Kotori

Kotori (japanese: songbird) is a server- and client-side application for managing and tagging local music-files, which offers features for detecting file duplicates by checking audio-fingerprints.

Concept

Managing application (Web UI)

• Web application for managing and editing track information
• Authorization required for editing
• Datasets can be filtered
• Snapshots are created on each dataset modification
• Modifications shall be explicitly saved, so multiple datasets may be executed at once
• Abusing users may be reported by authorized users
• Modifications may be reverted to a previous snapshot by privileged users
• Supported tags are described on: http://id3.org/id3v2.4.0-frames
• Supported file types for uploading album covers are:
  • *.jpg
  • *.png

Match tool (Local tool)

• Native client-side application

• Receives track information for local music-files

• Supported file types are:

  • *.mp3

• Capable of overwriting or merging ID3 tags of specified files with received values

• GUI and CLI

• CLI:

  • Node.js script
  • Usage: node run kotori.js [OPTIONS] file
  • Options can be specified by flags available:
    • -b, --backup Create backup of each edited file in specified directory
    • -h, --help Show help on how to use this monster
    • -l, --listduplicate Find and print paths of all duplicates to stdout, sorted by quality, best quality first
    • -m, --merge Merge existing ID3 tag values of specified tracks with received information
    • -o, --overwrite Overwrite existing ID3 tag values of specified files with received values
    • -r, --report List received information per specified track and print to stdout
    • -v, --verbose Print further information during progress
  • Examples:
    • List all duplicates of mp3-files in current path and write to file: node run kotori.js -l *.mp3 > output.txt
    • Merge received information with ID3 tags of mp3-files in current path and print status during progress: node run kotori.js -m -v *.mp3
    • Print all received track information for specified file node run kotori.js -r "foo - bar.mp3"

• GUI:

  • Wrapper for CLI
  • Drag-and-drop functionality for selecting tracks (folders / files)
  • Offers different actions for handling each duplicate:
    • Keep file with best quality only
    • Write paths of duplicates to file
    • Ignore
  • A selected action may be marked as default for current process

Use cases

• Detect duplicates of tracks on your HDD

Kotori is capable of detecting duplicates of locally stored tracks, by creating audio-fingerprints. By creating audio-fingerprints even unnamed duplicated files will be recognized or duplicates of different quality, size and track length.

• Add proper ID3 tags to your tracks

Missing or incomplete ID3 tags of tracks in a huge music collection can be easily set and merged from our database, which can be collaboratively completed on our web application.

Functionality

Backend

Technology

• PostgreSQL
• NodeJS
• NodeJS packages:
  • Express
• RESTful API

REST-API

Paths

• /v1/tracks/query
  • Method: POST
  • Accept: application/json
  • Description: Returns all available song information from DB for specific songs, which must be determined by its fingerprints. Fingerprints of songs must be passed as JSON in request-body.
  • Example:
    • Request-Body:

      
      [
          {
              "fingerprint": [0, 0, 0],
          },
          {
              "fingerprint": [0, 0],
          },
          ...
      ]

    • Response-Body:

      
      [
          {
              "fingerprint": [0, 0, 0],
              "results": [
                  {
                      "trackId": "1",
                      "artist": "John Doe",
                      ...
                  },
                  {
                      "trackId": "20",
                  },
                  ...
              ]
          },
          {
              "fingerprint": [0, 0],
              "results": [
                  {
                      "trackId": "12",
                      "artist": "Jane Doe",
                      ...
                  },
                  {
                      "trackId": "2",
                      ...
                  },
                  ...
              ]
          },
          ...
      ]

• /v1/tracks
  • Method: PUT
  • Accept: application/json
  • Description: Updates stored information of specific songs stored on server with passed values. Expects several song information grouped by its related fingerprints as JSON in request-body.
  • Example:
    • Request-Body:

      
      [
          {
              "fingerprint": [1, 1, 1],
              "tags": {
                  "year": 1981
              }
          },
          {
              "fingerprint": [0, 0],
              "tags": {
                  "artist": "Jane Doe",
                  "title": "Foo",
                  ...
              }
          },
          {
              "fingerprint": [0, 1],
              "tags": {
                  "album": "Baz Bat - Part 1"
                  ...
              }
          },
          {
              "fingerprint": [1, 0],
              "tags": {
                  ...
              }
          },
          ...
      ]