Skip to content

Commit

Permalink
adding post processing
Browse files Browse the repository at this point in the history
  • Loading branch information
tim-roethig-db committed May 26, 2024
1 parent a783f1b commit d122e56
Show file tree
Hide file tree
Showing 2 changed files with 15 additions and 12 deletions.
25 changes: 14 additions & 11 deletions amondin/post_processing.py
Original file line number Diff line number Diff line change
Expand Up @@ -24,15 +24,18 @@ def merge_rows_consecutive_speaker(transcript: pd.DataFrame) -> pd.DataFrame:
:param transcript:
:return:
"""
transcript['speaker_group'] = (transcript['speaker'] != transcript['speaker'].shift()).cumsum()

transcript = transcript.groupby(['speaker_group', 'speaker']).agg({
'start': "min",
# create a column speaker_group that signals if speakers have consecutive segments
transcript["speaker_group"] = (transcript["speaker"] != transcript["speaker"].shift()).cumsum()

# group by speaker_group and speaker to merge consecutive segments
transcript = transcript.groupby(["speaker_group", "speaker"]).agg({
"start": "min",
"end": "max",
'text': lambda x: ' '.join(x)
"text": lambda x: " ".join(x)
}).reset_index()

transcript = transcript.drop(columns='speaker_group')
# drop helper column speaker_group
transcript = transcript.drop(columns="speaker_group")

return transcript

Expand All @@ -43,15 +46,15 @@ def format_time_stamp(transcript: pd.DataFrame) -> pd.DataFrame:
:param transcript:
:return:
"""
transcript['start'] = transcript['start'].apply(_seconds_to_time_stamp)
transcript['end'] = transcript['end'].apply(_seconds_to_time_stamp)
transcript["start"] = transcript["start"].apply(_seconds_to_time_stamp)
transcript["end"] = transcript["end"].apply(_seconds_to_time_stamp)

transcript['time_stamp'] = transcript.apply(
transcript["time_stamp"] = transcript.apply(
lambda row: f"{row['start']} -> {row['end']}",
axis='columns'
axis="columns"
)

return transcript[['speaker', "time_stamp", "text"]]
return transcript[["speaker", "time_stamp", "text"]]


if __name__ == "__main__":
Expand Down
2 changes: 1 addition & 1 deletion amondin/speech2text.py
Original file line number Diff line number Diff line change
Expand Up @@ -52,5 +52,5 @@ def speech2text(
}
)

# return sting in list
# return string in a list
return [result["text"] for result in results]

0 comments on commit d122e56

Please sign in to comment.