setseeker

fileshazzer.py splits DJ sets (or any MP3 you throw at it) into segments, runs them through Shazam, and spits out a timestamped tracklist. seekspawner.py takes those IDs and hunts the files down on Soulseek via slsk-batchdl. Bring your own Soulseek credentials (setup will store them for you), or create new ones if you don't have them

Requirements

• Python 3.11
• ffmpeg command line tool available on your $PATH
• .NET 6 SDK so we can build slsk-batchdl
• git (only needed the first time setup.sh clones slsk-batchdl)

setup.sh tries to install ffmpeg and the .NET SDK for you on macOS or Debian/Ubuntu. If you're on anything else (or the script still complains that .NET is missing), install those pieces manually, then rerun ./setup.sh.

Setup

From the repo root:

chmod +x setup.sh
./setup.sh

What that script takes care of:

• Creates a fresh virtual environment at .venv and installs requirements.txt
• Checks for ffmpeg; will try to install it if the binary isn't found
• Makes sure .NET 6 is around, then clones and builds slsk-batchdl into slsk-batchdl/slsk-batchdl/bin/Release
• Sets up the working folders: sets, tracklists, spoils, user, logs, tmp/segments, tmp/queries
• Stores encrypted Soulseek credentials at user/slsk_cred.json + user/slsk.key (set SLSK_USERNAME and SLSK_PASSWORD in your shell to skip the prompt)

You can rerun setup.sh any time; it will reuse what already exists, offer to rotate credentials, and rebuild slsk-batchdl if needed.

Run it

1. Drop MP3s into sets/ or point fileshazzer.py at a SoundCloud link (the soundcloud_url variable near the top).

2. Activate the virtualenv:

   source .venv/bin/activate
   

3. Pick your workflow:

   • Full pipeline (Shazam + Soulseek download)

     chmod +x launcher.sh
     ./launcher.sh
     

     launcher.sh activates the venv, runs fileshazzer.py, then immediately hands the fresh queries to seekspawner.py.

   • Just fingerprint and build tracklists

     python3.11 fileshazzer.py
     

     Tracklists (and their original MP3s) end up under tracklists/<set_name>/. If Shazam throttles and things look frozen, give it a minute—it will continue. Want more reliable matches? Bump segment_length (default 60 seconds) at the top of the script so each chunk contains more audio.

   • Already have a tracklist and only need Soulseek

     python3.11 seekspawner.py
     

     The script looks for *.txt tracklists under tracklists/, parses them into tmp/queries/queries.txt, and fires those queries at Soulseek. You can drop your own tracklist files in there as long as they follow the same [hh:mm:ss] Artist - Title format.

seekspawner.py logs anything it had to skip to logs/skipped_queries.log, and downloads land in spoils/.

Example tracklist output

Final Tracklist:
[00:01:00] Umek - Center of Gravity
[00:13:30] Sade - Like Tattoo
[00:21:00] Zimmie Gix - Absolute Chill
[00:21:30] Andrea Frisina & Irregular Synth - Dub City
[00:22:00] Alan Fitzpatrick - Brian's Proper Dun One
[00:23:30] R.A.W. - Unbe (Erick 'More' Mix)
[00:24:00] Terrace - Magic O
...etc