import flask from flask import jsonify, request import requests from flask_cors import CORS,cross_origin from OpenSSL import SSL import boto3 from botocore.exceptions import ClientError import time from datetime import datetime from pytz import timezone import urllib import whisper import whisper_timestamped import moviepy.editor as mp from pydub import AudioSegment from pydub.silence import detect_silence, detect_nonsilent import json import collections from collections import defaultdict import re import string import cv2 import os import numpy as np import ssl import mysql.connector from gaze_tracking import GazeTracking from config import * # context = SSL.Context(SSL.TLSv1_2_METHOD) # context.use_privatekey_file('./crt/privatekey.pem') # context.use_certificate_file('./crt/fullchain.pem') ssl._create_default_https_context = ssl._create_unverified_context app = flask.Flask(__name__) CORS(app) @app.route('/', methods = ['GET']) @cross_origin() def hello(): # return "Hello, Python233!!" # os.environ['CURL_CA_BUNDLE']="" # session = requests.Session() # session.verify = False # HOST = DB_host # USER = DB_user # PASSWORD = DB_password # DATABASE = DB_name try: secret_name = "aws-secret" region_name = "us-east-1" # Create a Secrets Manager client session = boto3.session.Session() client = session.client( service_name='secretsmanager', region_name=region_name ) try: get_secret_value_response = client.get_secret_value( SecretId=secret_name ) get_secret_value_db = client.get_secret_value( SecretId='live-db' ) except ClientError as e: # For a list of exceptions thrown, see # https://docs.aws.amazon.com/secretsmanager/latest/apireference/API_GetSecretValue.html raise e secret = get_secret_value_response['SecretString'] secrt = json.loads(secret) secret_db = get_secret_value_db['SecretString'] secrt_db = json.loads(secret_db) # mydb = mysql.connector.connect( # host=secrt_db['db_host'], # user=secrt_db['db_user'], # password=secrt_db['db_pass'], # database=secrt_db['db_name'] # ) mydb = mysql.connector.connect( host='mip-prod-mysql.c87oewlzrnhl.us-east-1.rds.amazonaws.com', user='mip_prod_mysql', password='cNzymZll*2RTD41u', database='mip_prod_mysql' ) print(mydb) cursor = mydb.cursor() sql_check_query = "SELECT * FROM mip_ai_feedback_queue_tbl WHERE status = '1'" cursor.execute(sql_check_query) records = cursor.fetchall() print("Continue satus rows in table: ", cursor.rowcount) if cursor.rowcount == 0: sql_select_Query = "SELECT * FROM mip_ai_feedback_queue_tbl WHERE status = '0' AND data_process = '1' ORDER BY created_at ASC LIMIT 1" cursor.execute(sql_select_Query) records = cursor.fetchall() print("Total number of rows in table: ", cursor.rowcount) if cursor.rowcount > 0: for row in records: result_db = row[0] # start process time now_utc = datetime.now(timezone('America/Los_Angeles')) start_time = now_utc.strftime("%Y-%m-%d %H:%M:%S") print(start_time) sql_update_Query = "UPDATE mip_ai_feedback_queue_tbl SET status = '1', process_start_time = '"+start_time+"' WHERE id = '"+result_db+"'" cursor.execute(sql_update_Query) mydb.commit() print(cursor.rowcount, "record(s) affected") print(result_db) video_path = row[1] user_id = row[2] interview_id = row[3] question_id = row[4] interview_type = row[5] # video_path = request.args.get('video') # user_id = request.args.get('user_id') # interview_id = request.args.get('interview_id') # question_id = request.args.get('question_id') # interview_type = request.args.get('type') split_name = os.path.splitext(video_path)[0] # access_key_id = AWS_ACCESS_KEY_ID # secret_access_key = AWS_SECRET_ACCESS_KEY # region_name = REGION_NAME try: s3 = boto3.client('s3', aws_access_key_id=secrt['AWS_ACCESS_KEY_ID'], aws_secret_access_key=secrt['AWS_SECRET_ACCESS_KEY'], region_name=REGION_NAME) BUCKET_NAME = 'pub.myinterviewpractice.com' except Exception as e: print(f"Error creating S3 client: {e}") try: url = s3.generate_presigned_url( ClientMethod='get_object', Params={ 'Bucket': BUCKET_NAME, 'Key': 'live/' + video_path } ) if video_path.endswith(".mp3"): aud_file = s3.download_file( Filename="audio_files/"+video_path, Bucket= BUCKET_NAME, Key= 'live/' + video_path ) # print(aud_file) # def transcribe_file(job_name, file_uri, transcribe_client): # transcribe_client.start_transcription_job( # TranscriptionJobName=job_name, # Media={'MediaFileUri': file_uri}, # MediaFormat='mp4', # LanguageCode='en-US' # ) # max_tries = 60 # while max_tries > 0: # max_tries -= 1 # job = transcribe_client.get_transcription_job(TranscriptionJobName=job_name) # job_status = job['TranscriptionJob']['TranscriptionJobStatus'] # if job_status in ['COMPLETED', 'FAILED']: # print(f"Job {job_name} is {job_status}.") # if job_status == 'COMPLETED': # response = urllib.request.urlopen(job['TranscriptionJob']['Transcript']['TranscriptFileUri']) # data = json.loads(response.read()) # text = data['results']['transcripts'][0]['transcript'] # return text # break # else: # print(f"Waiting for {job_name}. Current status is {job_status}.") # time.sleep(10) # file_uri = ('s3://images.myinterviewpractice.com/ai_test/' + video_path) err_audio = 0 er_movie = 0 reslt = '' time_au = 0 audio_vol = 0 total_words = 0 pace = 0 vol = '' puse = 0 arr_ps = [] new_filler = 0 new_total_power_count = 0 ngtv_count = 0 um_count = 0 new_f_ar = [] filler_count = 0 filler_store = [] filler_time_arr = [] filler_word_counts = {} ph_filler_ar = [] pfiller_var = 0 filler_phrse_word = None filer_phrse = '' ph_filler_wrds = {} power_count = 0 power_store = [] power_time_arr = [] power_word_counts = {} ph_power_ar = [] ph_power_var = 0 power_phrse_word = None ph_power_wrds = {} powr_phrse = '' ngtv_count = 0 ngtv_store = [] ngtv_time_arr = [] ngtv_word_counts = {} um_count = 0 um_store = [] um_time_arr = [] um_word_counts = {} um_time = '' ngtv_time = '' filr_time = '' pwr_time = '' total_filler_words = 0 total_power_words = 0 new_p_ar = [] new_n_ar = [] err_eye = 0 err_smle = 0 arr1 = 0 lght = '' cntr = 0 lft = 0 rght = 0 count_s = 0 smile_count = 0 dis_total_eye_percentage = 0 per_eye_contact_center = 0 per_eye_contact_left = 0 per_eye_contact_right = 0 dis_total_smile_percentage = 0 eye_contact_behav = "" try: audio = whisper_timestamped.load_audio(url) model = whisper_timestamped.load_model("base", device="cpu") result = whisper_timestamped.transcribe(model,audio,beam_size=5,best_of=5) reslt = (result) rs = json.dumps(reslt, indent = 1) curr_time = time.localtime() end_time = time.strftime("%H-%M-%S", curr_time) # rs_aws = transcribe_file( split_name + '-job' + end_time, file_uri, transcribe_client) # print("Video CC:",rs_aws) # print(rs) def unique(list1): # insert the list to the set list_set = set(list1) # convert the set to the list unique_list = (list(list_set)) u_ar = [] for x in unique_list: u_ar.append(x) print (u_ar) combined_data = defaultdict(list) for word, value in u_ar: combined_data[word].append(value) combined_data[word].sort() # Convert the dictionary to a list of lists new_ar = [{key : values} for key, values in combined_data.items()] # obg[] print(new_ar) return new_ar def convert_to_preferred_format(sec): sec = sec % (24 * 3600) hour = sec // 3600 sec %= 3600 min = sec // 60 sec %= 60 # print("seconds value in hours:",hour) # print("seconds value in minutes:",min) return "%2d:%02d" % (min, sec) # search filler words with timestamp ld = json.loads(rs) with open(r'lib_file/filler.txt','r') as file: filler_con = file.read() with open(r'lib_file/filler_phrase.txt','r') as file: filler_phrse = file.read() ph_filler = filler_phrse.split(',') with open(r'lib_file/power.txt','r') as file: power_con = file.read() with open(r'lib_file/power_phrase.txt','r') as file: power_phrse = file.read() ph_power = power_phrse.split(',') with open(r'lib_file/negative.txt','r') as file: ngtv_con = file.read() with open(r'lib_file/um_words.txt','r') as file: um_con = file.read() for i in range(0, len(ld['segments'])): txt = ld['segments'][i]['text'] txt_lwr = txt.lower() # print(txt) new_s = txt_lwr.translate(str.maketrans('', '', string.punctuation)) # print(new_s) for word in new_s.split(): # print(word) for w in filler_con.split(','): if word == w: filler_count += 1 filler_store.append(word) if filler_count > 0: # print(filler_count) for j in range(0, len(ld['segments'][i]['words'])): wrd = ld['segments'][i]['words'][j]['text'] wrd_lwr = wrd.lower() new_w = wrd_lwr.translate(str.maketrans('', '', string.punctuation)) if new_w == w: pfc_lwr = (ld['segments'][i]['words'][j]['text']).lower() pfc_txt = pfc_lwr.translate(str.maketrans('', '', string.punctuation)) filr_time = (pfc_txt,(convert_to_preferred_format(ld['segments'][i]['words'][j]['start']))) filler_time_arr.append(filr_time) # print(ld['segments'][i]['words'][j]['text'],ld['segments'][i]['words'][j]['start']) filler_word_counts = collections.Counter(filler_store) new_f_ar = [] list1 = filler_time_arr new_f_ar = unique(list1) print(filler_count) if filler_count == 0: print(filler_count) print('No Filler Words') for pi in range(0, len(ld['segments'])): p_txt = ld['segments'][pi]['text'] p_txt_lwr = p_txt.lower() # print(txt) new_ps = p_txt_lwr.translate(str.maketrans('', '', string.punctuation)) # print(new_s) for pword in new_ps.split(): for p in power_con.split(','): if pword == p: power_count += 1 power_store.append(pword) if power_count > 0: # print(power_count) for k in range(0, len(ld['segments'][pi]['words'])): p_wrd = ld['segments'][pi]['words'][k]['text'] p_wrd_lwr = p_wrd.lower() new_pw = p_wrd_lwr.translate(str.maketrans('', '', string.punctuation)) if new_pw == p: ppc_lwr = (ld['segments'][pi]['words'][k]['text']).lower() ppc_txt = ppc_lwr.translate(str.maketrans('', '', string.punctuation)) pwr_time = (ppc_txt,(convert_to_preferred_format(ld['segments'][pi]['words'][k]['start']))) power_time_arr.append(pwr_time) # print(ld['segments'][pi]['words'][k]['text'],ld['segments'][pi]['words'][k]['start']) power_word_counts = collections.Counter(power_store) new_p_ar = [] list2 = power_time_arr new_p_ar = unique(list2) print(power_count) if power_count == 0: print(power_count) print('No Power Words') for ni in range(0, len(ld['segments'])): n_txt = ld['segments'][ni]['text'] n_txt_lwr = n_txt.lower() # print(txt) new_ns = n_txt_lwr.translate(str.maketrans('', '', string.punctuation)) # print(new_s) for nword in new_ns.split(): for n in ngtv_con.split(','): if nword == n: ngtv_count += 1 ngtv_store.append(nword) if ngtv_count > 0: # print(ngtv_count) for l in range(0, len(ld['segments'][ni]['words'])): n_wrd = ld['segments'][ni]['words'][l]['text'] n_wrd_lwr = n_wrd.lower() new_nw = n_wrd_lwr.translate(str.maketrans('', '', string.punctuation)) if new_nw == n: pnc_lwr = (ld['segments'][ni]['words'][l]['text']).lower() pnc_txt = pnc_lwr.translate(str.maketrans('', '', string.punctuation)) ngtv_time = (pnc_txt,(convert_to_preferred_format(ld['segments'][ni]['words'][l]['start']))) ngtv_time_arr.append(ngtv_time) # print(ld['segments'][i]['words'][l]['text'],ld['segments'][i]['words'][l]['start']) ngtv_word_counts = collections.Counter(ngtv_store) new_n_ar = [] list3 = ngtv_time_arr new_n_ar = unique(list3) print(ngtv_count) if ngtv_count == 0: print(ngtv_count) print('No Negative Words') # um counter analysis for ui in range(0, len(ld['segments'])): u_txt = ld['segments'][ui]['text'] u_txt_lwr = u_txt.lower() # print(txt) new_us = u_txt_lwr.translate(str.maketrans('', '', string.punctuation)) # print(new_s) for uword in new_us.split(): for u in um_con.split(','): if uword == u: um_count += 1 um_store.append(uword) if um_count > 0: # print(um_count) for m in range(0, len(ld['segments'][ui]['words'])): u_wrd = ld['segments'][ui]['words'][m]['text'] u_wrd_lwr = u_wrd.lower() new_uw = u_wrd_lwr.translate(str.maketrans('', '', string.punctuation)) if new_uw == u: um_time = (ld['segments'][ui]['words'][m]['text'],ld['segments'][ui]['words'][m]['start']) um_time_arr.append(um_time) um_word_counts = collections.Counter(um_store) new_u_ar = [] list4 = um_time_arr new_u_ar = unique(list4) print(um_count) if um_count == 0: print(um_count) print('No Um Words') # --------- # um analysis using aws transcript # u_txt = rs_aws # u_txt_lwr = u_txt.lower() # # print(txt) # um_time_arr = [] # new_us = u_txt_lwr.translate(str.maketrans('', '', string.punctuation)) # for uword in new_us.split(): # for u in um_con.split(','): # if uword == u: # um_count += 1 # um_store.append(uword) # # print(store) # um_word_counts = collections.Counter(um_store) # if um_count > 0 : # print(um_word_counts) # # print(filler_count + pfiller_var) # # print(filer_phrse) # else: # print(um_word_counts) # print('No Um Words') print("Filler Count:",filler_word_counts) print("Power Count:",power_word_counts) print("Negative Count:",ngtv_word_counts) print("Um Count:",um_word_counts) reslt = (ld['text']) print(reslt) rslt_lwr = reslt.lower() ph_var = '' f_count = 0 f_sum = 0 ph_pow_var = '' p_count = 0 p_sum = 0 # filler phrases detection for a in range(0, len(ph_filler)): new_ph = re.search(ph_filler[a], rslt_lwr) try: filler_phrse_word = new_ph.group() f_count = len(re.findall(filler_phrse_word, rslt_lwr)) except AttributeError: continue # print(filler_phrse_word) pfiller_var += 1 ph_var = filler_phrse_word.strip() # ph_var = (filler_phrse_word, f_count) data_f={} data_f[filler_phrse_word]=f_count print(data_f) ph_filler_ar.append(data_f) # filer_phrse = collections.Counter(ph_filler_ar) f_sum = f_sum + f_count if pfiller_var > 0: print(ph_filler_ar) print('Total filler Phrase:', f_sum) ph_filler_wrds = ph_filler_ar total_filler_words = pfiller_var else: total_filler_words = 'No Filler Phrases' # power phrases detection for b in range(0, len(ph_power)): pow_ph = re.search(ph_power[b], rslt_lwr) try: power_phrse_word = pow_ph.group() p_count = len(re.findall(power_phrse_word, rslt_lwr)) except AttributeError: continue # print(power_phrse_word) ph_power_var += 1 ph_pow_var = power_phrse_word.strip() # ph_pow_var = (power_phrse_word, p_count) data_p={} data_p[ph_pow_var]=p_count print(data_p) ph_power_ar.append(data_p) # powr_phrse = collections.Counter(ph_power_ar) p_sum = p_sum + p_count if ph_power_var > 0: # print(power_phrse) print('Total Power Phrase:', p_sum) ph_power_wrds = ph_power_ar total_power_words = ph_power_var else: total_power_words = 'No Power Phrases' new_total_filler_count = f_sum + filler_count print("New Total Filler:",new_total_filler_count) new_total_power_count = p_sum + power_count print("New Total Power:",new_total_power_count) # analysis of CC total_words = len(reslt.split()) print("Total words spoken:",total_words) new_filler_count = 0 new_filler = 0 # in 100 words if total_words > 0: new_filler_count = ((new_total_filler_count/total_words)*100) new_filler = (round(new_filler_count,1)) print(new_filler) except Exception as e: # error counting err_audio = err_audio + 1 print("Error in audio analysis:", err_audio) if video_path.endswith(".mp3"): # total_words = len(reslt.split()) # print("Total words spoken:",total_words) # audio analysis file_path = "audio_files/"+video_path audio_format = "mp3" audio = AudioSegment.from_file(file_path, format = audio_format) n = len(audio) tme = (n/1000) time_dec = ((n/1000)/60) # time_min = (str(round(time_dec, 2)) + ' Min') time_au = convert_to_preferred_format(tme) print("Audio Length:",time_au) pace_decimal = (total_words / time_dec) pace = (str(round(pace_decimal , 2))) print("Pace of Speech:",pace) # volume analysis print('Volume in dBFS:',audio.dBFS) audio_vol = (str(round(audio.dBFS, 2))) if audio.dBFS < -60: audio_vol = -100 if audio.dBFS > -10: vol = "loud" print(vol) elif audio.dBFS < -10 and audio.dBFS > -30: vol = "Normal" print(vol) elif audio.dBFS < -30: vol = "Low" print(vol) # pause counter audio_pause = [] chunk_ar = [] arr_ps = [] audio_pause_arr = detect_silence(audio, min_silence_len=3200, silence_thresh=-40) for chunks in audio_pause_arr: chunk_ar = ( [convert_to_preferred_format(chunk/1000) for chunk in chunks]) # print(chunk_ar) audio_pause.append(chunk_ar) for key, value in audio_pause: # print(key) # print(value) arr_ps.append(key) if not chunk_ar: print("Pause timimg array:", arr_ps) else: print("Pause timimg array bef-pop:", arr_ps) if arr_ps[0] == ' 0:00': arr_ps.pop(0) # print('yes') print("Pause timimg array pop:", arr_ps) puse = len(arr_ps) print("pause:", puse) os.remove('audio_files/'+video_path) result_text = {"video_cc": reslt, "audio_length": time_au, "audio_volume_dBFS": audio_vol, "total_words_spoken": total_words, "pace_of_speech": pace, "audio_volume": vol, "pause": puse, "pause_timestamp": arr_ps, "filler_word_count": new_filler, "power_word_count": new_total_power_count, "negative_word_count": ngtv_count, "um_word_count": um_count, "filler_word_timestamp": new_f_ar, "filler_count": filler_word_counts, "filler_phrases" :ph_filler_wrds, "filler_phrase_counter" :total_filler_words, "power_phrases" :ph_power_wrds, "power_phrase_counter" :total_power_words, "power_word_timestamp": new_p_ar, "power_count": power_word_counts, "negative_word_timestamp": new_n_ar, "negative_count": ngtv_word_counts, "um_word_timestamp": um_time_arr, "um_count": um_word_counts} # dm_ar = [] # dm_ar.append(result_text) # print(dm_ar) new_rsltxt = json.dumps(result_text) mycursor = mydb.cursor() sql = "INSERT INTO mip_ai_generate_feedback_data_tbl (user_id,interview_id,question_id,interview_type,content) VALUES (%s, %s, %s, %s, %s)" val = (user_id, interview_id, question_id, interview_type, new_rsltxt) else: # converting video to audio try: clip = mp.VideoFileClip(r"" + url) clip.audio.write_audiofile(r"audio_files/"+ split_name +".wav") print("Finished the conversion into audio...") # generating CC # model = whisper.load_model("base") # result = model.transcribe("/home/staging/dashboard-staging.myinterviewpractice.com/demo_ai/uploads/" + video_path , fp16=False) # audio analysis file_path = "audio_files/"+ split_name +".wav" audio_format = "wav" audio = AudioSegment.from_file(file_path , format = audio_format) n = len(audio) tme = (n/1000) time_dec = ((n/1000)/60) # time_min = (str(round(time_dec, 2)) + ' Min') time_au = convert_to_preferred_format(tme) print("Audio Length:",time_au) pace_decimal = (total_words / time_dec) pace = (str(round(pace_decimal , 2))) print("Pace of Speech:",pace) # volume analysis print('Volume in dBFS:',audio.dBFS) audio_vol = (str(round(audio.dBFS, 2))) if audio.dBFS < -60: audio_vol = -100 if audio.dBFS > -10: vol = "loud" print(vol) elif audio.dBFS < -10 and audio.dBFS > -30: vol = "Normal" print(vol) elif audio.dBFS < -30: vol = "Low" print(vol) # pause counter audio_pause = [] chunk_ar = [] arr_ps = [] audio_pause_arr = detect_silence(audio, min_silence_len=3200, silence_thresh=-40) for chunks in audio_pause_arr: chunk_ar = ([convert_to_preferred_format(chunk/1000) for chunk in chunks]) # print(convert_to_preferred_format(chunk_ar)) audio_pause.append(chunk_ar) for key, value in audio_pause: # print(key) # print(value) arr_ps.append(key) if not chunk_ar: print("Pause timimg array:", arr_ps) else: print("Pause timimg array bef-pop:", arr_ps) if arr_ps[0] == ' 0:00': arr_ps.pop(0) # print('yes') print("Pause timimg array pop:", arr_ps) puse = len(arr_ps) print("pause:", puse) os.remove("audio_files/"+ split_name +".wav") except BaseException as e: print(f"MoviePyError: {e}") er_movie = er_movie + 1 print("Moviepy Error Detected:", er_movie) # video light analysis face_casecade=cv2.CascadeClassifier('frontface_default.xml') smile_cascade=cv2.CascadeClassifier('haarcascade_smile.xml') cntr = 0 lft = 0 rght = 0 count_s = 0 cap = cv2.VideoCapture(url) gaze = GazeTracking() fps = cap.get(cv2.CAP_PROP_FPS) print(f"{fps} frames per second") total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT)) if total_frames > 0: print(total_frames) for ft in range(0, total_frames): ret,frame = cap.read() try: gaze.refresh(frame) new_frame = gaze.annotated_frame() if gaze.is_right(): rght = rght + 1 elif gaze.is_left(): lft = lft + 1 elif gaze.is_center(): cntr = cntr + 1 except Exception as e: err_eye = err_eye + 1 # print(f"Error during eye detection: {e}") try: gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) face = face_casecade.detectMultiScale(gray, scaleFactor=1.3, minNeighbors=5) for (x,y,w,h) in face: cv2.rectangle(frame,(x,y),(x+w,y+h),(0,255,0),2) roi_gray = gray[y:y+h, x:x+w] roi_img = frame[y:y+h, x:x+w] smile = smile_cascade.detectMultiScale(roi_gray, scaleFactor= 1.6, minNeighbors=20) for (x,y,w,h) in smile: cv2.rectangle(roi_img,(x,y),(x+w,y+h),(0,0,255),2) count_s = count_s+1 # c_ar.append(count) except Exception as e: err_smle = err_smle + 1 # print(f"Error during smile detection: {e}") smile_count = (count_s) print("Right",rght) print("Left",lft) print("Center",cntr) print("Smile",smile_count) print("Error during eye detection:",err_eye ) print("Error during smile detection:",err_smle ) print("Error during audio detection:",err_audio ) # percentage calculation of eye tracking total_eye_percentage = (((cntr + lft + rght) / total_frames) * 100) dis_total_eye_percentage = (str(round(total_eye_percentage, 2))) print(dis_total_eye_percentage) if(total_eye_percentage > 80): eye_cap_note = "Good eye captured by the camera." print(eye_cap_note) elif(total_eye_percentage < 80 and total_eye_percentage > 20): eye_cap_note = "Normal eye captured by the camera." print(eye_cap_note) elif(total_eye_percentage < 20): eye_cap_note = "Eye not properly captured by the camera." print(eye_cap_note) center_per_eye = ((cntr / total_frames) * 100) left_per_eye = ((lft / total_frames) * 100) right_per_eye = ((rght / total_frames) * 100) per_eye_contact_center = (str(round(center_per_eye, 2))) per_eye_contact_left = (str(round(left_per_eye, 2))) per_eye_contact_right = (str(round(right_per_eye, 2))) print(per_eye_contact_center) print(per_eye_contact_left) print(per_eye_contact_right) eye_contact_behav = "" if(center_per_eye > left_per_eye or right_per_eye): eye_contact_behav = "Good eye contact [based on the Total Eye Capture]" print(eye_contact_behav) elif(left_per_eye > center_per_eye or right_per_eye): eye_contact_behav = "Leftside eye contact [based on the Total Eye Capture]" print(eye_contact_behav) elif(right_per_eye > center_per_eye or left_per_eye): eye_contact_behav = "Rightside eye contact [based on the Total Eye Capture]" print(eye_contact_behav) # percentage calculation of smiling total_smile_percentage = (((smile_count) / total_frames) * 100) dis_total_smile_percentage = (str(round(total_smile_percentage, 2))) print(dis_total_smile_percentage) count = [] lght = '' arr1 = 0 for fp in range(0, total_frames): # ret,frame = cap.read() hsvImage = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV) h,s,v = cv2.split(hsvImage) live_value = cv2.mean(v) count.append(live_value[0]) arr = np.array(count) arr_decimal = np.mean(arr) arr1 = round(arr_decimal, 2) # print(arr1) if arr1 > (150): lght = "High Light Video" # print("High Light Video") elif arr1 < (150) and arr1 > (50): lght = "Normal Light Video" # print("Normal Light Video") elif arr1 < (50): lght = "Low Light Video" # print("Low Light Video") else: print("Video Frames not detected.") result_text = {"video_cc": reslt, "audio_length": time_au, "audio_volume_dBFS": audio_vol, "total_words_spoken": total_words, "pace_of_speech": pace, "audio_volume": vol, "pause": puse, "pause_timestamp": arr_ps, "video_light_value": arr1, "video_light": lght, "filler_word_count": new_filler, "power_word_count": new_total_power_count, "negative_word_count": ngtv_count, "um_word_count": um_count, "filler_word_timestamp": new_f_ar, "filler_count": filler_word_counts, "filler_phrases" :ph_filler_wrds, "filler_phrase_counter" :total_filler_words, "power_phrases" :ph_power_wrds, "power_phrase_counter" :total_power_words, "power_word_timestamp": new_p_ar, "power_count": power_word_counts, "negative_word_timestamp": new_n_ar, "negative_count": ngtv_word_counts, "um_word_timestamp": um_time_arr, "um_count": um_word_counts, "total_video_frames": total_frames, "eye_contact_leftside": lft, "eye_contact_rightside": rght, "eye_contact_center": cntr, "smiling_frames": smile_count, "total_eye_capture": dis_total_eye_percentage, "percentage_eye_contact_center": per_eye_contact_center, "percentage_eye_contact_left": per_eye_contact_left, "percentage_eye_contact_right": per_eye_contact_right, "eye_contact_behavior": eye_contact_behav, "total_smile_capture": dis_total_smile_percentage} # dm_ar = [] # dm_ar.append(result_text) # print(dm_ar) new_rsltxt = json.dumps(result_text) mycursor = mydb.cursor() usr_id = str(user_id) intrvw_id = str(interview_id) qsn_id = str(question_id) sql_search_Query = "SELECT * FROM mip_ai_generate_feedback_data_tbl WHERE user_id = '"+usr_id+"' AND interview_id = '"+intrvw_id+"' AND question_id = '"+qsn_id+"' ORDER BY id DESC LIMIT 1" mycursor.execute(sql_search_Query) records = mycursor.fetchall() # print(records) if mycursor.rowcount > 0: new_id = str(records[0][0]) print("Generate result id",new_id) # upChksql = 'UPDATE mip_ai_generate_feedback_data_tbl SET content = "'+new_rsltxt+'" WHERE id = "'+new_id+'"' delSql = "DELETE FROM mip_ai_generate_feedback_data_tbl WHERE id = '"+new_id+"'" mycursor.execute(delSql) print(mycursor.rowcount, "record deleted.") nw_sql = "INSERT INTO mip_ai_generate_feedback_data_tbl (user_id,interview_id,question_id,interview_type,content) VALUES (%s, %s, %s, %s, %s)" nw_val = (user_id, interview_id, question_id, interview_type, new_rsltxt) mycursor.execute(nw_sql, nw_val) mydb.commit() print(mycursor.rowcount, "record inserted.") else: sql = "INSERT INTO mip_ai_generate_feedback_data_tbl (user_id,interview_id,question_id,interview_type,content) VALUES (%s, %s, %s, %s, %s)" val = (user_id, interview_id, question_id, interview_type, new_rsltxt) mycursor.execute(sql, val) mydb.commit() print(mycursor.rowcount, "record inserted.") # end process time now_utc = datetime.now(timezone('America/Los_Angeles')) end_time = now_utc.strftime("%Y-%m-%d %H:%M:%S") print(end_time) str_time = str(time_au) if err_audio > 0: errMsg = (f"Audio error found.{err_audio}") print(errMsg) upErsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '3', process_end_time = '"+end_time+"', duration = '"+str_time+"', exception = '"+errMsg+"' WHERE id = '"+result_db+"'" cursor.execute(upErsql) mydb.commit() print(cursor.rowcount, "record(s) updated for audio") if er_movie > 0: errMsg = (f"Video error found.{er_movie}") print(errMsg) upErsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '3', process_end_time = '"+end_time+"', duration = '"+str_time+"', exception = '"+errMsg+"' WHERE id = '"+result_db+"'" cursor.execute(upErsql) mydb.commit() print(cursor.rowcount, "record(s) updated for Video") if err_eye > 0: errMsg = (f"Eye-detection error found. {err_eye} frames") upErsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '3', process_end_time = '"+end_time+"', duration = '"+str_time+"', exception = '"+errMsg+"' WHERE id = '"+result_db+"'" cursor.execute(upErsql) mydb.commit() print(cursor.rowcount, "record(s) updated") if err_smle > 0: errMsg = (f"Smile-detection error found. {err_smle} frames") upErsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '3', process_end_time = '"+end_time+"', duration = '"+str_time+"', exception = '"+errMsg+"' WHERE id = '"+result_db+"'" cursor.execute(upErsql) mydb.commit() print(cursor.rowcount, "record(s) updated") if ((err_audio > 0 and err_eye > 0) or (err_eye > 0 and err_smle > 0) or (err_audio > 0 and err_smle > 0) or (err_audio > 0 and err_eye > 0 and err_smle > 0)): errMsg = ("Multiple errors found. Please check the video") upErsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '3', process_end_time = '"+end_time+"', duration = '"+str_time+"', exception = '"+errMsg+"' WHERE id = '"+result_db+"'" cursor.execute(upErsql) mydb.commit() print(cursor.rowcount, "record(s) updated") # delsql = "DELETE FROM mip_ai_feedback_queue_tbl WHERE id='"+result_db+"'" upsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '3', process_end_time = '"+end_time+"', duration = '"+str_time+"' WHERE id = '"+result_db+"'" cursor.execute(upsql) mydb.commit() print(cursor.rowcount, "record(s) updated") int_id = str(interview_id) int_ty = str(interview_type) sql_check_Query = "SELECT * FROM mip_ai_feedback_queue_tbl WHERE status IN ('0','1') AND interview_id = '"+int_id+"' AND interview_type = '"+interview_type+"' ORDER BY created_at" # sql_check_Query = "SELECT * FROM mip_ai_feedback_queue_tbl WHERE interview_id='"+int_id+"' ORDER BY created_at" cursor.execute(sql_check_Query) records = cursor.fetchall() print("remaining rows in table: ", cursor.rowcount) if cursor.rowcount == 0: if interview_type == '1': sql_reupdate_Query = "UPDATE mip_interview_tbl SET ai_feedback = '2' WHERE id = '"+int_id+"'" cursor.execute(sql_reupdate_Query) mydb.commit() print(cursor.rowcount, "record(s) affected") else: sql_reupdate_Query = "UPDATE mip_giving_custom_interview_tbl SET ai_feedback = '2' WHERE id = '"+int_id+"'" cursor.execute(sql_reupdate_Query) mydb.commit() print(cursor.rowcount, "record(s) affected") url = 'https://myinterviewpractice.com/ajax/api/ai_feedbak_notify.php' payload = 'interview_id='+int_id+'&type='+int_ty headers = { 'Content-Type': 'application/x-www-form-urlencoded', } response = requests.request("POST", url, headers=headers, data=payload) print(response.text) # print(response.json()) return jsonify(result_text) except Exception as e: # Catch any other exceptions not handled above ex_err = (f"An error occurred: {e}") gt_err = ("Error: An unexpected error occurred.") print(ex_err) result_errtext = {"error": 'Video is not loaded properly.'} new_errsltxt = json.dumps(result_errtext) upexsql = "UPDATE mip_ai_feedback_queue_tbl SET status = '2', exception = '"+gt_err+"', process_start_time = NULL WHERE id = '"+result_db+"'" cursor.execute(upexsql) mydb.commit() print(cursor.rowcount, "record(s) updated") usr_id = str(user_id) intrvw_id = str(interview_id) qsn_id = str(question_id) sql_search_Query = "SELECT * FROM mip_ai_generate_feedback_data_tbl WHERE user_id = '"+usr_id+"' AND interview_id = '"+intrvw_id+"' AND question_id = '"+qsn_id+"' ORDER BY id DESC LIMIT 1" cursor.execute(sql_search_Query) records = cursor.fetchall() # print(records) if cursor.rowcount > 0: new_id = str(records[0][0]) print("Generate result id",new_id) # upChksql = 'UPDATE mip_ai_generate_feedback_data_tbl SET content = "'+new_rsltxt+'" WHERE id = "'+new_id+'"' delSql = "DELETE FROM mip_ai_generate_feedback_data_tbl WHERE id = '"+new_id+"'" cursor.execute(delSql) print(cursor.rowcount, "record deleted.") ersql = "INSERT INTO mip_ai_generate_feedback_data_tbl (user_id,interview_id,question_id,interview_type,content) VALUES (%s, %s, %s, %s, %s)" erval = (user_id, interview_id, question_id, interview_type, new_errsltxt) cursor.execute(ersql, erval) mydb.commit() print(cursor.rowcount, "record inserted.") else: er_nwsql = "INSERT INTO mip_ai_generate_feedback_data_tbl (user_id,interview_id,question_id,interview_type,content) VALUES (%s, %s, %s, %s, %s)" er_nwval = (user_id, interview_id, question_id, interview_type, new_errsltxt) cursor.execute(er_nwsql, er_nwval) mydb.commit() print(cursor.rowcount, "record inserted.") int_id = str(interview_id) int_ty = str(interview_type) sql_check_Query = "SELECT * FROM mip_ai_feedback_queue_tbl WHERE status IN ('0','1') AND interview_id = '"+int_id+"' AND interview_type = '"+interview_type+"' ORDER BY created_at" # sql_check_Query = "SELECT * FROM mip_ai_feedback_queue_tbl WHERE interview_id='"+int_id+"' ORDER BY created_at" cursor.execute(sql_check_Query) records = cursor.fetchall() print("remaining rows in table: ", cursor.rowcount) if cursor.rowcount == 0: if interview_type == '1': sql_reupdate_Query = "UPDATE mip_interview_tbl SET ai_feedback = '2' WHERE id = '"+int_id+"'" cursor.execute(sql_reupdate_Query) mydb.commit() print(cursor.rowcount, "record(s) affected") else: sql_reupdate_Query = "UPDATE mip_giving_custom_interview_tbl SET ai_feedback = '2' WHERE id = '"+int_id+"'" cursor.execute(sql_reupdate_Query) mydb.commit() print(cursor.rowcount, "record(s) affected") url = 'https://staging.myinterviewpractice.com/ajax/api/ai_feedbak_notify.php' payload = 'interview_id='+int_id+'&type='+int_ty headers = { 'Content-Type': 'application/x-www-form-urlencoded', } response = requests.request("POST", url, headers=headers, data=payload) print(response.text) else: err = "No record found" print(err) return (err) else: err1 = "Queue is in progress" print(err1) time_data = 0 for row in records: time_data = row[10] result_db = row[0] new_tme = (str(time_data)) print('new time:',new_tme) now_utc1 = datetime.now(timezone('America/Los_Angeles')) cur_time = now_utc1.strftime("%Y-%m-%d %H:%M:%S") print("current time:", cur_time) spend_time = ((datetime.strptime(cur_time, "%Y-%m-%d %H:%M:%S")) - (datetime.strptime(new_tme, "%Y-%m-%d %H:%M:%S"))) hours = spend_time.total_seconds() // 3600 minutes = (spend_time.total_seconds() % 3600) // 60 seconds = spend_time.total_seconds() % 60 print(f"The difference is {hours} hours, {minutes} minutes, and {seconds} seconds.") if hours >= 2: print("Processing time is exceed 2 hours.") for nrow in records: video_name = nrow[1] user_id = nrow[2] interview_id = nrow[3] question_id = nrow[4] interview_type = nrow[5] data_process = nrow[7] tries = nrow[8] new_resultid = round(time.time() * 1000) now_utc = datetime.now(timezone('America/Los_Angeles')) create_time = now_utc.strftime("%Y-%m-%d %H:%M:%S") created_at = create_time mycursor = mydb.cursor() inssql = "INSERT INTO mip_ai_feedback_queue_tbl (id,video_name,user_id,interview_id,question_id,interview_type,status,data_process,tries,created_at,process_start_time,process_end_time,duration,exception) VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s)" val = (new_resultid, video_name, user_id, interview_id, question_id, interview_type, '0', data_process, tries, created_at, None, None, None, None) mycursor.execute(inssql, val) mydb.commit() print(mycursor.rowcount, "record inserted.") tmeout_err = "Processing time is exceed 2 hours" uptmesql = "UPDATE mip_ai_feedback_queue_tbl SET status = '2', exception = '"+tmeout_err+"' WHERE id = '"+result_db+"'" cursor.execute(uptmesql) mydb.commit() print(cursor.rowcount, "record(s) updated") else: print("Program is processing...") return (err1) except mysql.connector.Error as e: print("Error reading data from MySQL table", e) if __name__ == '__main__': context = ('/etc/letsencrypt/live/ai-feedback.myinterviewpractice.com/fullchain.pem', '/etc/letsencrypt/live/ai-feedback.myinterviewpractice.com/privkey.pem') # Set the hostname and port number hostname = "0.0.0.0" port = 9001 # Start the Flask app # app.run(debug=True, port=port) # app.run(host=hostname, port=port, ssl_context='adhoc') app.run(host=hostname, port=port, ssl_context=context, threaded=True, debug=False)