README.md

PRAAT file aggregator

This script is used to aggregate the output of algorithms generated with PRAAT into a CSV

The scrip will find all files in a directory and its sub-directory that match the algorithm and perform the operation on them.

It reads all files with an extension equal to the chosen_algorithm and transform each of them in a single line of a CSV.

Installation

Install dependencies

Depending on your OS, do these following steps to install the required software dependencies.

MacOS systems

brew install python

Linux Systems (Debian)

sudo apt get install python

Windows System

• Download and install Python

Install the script

Open a terminal and navigate to the directory you want to install the script.

Clone this repository using this line:

git clone https://github.com/Captnlink/praat-aggregator-csv.git

How to use

Data files structure

Each file should be named participantID_sessionName_projectName_SentenceName.algorithm. Each file should be in a subfolder named after the frequency range of the PRAAT analysis

You should prepare the folder containing the PRAAT data like this:

.
└── PRAAT_data/
    └── frequency_range/
        └── participantID_sessionName_projectName_SentenceName.algorithm
        └── ...

.
└── PRAAT_data/
    └── 250-300/
        └── p12345_v1_projectX_s1.norntimef0
        └── ...

Executing the script

Execute the following command line replacing PATH/TO/DIRECTORY" with the path of the PRAAT file location and chosen_algorithm with the appropriate algorithm.

python praat-aggregator-csv/sources/main.py -d "PATH/TO/DIRECTORY" -a chosen_algorithm

The possible algorithms are:

• normtimef0
• normtime_f0acceleration
• normtime_f0velocity
• normtime_semitonef0

Generated files

The generated files will be located under praat-aggregator-csv/output

• algorithm_aggregation.csv: The aggregated values of all praat files with target algorithm
• algorithm_participants.json: The count of the number of occurrence of a certain algorithm