The Pandemic Archive of Voices - Database & API

Description

The Pandemic Archive of Voices is an audio archive constituted by 24 different speakers who contributed 185 audios in 18 different languages. Each of these audios represents a word or expression in the speaker's native language that would translate their own individual experience of the current context of the pandemic.

This repository was created especifically to manage the installation project based on this database, but not exclusivelly.

It consists of:

• Database API
• Content Manager
• Contribution Form
• Local server to conntect with the installation API

API URL

https://pandemic-archive-of-voices.herokuapp.com/api/

Dependencies

• Mongoose Node library dealing with MongoDB
• Amazon S3
• Socket.io
• node-osc
• node-ffmpeg
• SoX for audio conversion from .ogg to .wav and .wav.
• opus-tools for Opus .ogg file conversion.

API

GET

• Get all database data

  Returns whole JSON Object containing all audio objects

  /api/data

• *Get speaker's audios by speaker's id

  Returns all audio objects from speaker

  /api/speaker_id/:id

• *Get speaker's audios by speaker's id

  Returns all spakers who contain the :name string

  /api/speaker/:name

• Get audio by id

  Returns specific audio object from its id

  /api/audio_id/:id

• Get audios by text

  Returns all audios that contain the text in the transcription value text in audio object

  /api/audio_id/:text

• Get sample array from audio id

  Returns the sample array of a particular audio in the database

  /api/get_audio_samples/:audio_id/:bits/:sample_rate

• Get compressed audio file

  Compresses and return the audio file from the databse with a designated bit depth and sample rate

  /api/get_compressed_audio_file/:audio_id/:bits/:sample_rate

• Get encrypted text from audio id

  Returns the audio waveform data converted into text based on "alphabet" encryption

  /api/get_audio_samples_characters/:audio_id/:bits/:sample_rate

• Get generated audio file from text string

  Returns an audio file from a character string, translating the characters into numbers, and then waveform

  /api/get_audio_samples_characters/:text

POST

• POST new audio

  Submits new audio to the database, restrited use

  /api/audio

  {
    id: String,
    data: Blob,
    languageInput: String,
    lang_other: String,
    text: String,
    timestamp: Int,
    deleted: Boolean,
    disabled: Boolean
  }

PUT

• PUT audio data

  Updates a audio object in the database, restrited use

  /api/audio

  {
     id: String,
     deleted: Boolean,
     disabled: Boolean,
     text: String
  }

Content Manager

Live update content manager

Contribution Form

Contribition form

Sound Writing

https://pandemic-archive-of-voices.herokuapp.com/soundWriting/

References

• About converting Opus .ogg files used by telegram voice messages: SoX doesn't work with Opus Audio files

• Sample conversion on SoX SoX resample and convert

License

MIT