Basic Python script for fast extraction of stills from videos through calls to ffmpeg using filenames and timestamps from an SPSS file

extract_frames.py

import os
import subprocess
import pyreadstat
import logging
import datetime


# configuration
input_directory = 'C:/Users/robin/PycharmProjects/asmir_cry_project/in'
output_directory = 'C:/Users/robin/PycharmProjects/asmir_cry_project/out'
spss_file = "C:/Users/robin/PycharmProjects/asmir_cry_project/crying_points_and_periods.sav"


# setup logging

def setup_logger(name, log_file, level=logging.INFO):
    handler = logging.FileHandler(log_file, mode='w')
    formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s')
    handler.setFormatter(formatter)
    logger = logging.getLogger(name)
    logger.setLevel(level)
    logger.addHandler(handler)
    return logger


success_log = setup_logger('success', 'success.log')
fail_log = setup_logger('failure', 'fail.log')

# case insensitive file matching


def getfile_insensitive(path):
    directory, filename = os.path.split(path)
    directory, filename = (directory or '.'), filename.lower()
    for f in os.listdir(directory):
        newpath = os.path.join(directory, f)
        if os.path.isfile(newpath) and f.lower() == filename:
            return newpath


def isfile_insensitive(path):
    return getfile_insensitive(path) is not None


# main

def extract_frame(video_name, ss_time, in_dir, out_dir, name_append=""):
    if not os.path.exists(out_dir):
        os.makedirs(out_dir)
    output_name = os.path.splitext(video_name)[0] + name_append + ".jpg"
    video_with_path_and_ext = os.path.join(in_dir, video_name)

    if isfile_insensitive(video_with_path_and_ext + ".wmv"):
        video_with_path_and_ext = getfile_insensitive(video_with_path_and_ext + ".wmv")
    elif isfile_insensitive(video_with_path_and_ext + ".mp4"):
        video_with_path_and_ext = getfile_insensitive(video_with_path_and_ext + ".mp4")
    else:
        fail_log.info(video_name + " " + ss_time + " file_missing")
        return
    output_name = os.path.join(out_dir, output_name)
    try:
        subprocess.call(['ffmpeg', '-y', '-ss', ss_time, '-i', video_with_path_and_ext, '-vframes', '1',
                         '-q:v', '2', output_name])
    except subprocess.CalledProcessError:
        fail_log.info(video_name + " " + ss_time + " conversion_failed")
        return
    success_log.info(video_name + " " + ss_time + " conversion_completed")


df, meta = pyreadstat.read_sav(spss_file)

for index, row in df.iterrows():
    file_name = str.strip(row['file_name'])
    start_time = datetime.datetime.strptime(row["start_time"].lstrip(), '%M:%S').strftime('%H:%M:%S')
    string_time = start_time.replace(':', '_')
    crying_type = "__" + string_time + "__" + str(int(row["crying_type"]))
    extract_frame(file_name, start_time, input_directory, output_directory, crying_type)

logging.shutdown()