A small python script that will modify OLX to use an instance of Open edX without edX's mobile video features.

edx_id_vid.py

#!/bin/env python
# -*- coding: utf-8 -*-
"""
A number of functions to rename and modify edXml video tags in order to make them mobile ready.

.. note:: This assumes that the csv file is ordered youtube_id, edx_id
"""

import sys

import xml.etree.ElementTree as ET
import csv


def print_help():
    print('''usage: edx_id_vid -h
      edx_id_vid -a csv_filename xml_filename
      edx_id_vid -d xml_filename''')


def load_csv(csv_fn):
    '''Load a csv document.'''
    edx_id_file = open(csv_fn, 'r')
    edx_id_csv = csv.reader(edx_id_file)
    edx_id_dict = {}

    for line in edx_id_csv:
        if line[0] != '':
            edx_id_dict[line[0]] = line[1]

    return edx_id_dict


def load_xml(xml_fn):
    '''Load an xml document.'''
    video_xml = ET.parse(xml_fn)
    return video_xml


def remove_edx_ids(xml_fn):
    '''
    Check an xml file for video tags, if any edx_video_ids exist, delete them.
    '''    
    video_xml = load_xml(xml_fn)
    root = video_xml.getroot()
    for video in root.iter('video'):
        if 'edx_video_id' in video.keys():
            del video.attrib['edx_video_id']
            print('edx_video_id removed in {}'.format(xml_fn))
    video_xml.write(xml_fn)
    print ('Wrote {}'.format(xml_fn))


def add_edx_ids(csv_fn, xml_fn):
    '''
    Compare any youtube_x_x... or youtube attributes and gather youtube IDs.
    If any youtube ids exist that are in the csv, add an edx_video_id attribute.
    '''
    
    video_xml = load_xml(xml_fn)
    edx_id_dict = load_csv(csv_fn)
    root = video_xml.getroot()
    attribs = ['youtube_id_0_75',
               'youtube_id_1_0',
               'youtube_id_1_25',
               'youtube_id_1_5',
               ]
    for video in root.iter('video'):
        youtube_ids = []
        yt_ids = video.attrib['youtube'].split(',')
        for yt_id in yt_ids:
            youtube_ids.append(yt_id.split(':')[1])

        for attrib in attribs:
            if attrib in video.attrib.keys():
                youtube_ids.append(video.attrib[attrib])

        for youtube_id in youtube_ids:
            if youtube_id in edx_id_dict.keys():
                video.set('edx_video_id', edx_id_dict[youtube_id])
                video_xml.write(xml_fn)
                print ('Wrote {}'.format(xml_fn))
                break


if __name__ == '__main__':
    if ('-h' in sys.argv[1]):
        print_help()
    elif ('-a' in sys.argv[1]):
        add_edx_ids(sys.argv[2], sys.argv[3])
    elif ('-d' in sys.argv[1]):
        remove_edx_ids(sys.argv[2])

This script serves to add edx_video_id attributes to video files from a CSV, but also to remove them for our own instance of OpenEdX, which does not use the video abstraction layer (VAL).

To run it on a number of files I use this bash one-liner:
$ for fn in video/*.xml; do; python edx_id_vid.py -d $fn; done

or if you want to add edx_video_id attributes:
$ for fn in video/*.xml; do; python edx_id_vid.py -a csv_of_IDs.csv $fn; done

For adding edX ids, this worked for me
$ for fn in video/*.xml; do python edx_id_vid.py -a csv_of_IDs.csv $fn; done
without the ';' after 'do'

Saving .csv files in mac OS format produced an error:
_csv.Error: new-line character seen in unquoted field - do you need to open the file in universal-newline mode?

to overcome this error, I saved the csv file in 'windows comma separated file' per suggestions from stack overflow and it worked
http://stackoverflow.com/questions/17315635/csv-new-line-character-seen-in-unquoted-field-error

I think that line 47
if 'edx_video_id' in video.keys:

should be
if 'edx_video_id' in video.attrib:

Oops, I didn't notice your comments until now. Sorry, @srayyan. video.attrib or video.keys are instance methods, and I was treating them like objects here. I just fixed that.