PR #1: CSV naming convention changes + fixing audio file checks

[carmen-chau]

PR fix for issues: #7, #13

Here are a list of changes to the PR:

1. Changing CSV file name to follow the naming convention:

file_name_transcription_time.csv

OR

file_name_translation_time.csv

depending on selected choice.

NOTE: There is a bug with the naming, see comments below.

NOTE 2: Will re-add renaming for the dual transcription + translation features once PR #11 has been reviewed or feedback provided.

2. Expanding audio file metadata checks

All file extensions as listed in issue #7 should be implemented in method validate_audio_file.

[moonsdust]

We should consider more audio file types

[carmen-chau]

We should consider more audio file types

Yep I'm working on that. FYI we can really only add the audio file type checks based on what file types Whisper supports

[carmen-chau]

@moonsdust I made changes to the file naming. So now the backslash shouldn't appear in between the time digits (with regards to the final file name).

Current file naming convention is: audio_transcription1133.csv. There is no ":" due to a bug (see below)

One weird bug though that I cannot seem to understand, is that when you add a ":" to seperate the hour digits from the minute digits, the backslash appears instead.

Ie: If I tried coding up the file naming formatting to be: audio_transcription11:33.csv, it formats it to be audio_transcription11\33.csv.

Would appreciate if you could let me know if this happens on your end too

[moonsdust]

@moonsdust I made changes to the file naming. So now the backslash shouldn't appear in between the time digits (with regards to the final file name).

Current file naming convention is: audio_transcription1133.csv. There is no ":" due to a bug (see below)

One weird bug though that I cannot seem to understand, is that when you add a ":" to seperate the hour digits from the minute digits, the backslash appears instead.

Ie: If I tried coding up the file naming formatting to be: audio_transcription11:33.csv, it formats it to be audio_transcription11\33.csv.

Would appreciate if you could let me know if this happens on your end too

You can't include ":" in file names on MacOS. That's why it changes ":" to "\". @carmen-chau

[carmen-chau]

@moonsdust I made changes to the file naming. So now the backslash shouldn't appear in between the time digits (with regards to the final file name). Current file naming convention is: audio_transcription1133.csv. There is no ":" due to a bug (see below) One weird bug though that I cannot seem to understand, is that when you add a ":" to seperate the hour digits from the minute digits, the backslash appears instead. Ie: If I tried coding up the file naming formatting to be: audio_transcription11:33.csv, it formats it to be audio_transcription11\33.csv. Would appreciate if you could let me know if this happens on your end too

You can't include ":" in file names on MacOS. That's why it changes ":" to "\". @carmen-chau

I did not know that! I tried putting an underscore and it works. I hope that change is ok to you.

At the moment the file names are inching on the longer side... I think that's another improvement to this CLI we can work on in the future.

[carmen-chau]

@moonsdust. The reason why the .mp4 file you suggested didn't work is because.... it wasn't actually a .mp4 file (using python magic library reveals it is actually a ADTS file).

Repull this PR again, and try again using the same spanish .mp4 file. It should work now.

Note: I can also confirm that the .m4a spanish file works fine

This is an interesting edge case worth noting in the documentation (ie: Some audio files are disguised as .mp4 or other extensions when in reality, they are ADTS files)