Scripts Used By the Gardner Lab

Basic Info

A collection of python scripts people in the lab use to extract audio form recording room computers, and distribute these recordings to a users local computer for analysis.

Important: some of these scripts rely on the OSX file structure and may not work as intended on a Windows machine (for instance: Matlab install dir).

Prerequisites

1. Python installed on your computer
2. A Mac

The mose useful file is:

get_todays_wav_files.py

This is a python script that copies audio files from our recording room computer, moves them to the user's computer, and extracts only the segments of the file that contain bird calls and songs. It can be run as an executable.

Script summary

When run with the proper commandline arguments, this script will move files from the GardnerLab Recording Room Computer. These files are only associated with the recording channel(s) and the date of recording that you specify. The files this script moves are located in the [channel_number]-WAV folder. After moving, it deletes the images (located in the [channel_number]-IMG folder) associated with those audio files. If the script doesn't find any files, it quits and will let the user know there weren't any audio files found. Then it runs a matlab script in the folder that the audio files were moved to, called zftftb_song_chop. It 'chops' the files into small chunks of only a couple songs per file, which saves a lot of disk space. After this, the script removes the original, larger files, and will then let the user know it has run successfully.

Installation Prerequisites:

1. Make sure you have MATLAB also installed on your computer. This script assumes MATLAB is located in your /Applications folder
2. Download Jeff Markowitz's song extraction code jmarkow. Repos needed are:
   • markolab
   • zftftb
   • robofinch
3. Make sure the recording computer's 'recording' folder is mounted on your computer using cmd+shift+k in finder on your mac, and search for Recording Computer 2
   • It should be mounted at /Volumes/recording, because the script assumes this path.
4. To set up to run daily, open up your terminal and type (crontab -e). Then refer to the set-up
   • To edit the crontab in nano, type:

     env EDITOR=nano crontab -e

   • (Or use emacs and become a wizard like Ben ⚡)

Use a configuration file

Store all the information you need in this simple json format shown below. You can configure the recording channels to copy over, the bird names, and the output directories for each bird (or all the birds if you choose).

{
  "output": "/Volumes/Untitled/winSongData/", // default savepath
  "birds": [
    {"name": "lny29-post",
    "channel": "11"},
    {"name": "lny29-post",
    "channel": "11",
    "output": "/Volumes/Untitled/winSongData/"} // override the default savepath for specific birds
  ]
}

Crontab set-up

And add this to your crontab:

0 21 * * * python [path to this script]/get_todays_wav_files.py --config [path to config file]

0 21 * * * means it will run this script every day at 21:00 (or 11:00 PM).

and:

0 8 * * * python [path to this script]/send_log_file.py [email protected]

The second script will send an email to [email protected] at 8:00 AM to give a status update from the script. It will send you information on which audio files were copied over, and if the script could not find any new files for the day.

The Old Method

This method was the first implementation, and still works, put is not recommended to use, because it is tedious to write the following code for each bird that needs files copied. It allowed for more errors.

To run every day, (for each bird) at 10PM:

0 21 * * * python PY_PATH/get_todays_wav_files.py -b BIRDID -c BOXID -d DESTINATION_PATH

Where...

PY_PATH = the full path to where you saved labPy

BIRDID = the bird's ID/name

BOXID = the box ID i.e. the recording channel

DESTINATION PATH = the path where you want to save the processed data

Each bird should be entered in on a different line, separated by ~20 minutes, to assure the network where the files copy through does not choke.

The same summary email as above can be sent with the same command:

0 8 * * * python PY_PATH/send_log_file.py [email protected]

You need to supply an email for this script, or else it will throw an exception.