[1]))[0]
#         (fX, fY, fW, fH) = faces
#         # Extract the ROI of the face from the grayscale image, resize it to a fixed 28x28 pixels, and then prepare
#         # the ROI for classification via the CNN
#         roi = gray[fY:fY + fH, fX:fX + fW]
#         roi = cv2.resize(roi, (48, 48))
#         roi = roi.astype("float") / 255.0
#         roi = img_to_array(roi)
#         roi = np.expand_dims(roi, axis=0)

#         preds = emotion_classifier.predict(roi)[0]
#         label = EMOTIONS[preds.argmax()]
#     else:
#         return frameClone, "Can't find your face"

#     probs = {}
#     cv2.putText(frameClone, label, (fX, fY - 10),
#                 cv2.FONT_HERSHEY_DUPLEX, 1, (238, 164, 64), 1)
#     cv2.rectangle(frameClone, (fX, fY), (fX + fW, fY + fH),
#                   (238, 164, 64), 2)

#     for (i, (emotion, prob)) in enumerate(zip(EMOTIONS, preds)):
#         probs[emotion] = float(prob)

#     return frameClone, probs

# inp = gr.components.Image(sources="webcam", label="Your face")
# out = [
#     gr.components.Image(label="Predicted Emotion"),
#     gr.components.Label(num_top_classes=2, label="Top 2 Probabilities")
# ]
# title = "Facial Emotion Recognition"
# description = "How well can this model predict your emotions? Take a picture with your webcam, and it will guess if" \
#               " you are: happy, sad, angry, disgusted, scared, surprised, or neutral."
# thumbnail = "https://raw.githubusercontent.com/gradio-app/hub-emotion-recognition/master/thumbnail.png"

# # gr.Interface(predict, inp, out, capture_session=True, title=title, thumbnail=thumbnail,
# #              description=description).launch(inbrowser=True)
# gr.Interface(fn=predict, inputs=inp, outputs=out, title=title, thumbnail=thumbnail).launch()

######################################################################################################################################################
def predict(frame):

    frame = imutils.resize(frame, width=300)
    gray = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)
    faces = face_detection.detectMultiScale(gray, scaleFactor=1.1,
                                            minNeighbors=5, minSize=(30, 30),
                                            flags=cv2.CASCADE_SCALE_IMAGE)

    frameClone = frame.copy()
    if len(faces) > 0:
        faces = sorted(faces, reverse=True,
                       key=lambda x: (x[2] - x[0]) * (x[3] - x[1]))[0]
        (fX, fY, fW, fH) = faces
        # Extract the ROI of the face from the grayscale image, resize it to a fixed 28x28 pixels, and then prepare
        # the ROI for classification via the CNN
        roi = gray[fY:fY + fH, fX:fX + fW]
        roi = cv2.resize(roi, (48, 48))
        roi = roi.astype("float") / 255.0
        roi = img_to_array(roi)
        roi = np.expand_dims(roi, axis=0)

        preds = emotion_classifier.predict(roi)[0]
        label = EMOTIONS[preds.argmax()]
    else:
        return frameClone, "Can't find your face"

    probs = {}
    cv2.putText(frameClone, label, (fX, fY - 10),
                cv2.FONT_HERSHEY_DUPLEX, 1, (238, 164, 64), 1)
    cv2.rectangle(frameClone, (fX, fY), (fX + fW, fY + fH),
                  (238, 164, 64), 2)

    for (i, (emotion, prob)) in enumerate(zip(EMOTIONS, preds)):
        probs[emotion] = float(prob)

    return frameClone, probs

inp = gr.components.Video(sources="webcam", label="Your face")
out = [
    gr.components.Video(label="Predicted Emotion"),
    gr.components.Label(num_top_classes=2, label="Top 2 Probabilities")
    ]

title = "Facial Emotion Recognition"
description = "How well can this model predict your emotions? Take a picture with your webcam, and it will guess if" \
              " you are: happy, sad, angry, disgusted, scared, surprised, or neutral."
thumbnail = "https://raw.githubusercontent.com/gradio-app/hub-emotion-recognition/master/thumbnail.png"

gr.Interface(fn=predict, inputs=inp, outputs=out, title=title, thumbnail=thumbnail).launch()