The GreenKey ASRToolkit provides tools for file conversion and ASR corpora organization. These are intended to simplify the workflow for building, customizing, and analyzing ASR models, useful for scientists, engineers, and other technologists in speech recognition.
File formats supported
File formats have format-specific handlers in asrtoolkit/data_handlers. The scripts
txt, and GreenKey
json formatted transcripts. A custom
html format is also available, though this should not be considered a stable format for long term storage as it is subject to change without notice.
usage: convert_transcript [-h] input_file output_file convert a single transcript from one text file format to another positional arguments: input_file input file output_file output file optional arguments: -h, --help show this help message and exit
This tool allows for easy conversion among file formats listed above.
Note: Attributes of a segment object not present in a parsed file retain their default values
- For example, a
segmentobject is created for each line of an STM line
- each is initialized with the following default values which are not encoded in STM files:
usage: wer [-h] [--char-level] [--ignore-nsns] reference_file transcript_file Compares a reference and transcript file and calculates word error rate (WER) between these two files positional arguments: reference_file reference "truth" file transcript_file transcript possibly containing errors optional arguments: -h, --help show this help message and exit --char-level calculate character error rate instead of word error rate --ignore-nsns ignore non silence noises like um, uh, etc. This tool allows for easy comparison of reference and hypothesis transcripts in any format listed above.
usage: clean_formatting.py [-h] files [files ...] cleans input *.txt files and outputs *_cleaned.txt positional arguments: files list of input files optional arguments: -h, --help show this help message and exit
This script standardizes how abbreviations, numbers, and other formatted text is expressed so that ASR engines can easily use these files as training or testing data. Standardizing the formatting of output is essential for reproducible measurements of ASR accuracy.
usage: split_audio_file [-h] [--target-dir TARGET_DIR] audio_file transcript Split an audio file using valid segments from a transcript file. For this utility, transcript files must contain start/stop times. positional arguments: audio_file input audio file transcript transcript optional arguments: -h, --help show this help message and exit --target-dir TARGET_DIR Path to target directory
usage: prepare_audio_corpora [-h] [--target-dir TARGET_DIR] corpora [corpora ...] Copy and organize specified corpora into a target directory. Training, testing, and development sets will be created automatically if not already defined. positional arguments: corpora Name of one or more directories in directory this script is run optional arguments: -h, --help show this help message and exit --target-dir TARGET_DIR Path to target directory
This script scrapes a list of directories for paired STM and SPH files. If
dev folders are present, these labels are used for the output folder. By default, a target directory of 'input-data' will be created. Note that filenames with hyphens will be sanitized to underscores and that audio files will be forced to single channel, 16 kHz, signed PCM format. If two channels are present, only the first will be used.
usage: degrade_audio_file input_file1.wav input_file2.wav Degrade audio files to 8 kHz format similar to G711 codec
This script reduces audio quality of input audio files so that acoustic models can learn features from telephony with the G711 codec.
Note that the use of this function requires the separate installation of
pandas. This can be done via
pip install pandas.
usage: extract_excel_spreadsheets.py [-h] [--input-folder INPUT_FOLDER] [--output-corpus OUTPUT_CORPUS] convert a folder of excel spreadsheets to a corpus of text files optional arguments: -h, --help show this help message and exit --input-folder INPUT_FOLDER input folder of excel spreadsheets ending in .xls or .xlsx --output-corpus OUTPUT_CORPUS output folder for storing text corpus
- Python >= 3.5 with
Code of Conduct
Please make sure you read and observe our Code of Conduct.
Pull Request process
- Fork it
- Create your feature branch (
git checkout -b feature/fooBar)
- Commit your changes (
git commit -am 'Add some fooBar')
- Push to the branch (
git push origin feature/fooBar)
- Create a new Pull Request
NOTE: Commits and pull requests to FINOS repositories will only be accepted from those contributors with an active, executed Individual Contributor License Agreement (ICLA) with FINOS OR who are covered under an existing and active Corporate Contribution License Agreement (CCLA) executed with FINOS. Commits from individuals not covered under an ICLA or CCLA will be flagged and blocked by the FINOS Clabot tool. Please note that some CCLAs require individuals/employees to be explicitly named on the CCLA.
Need an ICLA? Unsure if you are covered under an existing CCLA? Email email@example.com
The code in this repository is distributed under the Apache License, Version 2.0.
Copyright 2020 GreenKey Technologies