Merge pull request #7 from vjbytes102/master

Integration of NLP feature calculation

Dockerfile

@@ -12,7 +12,9 @@ RUN apt-get update && apt-get install -y python3-pip \
                    && apt-get install -y libavcodec-dev \
                    && apt-get install -y libavformat-dev \
                    && apt-get install -y libavdevice-dev \
-                   && apt-get install -y libboost-all-dev
+                   && apt-get install -y libboost-all-dev \
+                   && apt-get install -y git \
+                   && apt-get install -y sox
 RUN ln -sfn /usr/bin/pip3 /usr/bin/pip
 
 COPY . /app
@@ -24,8 +26,15 @@ RUN dpkg --configure -a
 RUN su -c ./install.sh
 RUN echo "Done OpenFace!"
 
-WORKDIR /app
+RUN echo "Cloning DeepSpeech..."
+WORKDIR /app/pkg
+RUN git clone https://github.com/mozilla/DeepSpeech.git
 
+WORKDIR /app/pkg/DeepSpeech
+RUN wget https://github.com/mozilla/DeepSpeech/releases/download/v0.9.1/deepspeech-0.9.1-models.pbmm
+RUN wget https://github.com/mozilla/DeepSpeech/releases/download/v0.9.1/deepspeech-0.9.1-models.scorer
+
+WORKDIR /app
 RUN pip install --upgrade pip
 RUN pip install -r requirements.txt
 RUN echo "Requirement txt done!"

dbm_lib/config/config_raw_feature.py

@@ -222,4 +222,32 @@ class ConfigRawReader(object):
             self.mov_Hpose_Yaw = config['raw_feature']['mov_Hpose_Yaw']
             self.mov_Hpose_Roll = config['raw_feature']['mov_Hpose_Roll']
             self.mov_Hpose_Dist = config['raw_feature']['mov_Hpose_Dist']
+            self.mov_leye_x = config['raw_feature']['mov_leye_x']
+            self.mov_leye_y = config['raw_feature']['mov_leye_y']
+            self.mov_leye_z = config['raw_feature']['mov_leye_z']
+            self.mov_reye_x = config['raw_feature']['mov_reye_x']
+            self.mov_reye_y = config['raw_feature']['mov_reye_y']
+            self.mov_reye_z = config['raw_feature']['mov_reye_z']
+            self.mov_eleft_disp = config['raw_feature']['mov_eleft_disp']
+            self.mov_eright_disp = config['raw_feature']['mov_eright_disp']
+
+            #NLP features
+            self.nlp_transcribe = config['raw_feature']['nlp_transcribe']
+            self.nlp_numSentences = config['raw_feature']['nlp_numSentences']
+            self.nlp_singPronPerAns = config['raw_feature']['nlp_singPronPerAns']
+            self.nlp_singPronPerSen = config['raw_feature']['nlp_singPronPerSen']
+            self.nlp_pastTensePerAns = config['raw_feature']['nlp_pastTensePerAns']
+            self.nlp_pastTensePerSen = config['raw_feature']['nlp_pastTensePerSen']
+            self.nlp_pronounsPerAns = config['raw_feature']['nlp_pronounsPerAns']
+            self.nlp_pronounsPerSen = config['raw_feature']['nlp_pronounsPerSen']
+            self.nlp_verbsPerAns = config['raw_feature']['nlp_verbsPerAns']
+            self.nlp_verbsPerSen = config['raw_feature']['nlp_verbsPerSen']
+            self.nlp_adjectivesPerAns = config['raw_feature']['nlp_adjectivesPerAns']
+            self.nlp_adjectivesPerSen = config['raw_feature']['nlp_adjectivesPerSen']
+            self.nlp_nounsPerAns = config['raw_feature']['nlp_nounsPerAns']
+            self.nlp_nounsPerSen = config['raw_feature']['nlp_nounsPerSen']
+            self.nlp_sentiment_mean = config['raw_feature']['nlp_sentiment_mean']
+            self.nlp_mattr = config['raw_feature']['nlp_mattr']
+            self.nlp_wordsPerMin = config['raw_feature']['nlp_wordsPerMin']
+            self.nlp_totalTime = config['raw_feature']['nlp_totalTime']
             

dbm_lib/controller/process_feature.py

@@ -7,7 +7,8 @@ created: 2020-20-07
 from dbm_lib.dbm_features.raw_features.audio import intensity, pitch_freq, hnr, gne, voice_frame_score, formant_freq
 from dbm_lib.dbm_features.raw_features.audio import pause_segment, jitter, shimmer, mfcc
 from dbm_lib.dbm_features.raw_features.video import face_asymmetry, face_au, face_emotion_expressivity, face_landmark
-from dbm_lib.dbm_features.raw_features.movement import head_motion, eye_blink
+from dbm_lib.dbm_features.raw_features.movement import head_motion, eye_blink, eye_gaze
+from dbm_lib.dbm_features.raw_features.nlp import transcribe, speech_features
 
 import subprocess
 import logging
@@ -122,6 +123,24 @@ def process_movement(video_uri, out_dir, dbm_group, r_config, dlib_model):
     
     logger.info('processing eye blink....')
     eye_blink.run_eye_blink(video_uri, out_dir, r_config, dlib_model)
+
+    logger.info('processing eye gaze....')
+    eye_gaze.run_eye_gaze(video_uri, out_dir, r_config)
+    
+def process_nlp(video_uri, out_dir, dbm_group, r_config, deep_path):
+    """
+    processing nlp features
+    Args:
+        video_uri: video path; out_dir: raw variable output dir
+        dbm_group: list of features to process; r_config: raw feature config object
+        deep_path: deep speech build path
+    """
+    if dbm_group != None and len(dbm_group)>0 and 'nlp' not in dbm_group:
+        return
+    
+    logger.info('Processing nlp variables from data in {}'.format(video_uri))
+    transcribe.run_transcribe(video_uri, out_dir, r_config, deep_path)
+    speech_features.run_speech_feature(video_uri, out_dir, r_config)
     
 def remove_file(file_path):
     """

dbm_lib/dbm_features/raw_features/movement/eye_gaze.py

@@ -0,0 +1,148 @@
+"""
+file_name: eye_gaze
+project_name: DBM
+created: 2020-30-11
+"""
+
+import os
+import glob
+import pandas as pd
+import numpy as np
+from scipy.spatial import distance
+from os.path import join
+import logging
+
+from dbm_lib.dbm_features.raw_features.util import util as ut
+
+logging.basicConfig(level=logging.INFO)
+logger=logging.getLogger()
+
+eye_pose_dir = 'movement/gaze'
+eye_pose_ext = '_eyegaze.csv'
+
+def eye_motion_df(l_disp, r_disp, error_list, r_config):
+    """
+    Generating eye movement dataframe
+
+    Args:
+        l_disp: displacement list(left eye); l_disp: displacement list(right eye)
+        r_config: raw variable config file object
+
+    Reutrns:
+        Final eye displacement dataframe
+    """
+    df_eye_left = pd.DataFrame(l_disp, columns=[r_config.mov_eleft_disp])
+    df_eye_right = pd.DataFrame(r_disp, columns=[r_config.mov_eright_disp])
+    
+    df_eye_motion = pd.concat([df_eye_left, df_eye_right], axis=1, sort=False)
+    df_eye_motion[r_config.err_reason] = error_list
+    return df_eye_motion
+
+def filter_motion(df_of, df_disp, col_l, col_r, r_config):
+    """
+    Filtering final eye movement dataframe
+    
+    Args:
+        df_of: Openface raw out dataframe; col_r: right eye column
+        col_l: left eye column; r_config: raw variable config file object
+    """
+    
+    df_of = df_of[col_l + col_r + [' confidence']]
+    df_of.loc[(df_of[' confidence'].astype(float) < 0.8), col_l + col_r] = np.nan
+    
+    df_filter =  df_of[col_l + col_r]
+    df_filter.columns = [r_config.mov_leye_x, r_config.mov_leye_y, r_config.mov_leye_z,
+                      r_config.mov_reye_x, r_config.mov_reye_y, r_config.mov_reye_z]
+    
+    df_motion = pd.concat([df_filter, df_disp], axis=1, sort=False)
+    return df_motion
+
+def eye_disp(of_results, col, r_config):
+    """
+    Computing head velocity frame by frame
+
+    Args:
+        of_results: Openface raw out dataframe
+        r_config: Face config file object
+
+    Reutrns:
+        Final head velocity frame by frame output
+    """
+    distance_list = []
+    error_list = []
+    
+    of_results = of_results[col+ [' confidence']]
+    for index, row in of_results.iterrows():
+        dst = np.nan
+        
+        if index == 0 or float(row[' confidence']) < 0.8:  #Threshold < 0.8
+            distance_list.append(dst)
+            
+            if float(row[' confidence']) < 0.8:
+                error_list.append('confidence less than 80%')
+                
+            else:
+                error_list.append('Pass')
+            continue
+            
+        if index > 0:
+            
+            point_x = (of_results[col[0]][index-1], of_results[col[1]][index-1], of_results[col[2]][index-1])
+            point_y = (row[col[0]],row[col[1]],row[col[2]])
+            try:
+                dst = distance.euclidean(point_x, point_y)
+            except:
+                pass
+            
+            distance_list.append(abs(dst))
+            error_list.append('Pass')
+            
+    return distance_list, error_list
+        
+def calc_eye_mov(video_uri, df_of, out_loc, fl_name, r_config):
+    """
+    Computing eye motion variables
+    Args:
+        df_of: Openface dataframe
+        out_loc: Output path for saving output csv's
+        fl_name: file name for output csv
+        r_config: raw variable config file object
+        
+    """
+    
+    col_l = [ ' gaze_0_x', ' gaze_0_y', ' gaze_0_z'] 
+    col_r = [ ' gaze_1_x', ' gaze_1_y', ' gaze_1_z']
+    
+    gazel_disp, err_l = eye_disp(df_of, col_l, r_config)
+    gazer_disp, err_r = eye_disp(df_of, col_r, r_config)
+    
+    df_disp = eye_motion_df(gazel_disp, gazer_disp, err_l, r_config)
+    df_disp['dbm_master_url'] = video_uri
+    
+    df_motion = filter_motion(df_of, df_disp, col_l, col_r, r_config)
+    ut.save_output(df_motion, out_loc, fl_name, eye_pose_dir, eye_pose_ext)
+    
+def run_eye_gaze(video_uri, out_dir, r_config):
+    """
+    Processing all patient's for getting eye movement artifacts 
+    --------------------------------
+    --------------------------------
+    Args:
+        video_uri: video path; input_dir : input directory for video's
+        out_dir: (str) Output directory for processed output; r_config: raw variable config object
+    """
+    try:
+        
+        #filtering path to generate input & output path
+        input_loc, out_loc, fl_name = ut.filter_path(video_uri, out_dir)
+        of_csv_path = glob.glob(join(out_loc, fl_name + '_OF_features/*.csv'))
+
+        if len(of_csv_path)>0:
+
+            of_csv = of_csv_path[0]
+            df_of = pd.read_csv(of_csv, error_bad_lines=False)
+
+            logger.info('Processing Output file {} '.format(os.path.join(out_loc, fl_name)))
+            calc_eye_mov(video_uri, df_of, out_loc, fl_name, r_config)
+    except Exception as e:
+        logger.error('Failed to process video file')

dbm_lib/dbm_features/raw_features/nlp/speech_features.py

@@ -0,0 +1,47 @@
+"""
+file_name: speech_features
+project_name: DBM
+created: 2020-13-11
+"""
+
+import os
+import numpy as np
+import pandas as pd
+import glob
+from os.path import join
+import logging
+
+from dbm_lib.dbm_features.raw_features.util import util as ut
+from dbm_lib.dbm_features.raw_features.util import nlp_util as n_util
+
+logging.basicConfig(level=logging.INFO)
+logger=logging.getLogger()
+
+speech_dir = 'nlp/speech_feature'
+speech_ext = '_nlp.csv'
+transcribe_ext = 'nlp/transcribe/*_transcribe.csv'
+
+def run_speech_feature(video_uri, out_dir, r_config):
+    """
+    Processing all patient's for fetching nlp features
+    -------------------
+    -------------------
+    Args:
+        video_uri: video path; r_config: raw variable config object
+        out_dir: (str) Output directory for processed output
+    """
+    try:
+        
+        input_loc, out_loc, fl_name = ut.filter_path(video_uri, out_dir)
+
+        transcribe_path = glob.glob(join(out_loc, transcribe_ext))
+        if len(transcribe_path)>0:
+
+            transcribe_df = pd.read_csv(transcribe_path[0])
+            df_speech= n_util.process_speech(transcribe_df, r_config)
+
+            logger.info('Saving Output file {} '.format(out_loc))
+            ut.save_output(df_speech, out_loc, fl_name, speech_dir, speech_ext)
+            
+    except Exception as e:
+        logger.error('Failed to process video file')

dbm_lib/dbm_features/raw_features/nlp/transcribe.py

@@ -0,0 +1,84 @@
+"""
+file_name: transcribe
+project_name: DBM
+created: 2020-10-11
+"""
+
+import pandas as pd
+import numpy as np
+import librosa
+import glob
+from os.path import join
+import logging
+
+from dbm_lib.dbm_features.raw_features.util import util as ut
+from dbm_lib.dbm_features.raw_features.util import nlp_util as n_util
+
+logging.basicConfig(level=logging.INFO)
+logger=logging.getLogger()
+
+formant_dir = 'nlp/transcribe'
+csv_ext = '_transcribe.csv'
+error_txt = 'error: length less than 0.1'
+
+def calc_transcribe(video_uri, audio_file, out_loc, fl_name, r_config, deep_path, aud_dur):
+    """
+    Preparing Formant freq matrix
+    Args:
+        audio_file: (.wav) parsed audio file; fl_name: input file name
+        out_loc: (str) Output directory; r_config: raw variable config
+    """
+    
+    text = n_util.process_deepspeech(audio_file, deep_path)
+    df_formant = pd.DataFrame([text], columns=[r_config.nlp_transcribe])
+    
+    df_formant.replace('', np.nan, regex=True,inplace=True)
+    df_formant[r_config.nlp_totalTime] = aud_dur
+    df_formant[r_config.err_reason] = 'Pass'# will replace with threshold in future release
+    df_formant['dbm_master_url'] = video_uri
+    
+    logger.info('Saving Output file {} '.format(out_loc))
+    ut.save_output(df_formant, out_loc, fl_name, formant_dir, csv_ext)
+    
+def empty_transcribe(video_uri, out_loc, fl_name, r_config):
+    
+    """
+    Preparing empty formant frequency matrix if something fails
+    """
+    cols = [r_config.nlp_transcribe, r_config.nlp_totalTime, r_config.err_reason]
+    out_val = [[np.nan, np.nan, error_txt]]
+    df_fm = pd.DataFrame(out_val, columns = cols)
+    df_fm['dbm_master_url'] = video_uri
+    
+    logger.info('Saving Output file {} '.format(out_loc))
+    ut.save_output(df_fm, out_loc, fl_name, formant_dir, csv_ext)
+
+def run_transcribe(video_uri, out_dir, r_config, deep_path):
+    
+    """
+    Processing all patient's for fetching Formant freq
+    ---------------
+    ---------------
+    Args:
+        video_uri: video path; r_config: raw variable config object
+        out_dir: (str) Output directory for processed output; deep_path: deepspeech build path
+    """
+    try:
+        
+        input_loc, out_loc, fl_name = ut.filter_path(video_uri, out_dir)
+        aud_filter = glob.glob(join(input_loc, fl_name + '.wav'))
+        if len(aud_filter)>0:
+
+            audio_file = aud_filter[0]
+            aud_dur = librosa.get_duration(filename=audio_file)
+
+            if float(aud_dur) < 0.1:
+                logger.info('Output file {} size is less than 0.1 sec'.format(audio_file))
+
+                empty_transcribe(video_uri, out_loc, fl_name, r_config)
+                return
+
+            calc_transcribe(video_uri, audio_file, out_loc, fl_name, r_config, deep_path, aud_dur)
+    except Exception as e:
+        logger.error('Failed to process audio file')
+        

dbm_lib/dbm_features/raw_features/util/nlp_util.py

@@ -0,0 +1,212 @@
+"""
+file_name: nlp_util
+project_name: DBM
+created: 2020-10-11
+"""
+
+import subprocess
+import json
+import numpy as np
+import pandas as pd
+import os
+import logging
+
+import nltk
+import re
+from lexicalrichness import LexicalRichness
+from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer 
+
+logging.basicConfig(level=logging.INFO)
+logger=logging.getLogger()
+
+#Speech to text using Deepspeech 0.9.1
+def deepspeech(AUDIO_FILE,deep_path):
+    """
+        Extracting text from audio using Deep Speech neural network trained model
+        Returns:
+            Text: text which is extracted from audio
+    """
+    api = 'deepspeech'
+    arg_speech0 = '--model'
+    arg_speech_path0 = os.path.join(deep_path, 'deepspeech-0.9.1-models.pbmm')
+    arg_speech1 = '--scorer'
+    arg_speech_path1 = os.path.join(deep_path, 'deepspeech-0.9.1-models.scorer')
+    arg_audio = "--audio"
+    
+    out = subprocess.Popen([api, arg_speech0, arg_speech_path0, arg_speech1, arg_speech_path1, arg_audio, AUDIO_FILE],
+                           stdout=subprocess.PIPE, 
+                           stderr=subprocess.STDOUT)
+    logger.info('Deepspeech output...... {}'.format(out))
+    try:
+        stdout,stderr = out.communicate()
+    except:
+        return "error", "error"
+    print(stderr)
+    return stdout,stderr
+
+def deep_speech_output_clean(result):
+    """
+        Parsing deep speech output(text)
+        Return: 
+            Text from speech
+    """
+    text = ""
+    if len(result)>0:
+        res_split = str(result[0]).split('\\n')
+        
+        if len(res_split)>0:
+            for i in range(len(res_split)):
+                if 'Inference took' in res_split[i]:
+                    text = res_split[i + 1]
+                    return text
+    return text
+
+def process_deepspeech(audio_file,deep_path):
+    """
+    Transcribing audio to extract text from speech
+    """
+    deep_output = deepspeech(audio_file,deep_path)
+    deep_text= deep_speech_output_clean(deep_output)
+    
+    return deep_text
+
+def nltk_download():
+    
+    try:
+        nltk.data.find('tokenizers/punkt')
+        
+    except LookupError:
+        logger.info('punkt is not available')
+        nltk.download('punkt')
+        
+    try:
+        nltk.data.find('averaged_perceptron_tagger')
+        
+    except LookupError:
+        logger.info('averaged_perceptron_tagger is not available')
+        nltk.download('averaged_perceptron_tagger')
+
+def empty_speech(r_config, master_url, error_txt):
+    """
+    Preparing empty speech matrix with error
+    Args:
+        r_config: raw config file object
+        error_txt: Error message during transcription
+        
+    Returns:
+            Empty dataframe for speech features with error
+    """
+    
+    col = [r_config.nlp_numSentences, r_config.nlp_singPronPerAns, r_config.nlp_singPronPerSen, r_config.nlp_pastTensePerAns,
+          r_config.nlp_pastTensePerSen, r_config.nlp_pronounsPerAns, r_config.nlp_pronounsPerSen, r_config.nlp_verbsPerAns,
+          r_config.nlp_verbsPerSen, r_config.nlp_adjectivesPerAns, r_config.nlp_adjectivesPerSen, r_config.nlp_nounsPerAns,
+          r_config.nlp_nounsPerSen, r_config.nlp_sentiment_mean, r_config.nlp_mattr, r_config.nlp_wordsPerMin,
+          r_config.nlp_totalTime, r_config.err_reason]
+    
+    df_speech = pd.DataFrame([[np.nan] * len(col) + [error_txt]], columns = col)
+    df_speech['dbm_master_url'] = master_url
+    
+    return df_speech
+
+def divide_var(speech_var1, spech_var2):
+    """
+    divide variables
+    """
+    speech_var = np.nan
+    if spech_var2!=0:
+        speech_var = speech_var1/spech_var2
+    return speech_var
+
+def process_speech(transcribe_df,r_config):
+    """
+        Preparing speech features
+        Args:
+            transcribe_df: Transcribed dataframe
+            r_config: raw config file object
+        Returns:
+            Dataframe for speech features
+    """
+    
+    err_transcribe = transcribe_df[r_config.err_reason].iloc[0]
+    transcribe = transcribe_df[r_config.nlp_transcribe].iloc[0]
+    total_time = transcribe_df[r_config.nlp_totalTime].iloc[0]
+    master_url = transcribe_df['dbm_master_url'].iloc[0]
+    
+    #clean transcribe
+    transcribe = transcribe.replace(",", "")
+    transcribe = " ".join(re.findall(r"[\w']+|[.!?]", transcribe))
+    
+    if err_transcribe != 'Pass':
+        df_speech = empty_speech(r_config, master_url, error_txt)
+        
+        return df_speech
+    
+    speech_dict = {}
+    nltk_download()
+    
+    sentences = nltk.tokenize.sent_tokenize(transcribe)
+    words_all = nltk.tokenize.word_tokenize(transcribe)
+    num_sentences = len(sentences)
+    
+    speech_dict[r_config.nlp_numSentences] = num_sentences
+    
+    #nlp_singPron
+    i_s = transcribe.count('I')
+    me_s = transcribe.count('me')
+    my_s = transcribe.count('my')
+    sing_count = i_s + me_s + my_s
+    
+    speech_dict[r_config.nlp_singPronPerAns] = sing_count if len(words_all)>0 else np.nan
+    speech_dict[r_config.nlp_singPronPerSen] = divide_var(speech_dict[r_config.nlp_singPronPerAns], num_sentences)
+    
+    tagged = nltk.pos_tag(transcribe.split())
+    tagged_df = pd.DataFrame(tagged, columns=['word', 'pos_tag'])
+    
+    #Past tense per answer
+    all_POSs = tagged_df['pos_tag'].tolist()
+    speech_dict[r_config.nlp_pastTensePerAns] = all_POSs.count('VBD') if len(words_all)>0 else np.nan
+    speech_dict[r_config.nlp_pastTensePerSen] = divide_var(speech_dict[r_config.nlp_pastTensePerAns], num_sentences)
+    
+    #Pronoun per answer
+    pronounsPerAns = all_POSs.count('PRP') + all_POSs.count('PRP$')
+    speech_dict[r_config.nlp_pronounsPerAns] = pronounsPerAns if len(words_all)>0 else np.nan
+    speech_dict[r_config.nlp_pronounsPerSen] = divide_var(speech_dict[r_config.nlp_pronounsPerAns], num_sentences)
+    
+    #Verb per answer
+    verbPerAns = all_POSs.count('VB') + all_POSs.count('VBD') + all_POSs.count('VBG') \
+                      + all_POSs.count('VBN') + all_POSs.count('VBP') + all_POSs.count('VBZ')
+    speech_dict[r_config.nlp_verbsPerAns] = verbPerAns if len(words_all) > 0 else np.nan
+    speech_dict[r_config.nlp_verbsPerSen] = divide_var(speech_dict[r_config.nlp_verbsPerAns], num_sentences)
+    
+    #Adjective per answer
+    adjectivesAns = all_POSs.count('JJ') + all_POSs.count('JJR') + all_POSs.count('JJS')
+    speech_dict[r_config.nlp_adjectivesPerAns] = adjectivesAns if len(words_all) > 0 else np.nan
+    speech_dict[r_config.nlp_adjectivesPerSen] = divide_var(speech_dict[r_config.nlp_adjectivesPerAns], num_sentences)
+    
+    #Noun per answer
+    nounsAns = all_POSs.count('NN') + all_POSs.count('NNP') + all_POSs.count('NNS')
+    speech_dict[r_config.nlp_nounsPerAns] = nounsAns if len(words_all) > 0 else np.nan
+    speech_dict[r_config.nlp_nounsPerSen] = divide_var(speech_dict[r_config.nlp_nounsPerAns], num_sentences)
+    
+    #Sentiment analysis
+    vader = SentimentIntensityAnalyzer()
+    sentence_valences = []
+    
+    for s in sentences:
+        sentiment_dict = vader.polarity_scores(s) 
+        sentence_valences.append(sentiment_dict['compound'])
+        
+    speech_dict[r_config.nlp_sentiment_mean] = np.mean(sentence_valences) if len(sentence_valences) > 0 else np.nan
+    non_punc = list(value for value in words_all if value not in ['.','!','?'])
+    
+    non_punc_as_str = " ".join(str(non_punc))
+    lex = LexicalRichness(non_punc_as_str)
+    speech_dict[r_config.nlp_mattr] = lex.mattr(window_size=lex.words) if lex.words > 0 else np.nan
+    
+    #Number of words per minute
+    speech_dict[r_config.nlp_wordsPerMin] = divide_var(len(non_punc), total_time)*60
+    speech_dict[r_config.nlp_totalTime] = total_time
+    speech_dict['dbm_master_url'] = master_url
+    
+    df_speech = pd.DataFrame([speech_dict])
+    return df_speech

dbm_lib/dbm_features/raw_features/video/open_face_process.py

@@ -62,11 +62,14 @@ def process_open_face(video_uri, input_dir, out_dir, of_path, dbm_group):
     """
     try:
         
-        if dbm_group != None and len(dbm_group) == 1 and 'acoustic' in dbm_group:
+        if dbm_group != None:
-            return
+            check_group = ['facial','movement'] #add group here: if you want to use openface output for raw variable calculation
-
+            check_val = bool(len({*check_group} & {*dbm_group}))
+            if not check_val:
+                return
+        
         filepaths = [video_uri]
         csv_filepaths = batch_open_face(filepaths, video_uri, input_dir, out_dir, of_path)
         
     except Exception as e:
-        logger.error('Failed to process video file')
+        logger.error('Failed to process video file')

process_data.py

@@ -20,6 +20,7 @@ logging.basicConfig(level=logging.INFO)
 logger=logging.getLogger()
 
 OPENFACE_PATH = 'pkg/OpenFace/build/bin/FeatureExtraction'
+DEEP_SPEECH = 'pkg/DeepSpeech'
 DLIB_SHAPE_MODEL = 'pkg/shape_detector/shape_predictor_68_face_landmarks.dat'
 
 def common_video(video_file, args, r_config):
@@ -36,6 +37,8 @@ def common_video(video_file, args, r_config):
 
     pf.process_facial(video_file, out_path, args.dbm_group, r_config)
     pf.process_acoustic(video_file, out_path, args.dbm_group, r_config)
+    pf.process_nlp(video_file, out_path, args.dbm_group, r_config, DEEP_SPEECH)
+
     pf.remove_file(video_file)
     pf.process_movement(video_file, out_path, args.dbm_group, r_config, DLIB_SHAPE_MODEL)
 
@@ -79,6 +82,7 @@ def process_raw_audio_file(args, s_config, r_config):
                 
                 out_path = os.path.join(args.output_path, 'raw_variables')
                 pf.process_acoustic(audio_file[0], out_path, args.dbm_group, r_config)
+                pf.process_nlp(audio_file[0], out_path, args.dbm_group, r_config, DEEP_SPEECH)
                 
             else:
                 logger.info('Enter correct audio(*.wav) file path.')
@@ -130,6 +134,8 @@ def process_raw_audio_dir(args, s_config, r_config):
                 
                 out_path = os.path.join(args.output_path, 'raw_variables')
                 pf.process_acoustic(audio, out_path, args.dbm_group, r_config)
+                pf.process_nlp(audio, out_path, args.dbm_group, r_config, DEEP_SPEECH)
+                
             except Exception as e:
                 logger.error('Failed to process wav file.')
         

process_dbm.sh

@@ -55,6 +55,9 @@ fi
 if [[ $dbm_group == *"movement"* ]]; then
     dbm_new="$dbm_new movement"
 fi
+if [[ $dbm_group == *"nlp"* ]]; then
+    dbm_new="$dbm_new nlp"
+fi
 
 #docker commands to run container
 docker create -ti --name dbm_container dbm bash

requirements.txt

@@ -20,3 +20,8 @@ more_itertools
 scipy==1.2.0
 pyyaml
 pydub
+deepspeech
+nltk
+lexicalrichness
+vaderSentiment
+textblob

resources/features/derived_feature.yml

@@ -2,7 +2,8 @@ derive_feature:
     
     #DBM Feature Group
     FEATURE_GROUP: ['FAC_ASYM', 'FAC_AU', 'FAC_EXP', 'FAC_LMK', 'ACO_INT', 'ACO_FF', 'ACO_HNR', 'ACO_GNE', 'ACO_FM', 
-    'ACO_JITTER','ACO_SHIMMER', 'ACO_PAUSE', 'ACO_VFS', 'ACO_MFCC', 'MOV_HM', 'MOV_HP', 'EYE_BLINK']
+    'ACO_JITTER','ACO_SHIMMER', 'ACO_PAUSE', 'ACO_VFS', 'ACO_MFCC', 'MOV_HM', 'MOV_HP', 'EYE_BLINK', 'NLP_SPEECH',
+    'EYE_GAZE']
     
     #Feature group output file extensions
     FAC_ASYM_LOC: _facasym
@@ -22,6 +23,8 @@ derive_feature:
     MOV_HM_LOC: _headmov
     MOV_HP_LOC: _headpose
     EYE_BLINK_LOC: _eyeblinks
+    NLP_SPEECH_LOC: _nlp
+    EYE_GAZE_LOC: _eyegaze
     
     #Facial category feature group
     FAC_ASYM: ['fac_AsymMaskMouth', 'fac_AsymMaskEyebrow', 'fac_AsymMaskEye', 'fac_AsymMaskCom']
@@ -64,6 +67,14 @@ derive_feature:
     MOV_HM: ['head_vel']
     MOV_HP: ['mov_Hpose_Dist','mov_Hpose_Pitch','mov_Hpose_Yaw','mov_Hpose_Roll']
     EYE_BLINK: ['mov_blink_ear', 'vid_dur', 'mov_blinkdur']
+    EYE_GAZE: ['mov_leye_x', 'mov_leye_y', 'mov_leye_z', 'mov_reye_x', 'mov_reye_y', 'mov_reye_z', 'mov_eleft_disp',
+    'mov_eright_disp']
+    
+    #NLP category feature group
+    NLP_SPEECH: ['nlp_numSentences', 'nlp_singPronPerAns', 'nlp_singPronPerSen', 'nlp_pastTensePerAns', 'nlp_pastTensePerSen',
+    'nlp_pronounsPerAns', 'nlp_pronounsPerSen', 'nlp_verbsPerAns', 'nlp_verbsPerSen', 'nlp_adjectivesPerAns', 
+    'nlp_adjectivesPerSen', 'nlp_nounsPerAns', 'nlp_nounsPerSen', 'nlp_sentiment_mean', 'nlp_mattr', 'nlp_wordsPerMin',
+    'nlp_totalTime']
     
     #Calculation for variables
     # Facial Asymmetry
@@ -248,3 +259,30 @@ derive_feature:
     mov_blink_ear: ['mean', 'std']
     vid_dur: ['count']
     mov_blinkdur: ['mean', 'std']
+    mov_leye_x: ['mean', 'std']
+    mov_leye_y: ['mean', 'std']
+    mov_leye_z: ['mean', 'std']
+    mov_reye_x: ['mean', 'std']
+    mov_reye_y: ['mean', 'std']
+    mov_reye_z: ['mean', 'std']
+    mov_eleft_disp: ['mean', 'std']
+    mov_eright_disp: ['mean', 'std']
+    
+    #NLP feature
+    nlp_numSentences: ['mean']
+    nlp_singPronPerAns: ['mean']
+    nlp_singPronPerSen: ['mean']
+    nlp_pastTensePerAns: ['mean']
+    nlp_pastTensePerSen: ['mean']
+    nlp_pronounsPerAns: ['mean']
+    nlp_pronounsPerSen: ['mean']
+    nlp_verbsPerAns: ['mean']
+    nlp_verbsPerSen: ['mean']
+    nlp_adjectivesPerAns: ['mean']
+    nlp_adjectivesPerSen: ['mean']
+    nlp_nounsPerAns: ['mean']
+    nlp_nounsPerSen: ['mean']
+    nlp_sentiment_mean: ['mean']
+    nlp_mattr: ['mean']
+    nlp_wordsPerMin: ['mean']
+    nlp_totalTime: ['mean']

resources/features/raw_feature.yml

@@ -196,3 +196,32 @@ raw_feature:
     mov_Hpose_Yaw: mov_hposeyaw
     mov_Hpose_Roll: mov_hposeroll
     mov_Hpose_Dist: mov_hposedist
+    mov_leye_x: mov_lefteyex
+    mov_leye_y: mov_lefteyey
+    mov_leye_z: mov_lefteyez
+    mov_reye_x: mov_righteyex
+    mov_reye_y: mov_righteyey
+    mov_reye_z: mov_righteyez
+    mov_eleft_disp: mov_leyedisp
+    mov_eright_disp: mov_reyedisp
+
+    #NLP markers
+    nlp_transcribe: nlp_transcribe
+    nlp_numSentences: nlp_numSentences
+    nlp_singPronPerAns: nlp_singPronPerAns
+    nlp_singPronPerSen: nlp_singPronPerSen
+    nlp_pastTensePerAns: nlp_pastTensePerAns
+    nlp_pastTensePerSen: nlp_pastTensePerSen
+    nlp_pronounsPerAns: nlp_pronounsPerAns
+    nlp_pronounsPerSen: nlp_pronounsPerSen
+    nlp_verbsPerAns: nlp_verbsPerAns
+    nlp_verbsPerSen: nlp_verbsPerSen
+    nlp_adjectivesPerAns: nlp_adjectivesPerAns
+    nlp_adjectivesPerSen: nlp_adjectivesPerSen
+    nlp_nounsPerAns: nlp_nounsPerAns
+    nlp_nounsPerSen: nlp_nounsPerSen
+    nlp_sentiment_mean: nlp_sentiment_mean
+    nlp_mattr: nlp_mattr
+    nlp_wordsPerMin: nlp_wordsPerMin
+    nlp_totalTime: nlp_totalTime
+