Misc transcript fixes + new scripts
I was playing with the functions that built our transcription (aka talking to snakes again) and ended up with a script that shows mismatched names with the dictionary. Some names were manually translated in the transcript (by translators), some were updated in the dictionary and the old name was left in the transcript.
I could see and update/fix a lot of them. (I analyzed what is majorly used to see what was old. In doubt I checked with the comic.)
Note: There is a "typo" in commit's name: JB to match jb.po
And I also plan to port ported my mass transcript regeneration script (.py) to the project. It will be useful for pipeline fixes.
It regenerates only existing .md files and is used as python 0_transcripts/regenerate_transcripts.py <locale/EPnum>.
If there is no locale, it will regenerate them all. It works on Linux and Windows, just like the extract_text.py.
All that loud output is suppressed:
$ python 0_transcripts/regenerate_transcripts.py fr
Regenerating ep01_fr_transcript.md
Regenerating ep02_fr_transcript.md
Regenerating ep03_fr_transcript.md
Regenerating ep04_fr_transcript.md
...
Regenerating ep37_fr_transcript.md
Regenerating ep38_fr_transcript.md
0_transcripts]$ python regenerate_transcripts.py 25
Regenerating ep25_da_transcript.md
Regenerating ep25_de_transcript.md
Regenerating ep25_en_transcript.md
Regenerating ep25_fr_transcript.md
...I prepared the script used to find anomalies for general use and added the functionality of the !999 (comment 2002032)
Current output:
Searching for anomalies in the folder structure:
ep01_Potion-of-Flight/lang/jb/todo.md
ep22_The-Voting-System/lang/vi/README.txt
ep38_The-Healer/lang/vi/README.md
Searching for anomalies in transcript names:
24 jb: <unknown>
27 jb: <unknown>
28 jb: <unknown>
32 kr: 병사
37 it: Testo scrittoBut there was way more transcript anomalies (60 distinct lines) before cleanup/fixes.
Next stop: making all unneeded and unwanted nowhitespace go away. They can have their space in /dev/null ...