ChildRecordData provides specifications and tools for the storage and management of day-long recordings of children and their associated meta-data and annotations.
See the formatting instructions and specifications
- Install the package via pip
The package can be installed using pip :
pip install git+https://github.com/LAAC-LSCP/ChildRecordsData.git
(The package requires python 3, so make sure pip points to a python3 installation, otherwise try pip3.)
If you are having permissions issues - or any other issue -, you can create and activate a python environment beforehand :
python3.6 -m venv ~/ChildProjectVenv
source ~/ChildProjectVenv/bin/activate
pip install git+https://github.com/LAAC-LSCP/ChildRecordsData.git
- Install git-annex
The package also requires git-annex in order to retrieve the datasets. It can be installed via apt install git-annex or brew install git-annex. It is encouraged to use a recent version of git-annex (8.2+).
The list of available datasets with instructions to download them can be found here.
child-project validate /path/to/dataset
Looks for errors and inconsistency in the metadata, or for missing audios. The validation will pass if the formatting instructions are met.
child-project convert /path/to/dataset --name=16kHz --format=wav --sampling=16000 --codec=pcm_s16le
Converts all recordings in a dataset to a given encoding. Converted audios are stored into converted_recordings/$name.
With audio splitting every 15 hours :
child-project convert /path/to/dataset --name=16kHz --split=15:00:00 --format=wav --sampling=16000 --codec=pcm_s16le
sbatch --mem=64G --time=5:00:00 --cpus-per-task=4 --ntasks=1 -o namibia.txt child-project convert /path/to/dataset --name standard --format WAV --codec pcm_s16le --sampling 16000 --threads 4`
Annotations can be imported one by one or in bulk. Annotation importation does the following :
- Convert all input annotations from their original format (e.g. rttm, eaf, textgrid..) into the CSV format defined here and stores them into
annotations/. - Registers them to the annotation index at
metadata/annotations.csv
child-project import-annotations /path/to/dataset --set eaf --recording_filename sound.wav --time_seek 0 --raw_filename example.eaf --range_onset 0 --range_offset 300 --format eaf
child-project import-annotations /path/to/dataset --annotations /path/to/dataframe.csv
The input dataframe /path/to/dataframe.csv must have one entry per annotation to import, according to the format specified here.
Compute recordings duration and store in into a column named 'duration' in the metadata.
child-project compute-durations [--force] /path/to/dataset