Audio Split line by line

(should be ready to use Google Speech Recognition API)

0. Prepare Google API

gcloud auth login

1. Prepare data and directory

Make a directory for the audio

mkdir ./{title} ./{title}/audio ./{title}/pre_audio

Change the name of files and locate them in the directory above (./{title})

- original text file: {title}_ORIGINAL.txt
  (eg. androcles-shorter_ORIGINAL.txt)

- original audio file: {title}.mp3  
  (eg. androcles-shorter.mp3)

2. Change ROOT in main.py

Title above and the ROOT below should be the same

ROOT = "title"

3. Run main.py

python main.py

Final audio files will be created in audio directory

Caution

• This code cannot make sentence audio files perfectly. (about maximum 85% can be processed)

• Success rate can be dramatically lower if the audio file has more sentences than that of the text file.

• Add . after the title of the text IN THE TEXT FILE if there is any. (To split sentences correctly)

• If you want to observe the analyzing process, change DEBUG in each python file to 1 (Currently 0)

---

Future Works

• Solve the problem of having two candidate indices (eg. when 'me' appears twice in a sentence)

• Add a function to check the unknown chunks depending on its number

• Add a function to update once more with the way of increasing the index of chunk, rather than sentence number

• Process Bar (tqdm)

• Add docopt