Add tool for collecting episode titles

New tool 0_transcripts/collect_titles.py for extracting text with id="episode-title" from SVG files.

If you want to use the title panel instead of the thumbnail, use the regex EPISODE_REGEX = re.compile(r'E\d+P00\.svg')

I have pulled out common code to 0_transcripts/svg.py - have a look at the function def check_filter(tag, attrib_filter)

I have also added a new validation step to the CI pipeline, since the script has needs that are more strict than the title matching check.

Re. #86 (closed)