%global _empty_manifest_terminate_build 0 Name: python-praatio Version: 6.0.0 Release: 1 Summary: A library for working with praat, textgrids, time aligned audio transcripts, and audio files. License: LICENSE URL: https://github.com/timmahrt/praatIO Source0: https://mirrors.nju.edu.cn/pypi/web/packages/89/05/d93ad7beaa9ee6183151f3b5eb4d3a5a164f9869e8b9ab3bc6f8ff9522ad/praatio-6.0.0.tar.gz BuildArch: noarch Requires: python3-typing-extensions %description A library for working with praat, time aligned audio transcripts, and audio files *that comes with batteries included*. Praat uses a file format called textgrids, which are time aligned speech transcripts. This library isn't just a data struct for reading and writing textgrids--many utilities are provided to make it easy to work with with transcripts and associated audio files. This library also provides some other tools for use with praat. Praat is an open source software program for doing phonetic analysis and annotation of speech. [Praat can be downloaded here]() # Table of contents 1. [Documentation](#documentation) 2. [Tutorials](#tutorials) 3. [Version history](#version-history) 4. [Requirements](#requirements) 5. [Installation](#installation) 6. [Upgrading major versions](#upgrading) 7. [Usage](#usage) 8. [Common use cases](#common-use-cases) 9. [Output types](#output-types) 10. [Tests](#tests) 11. [Citing praatIO](#citing-praatio) 12. [Acknowledgements](#acknowledgements) ## Documentation Automatically generated pdocs can be found here: http://timmahrt.github.io/praatIO/ ## Tutorials There are tutorials available for learning how to use PraatIO. These are in the form of IPython Notebooks which can be found in the /tutorials/ folder distributed with PraatIO. You can view them online using the external website Jupyter: [Tutorial 1: An introduction and tutorial]() ## Version History *Praatio uses semantic versioning (Major.Minor.Patch)* Please view [CHANGELOG.md](https://github.com/timmahrt/praatIO/blob/main/CHANGELOG.md) for version history. ## Requirements Python module `https://pypi.org/project/typing-extensions/`. It should be installed automatically with praatio but you can install it manually if you have any problems. ``Python 3.7.*`` or above [Click here to visit travis-ci and see the specific versions of python that praatIO is currently tested under]() If you are using ``Python 2.x`` or ``Python < 3.7``, you can use `PraatIO 4.x`. ## Installation PraatIO is on pypi and can be installed or upgraded from the command-line shell with pip like so python -m pip install praatio --upgrade Otherwise, to manually install, after downloading the source from github, from a command-line shell, navigate to the directory containing setup.py and type python setup.py install If python is not in your path, you'll need to enter the full path e.g. C:\Python37\python.exe setup.py install ## Upgrading Please view [UPGRADING.md](https://github.com/timmahrt/praatIO/blob/main/UPGRADING.md) for detailed information about how to upgrade from earlier versions. ## Usage 99% of the time you're going to want to run ```python from praatio import textgrid tg = textgrid.openTextgrid(r"C:\Users\tim\Documents\transcript.TextGrid", False) ``` Or if you want to work with KlattGrid files ```python from praatio import klattgrid kg = klattgrid.openKlattGrid(r"C:\Users\tim\Documents\transcript.KlattGrid") ``` See /test for example usages ## Common Use Cases What can you do with this library? - query a textgrid to get information about the tiers or intervals contained within ```python tg = textgrid.openTextgrid("path_to_textgrid", False) entries = tg.tierDict["speaker_1_tier"].entries # Get all intervals entries = tg.tierDict["phone_tier"].find("a") # Get the indicies of all occurrences of 'a' ``` - create or augment textgrids using data from other sources - found that you clipped your audio file five seconds early and have added it back to your wavefile but now your textgrid is misaligned? Add five seconds to every interval in the textgrid ```python tg = textgrid.openTextgrid("path_to_textgrid", False) moddedTG = tg.editTimestamps(5) moddedTG.save('output_path_to_textgrid', 'long_textgrid', True) ``` - utilize the klattgrid interface to raise all speech formants by 20% ```python kg = klattgrid.openKlattGrid("path_to_klattgrid") incrTwenty = lambda x: x * 1.2 kg.tierDict["oral_formants"].modifySubtiers("formants",incrTwenty) kg.save(join(outputPath, "bobby_twenty_percent_less.KlattGrid")) ``` - replace labeled segments in a recording with silence or delete them - see /examples/deleteVowels.py - use set operations (union, intersection, difference) on textgrid tiers - see /examples/textgrid_set_operations.py - see /praatio/praatio_scripts.py for various ready-to-use functions such as - `splitAudioOnTier()`: split an audio file into chunks specified by intervals in one tier - `spellCheckEntries()`: spellcheck a textgrid tier - `tgBoundariesToZeroCrossings()`: adjust all boundaries and points to fall at the nearest zero crossing in the corresponding audio file - `alignBoundariesAcrossTiers()`: for handmade textgrids, sometimes entries may look as if they are aligned at the same time but actually are off by a small amount, this will correct them ## Output types PraatIO supports 4 textgrid output file types: short textgrid, long textgrid, json, and textgrid-like json. Short textgrids and long textgrids are both formats that are natively supported by praat. Short textgrids are meant to be more concise while long textgrids are meant to be more human-readable. For more information on these file formats, please see [praat's official documentation](https://www.fon.hum.uva.nl/praat/manual/TextGrid_file_formats.html) JSON and textgrid-like JSON are more developer-friendly formats, but they are not supported by praat. The default JSON format is more minimal while the textgrid-like JSON is formatted with information similar to a textgrid file. The default JSON format does not support one use-case: a textgrid has a specified minimum and maximum timestamp. The textgrid's tiers also have a specified minimum and maximum timestamp. Under most circumstances, they are the same, but the user can specify them to be different and praat will respect this. If you have such textgrids, you should use the textgrid-like JSON. Here is the schema for the JSON output file: ``` { "start": 0.0, "end": 1.8, "tiers": { "phone": { "type": "IntervalTier", "entries": [[0.0, 0.3, ""], [0.3, 0.38, "m"]] }, "pitch": { "type": "TextTier", "entries": [[0.32, "120"], [0.37, "85"]] } } } ``` Here is the schema for the Textgrid-like JSON output file. Notably, `tiers` is a list of hashes, rather than a hash of hashes. Also, each tier specifies it's name, and a min and max time. ``` { "xmin": 0.0, "xmax": 1.8, "tiers": [ { "class": "IntervalTier", "name": "phone", "xmin": 0.0, "xmax": 1.8, "entries": [[0.0, 0.3, ""], [0.3, 0.38, "m"]] }, { "class": "TextTier", "name": "pitch", "xmin": 0.0, "xmax": 1.8, "entries": [[0.32, "120"], [0.37, "85"]] } ] } ``` ## Citing praatIO PraatIO is general purpose coding and doesn't need to be cited but if you would like to, it can be cited like so: Tim Mahrt. PraatIO. https://github.com/timmahrt/praatIO, 2016. ## Acknowledgements Development of PraatIO was possible thanks to NSF grant **BCS 12-51343** to Jennifer Cole, José I. Hualde, and Caroline Smith and to the A\*MIDEX project (n° **ANR-11-IDEX-0001-02**) to James Sneed German funded by the Investissements d'Avenir French Government program, managed by the French National Research Agency (ANR). %package -n python3-praatio Summary: A library for working with praat, textgrids, time aligned audio transcripts, and audio files. Provides: python-praatio BuildRequires: python3-devel BuildRequires: python3-setuptools BuildRequires: python3-pip %description -n python3-praatio A library for working with praat, time aligned audio transcripts, and audio files *that comes with batteries included*. Praat uses a file format called textgrids, which are time aligned speech transcripts. This library isn't just a data struct for reading and writing textgrids--many utilities are provided to make it easy to work with with transcripts and associated audio files. This library also provides some other tools for use with praat. Praat is an open source software program for doing phonetic analysis and annotation of speech. [Praat can be downloaded here]() # Table of contents 1. [Documentation](#documentation) 2. [Tutorials](#tutorials) 3. [Version history](#version-history) 4. [Requirements](#requirements) 5. [Installation](#installation) 6. [Upgrading major versions](#upgrading) 7. [Usage](#usage) 8. [Common use cases](#common-use-cases) 9. [Output types](#output-types) 10. [Tests](#tests) 11. [Citing praatIO](#citing-praatio) 12. [Acknowledgements](#acknowledgements) ## Documentation Automatically generated pdocs can be found here: http://timmahrt.github.io/praatIO/ ## Tutorials There are tutorials available for learning how to use PraatIO. These are in the form of IPython Notebooks which can be found in the /tutorials/ folder distributed with PraatIO. You can view them online using the external website Jupyter: [Tutorial 1: An introduction and tutorial]() ## Version History *Praatio uses semantic versioning (Major.Minor.Patch)* Please view [CHANGELOG.md](https://github.com/timmahrt/praatIO/blob/main/CHANGELOG.md) for version history. ## Requirements Python module `https://pypi.org/project/typing-extensions/`. It should be installed automatically with praatio but you can install it manually if you have any problems. ``Python 3.7.*`` or above [Click here to visit travis-ci and see the specific versions of python that praatIO is currently tested under]() If you are using ``Python 2.x`` or ``Python < 3.7``, you can use `PraatIO 4.x`. ## Installation PraatIO is on pypi and can be installed or upgraded from the command-line shell with pip like so python -m pip install praatio --upgrade Otherwise, to manually install, after downloading the source from github, from a command-line shell, navigate to the directory containing setup.py and type python setup.py install If python is not in your path, you'll need to enter the full path e.g. C:\Python37\python.exe setup.py install ## Upgrading Please view [UPGRADING.md](https://github.com/timmahrt/praatIO/blob/main/UPGRADING.md) for detailed information about how to upgrade from earlier versions. ## Usage 99% of the time you're going to want to run ```python from praatio import textgrid tg = textgrid.openTextgrid(r"C:\Users\tim\Documents\transcript.TextGrid", False) ``` Or if you want to work with KlattGrid files ```python from praatio import klattgrid kg = klattgrid.openKlattGrid(r"C:\Users\tim\Documents\transcript.KlattGrid") ``` See /test for example usages ## Common Use Cases What can you do with this library? - query a textgrid to get information about the tiers or intervals contained within ```python tg = textgrid.openTextgrid("path_to_textgrid", False) entries = tg.tierDict["speaker_1_tier"].entries # Get all intervals entries = tg.tierDict["phone_tier"].find("a") # Get the indicies of all occurrences of 'a' ``` - create or augment textgrids using data from other sources - found that you clipped your audio file five seconds early and have added it back to your wavefile but now your textgrid is misaligned? Add five seconds to every interval in the textgrid ```python tg = textgrid.openTextgrid("path_to_textgrid", False) moddedTG = tg.editTimestamps(5) moddedTG.save('output_path_to_textgrid', 'long_textgrid', True) ``` - utilize the klattgrid interface to raise all speech formants by 20% ```python kg = klattgrid.openKlattGrid("path_to_klattgrid") incrTwenty = lambda x: x * 1.2 kg.tierDict["oral_formants"].modifySubtiers("formants",incrTwenty) kg.save(join(outputPath, "bobby_twenty_percent_less.KlattGrid")) ``` - replace labeled segments in a recording with silence or delete them - see /examples/deleteVowels.py - use set operations (union, intersection, difference) on textgrid tiers - see /examples/textgrid_set_operations.py - see /praatio/praatio_scripts.py for various ready-to-use functions such as - `splitAudioOnTier()`: split an audio file into chunks specified by intervals in one tier - `spellCheckEntries()`: spellcheck a textgrid tier - `tgBoundariesToZeroCrossings()`: adjust all boundaries and points to fall at the nearest zero crossing in the corresponding audio file - `alignBoundariesAcrossTiers()`: for handmade textgrids, sometimes entries may look as if they are aligned at the same time but actually are off by a small amount, this will correct them ## Output types PraatIO supports 4 textgrid output file types: short textgrid, long textgrid, json, and textgrid-like json. Short textgrids and long textgrids are both formats that are natively supported by praat. Short textgrids are meant to be more concise while long textgrids are meant to be more human-readable. For more information on these file formats, please see [praat's official documentation](https://www.fon.hum.uva.nl/praat/manual/TextGrid_file_formats.html) JSON and textgrid-like JSON are more developer-friendly formats, but they are not supported by praat. The default JSON format is more minimal while the textgrid-like JSON is formatted with information similar to a textgrid file. The default JSON format does not support one use-case: a textgrid has a specified minimum and maximum timestamp. The textgrid's tiers also have a specified minimum and maximum timestamp. Under most circumstances, they are the same, but the user can specify them to be different and praat will respect this. If you have such textgrids, you should use the textgrid-like JSON. Here is the schema for the JSON output file: ``` { "start": 0.0, "end": 1.8, "tiers": { "phone": { "type": "IntervalTier", "entries": [[0.0, 0.3, ""], [0.3, 0.38, "m"]] }, "pitch": { "type": "TextTier", "entries": [[0.32, "120"], [0.37, "85"]] } } } ``` Here is the schema for the Textgrid-like JSON output file. Notably, `tiers` is a list of hashes, rather than a hash of hashes. Also, each tier specifies it's name, and a min and max time. ``` { "xmin": 0.0, "xmax": 1.8, "tiers": [ { "class": "IntervalTier", "name": "phone", "xmin": 0.0, "xmax": 1.8, "entries": [[0.0, 0.3, ""], [0.3, 0.38, "m"]] }, { "class": "TextTier", "name": "pitch", "xmin": 0.0, "xmax": 1.8, "entries": [[0.32, "120"], [0.37, "85"]] } ] } ``` ## Citing praatIO PraatIO is general purpose coding and doesn't need to be cited but if you would like to, it can be cited like so: Tim Mahrt. PraatIO. https://github.com/timmahrt/praatIO, 2016. ## Acknowledgements Development of PraatIO was possible thanks to NSF grant **BCS 12-51343** to Jennifer Cole, José I. Hualde, and Caroline Smith and to the A\*MIDEX project (n° **ANR-11-IDEX-0001-02**) to James Sneed German funded by the Investissements d'Avenir French Government program, managed by the French National Research Agency (ANR). %package help Summary: Development documents and examples for praatio Provides: python3-praatio-doc %description help A library for working with praat, time aligned audio transcripts, and audio files *that comes with batteries included*. Praat uses a file format called textgrids, which are time aligned speech transcripts. This library isn't just a data struct for reading and writing textgrids--many utilities are provided to make it easy to work with with transcripts and associated audio files. This library also provides some other tools for use with praat. Praat is an open source software program for doing phonetic analysis and annotation of speech. [Praat can be downloaded here]() # Table of contents 1. [Documentation](#documentation) 2. [Tutorials](#tutorials) 3. [Version history](#version-history) 4. [Requirements](#requirements) 5. [Installation](#installation) 6. [Upgrading major versions](#upgrading) 7. [Usage](#usage) 8. [Common use cases](#common-use-cases) 9. [Output types](#output-types) 10. [Tests](#tests) 11. [Citing praatIO](#citing-praatio) 12. [Acknowledgements](#acknowledgements) ## Documentation Automatically generated pdocs can be found here: http://timmahrt.github.io/praatIO/ ## Tutorials There are tutorials available for learning how to use PraatIO. These are in the form of IPython Notebooks which can be found in the /tutorials/ folder distributed with PraatIO. You can view them online using the external website Jupyter: [Tutorial 1: An introduction and tutorial]() ## Version History *Praatio uses semantic versioning (Major.Minor.Patch)* Please view [CHANGELOG.md](https://github.com/timmahrt/praatIO/blob/main/CHANGELOG.md) for version history. ## Requirements Python module `https://pypi.org/project/typing-extensions/`. It should be installed automatically with praatio but you can install it manually if you have any problems. ``Python 3.7.*`` or above [Click here to visit travis-ci and see the specific versions of python that praatIO is currently tested under]() If you are using ``Python 2.x`` or ``Python < 3.7``, you can use `PraatIO 4.x`. ## Installation PraatIO is on pypi and can be installed or upgraded from the command-line shell with pip like so python -m pip install praatio --upgrade Otherwise, to manually install, after downloading the source from github, from a command-line shell, navigate to the directory containing setup.py and type python setup.py install If python is not in your path, you'll need to enter the full path e.g. C:\Python37\python.exe setup.py install ## Upgrading Please view [UPGRADING.md](https://github.com/timmahrt/praatIO/blob/main/UPGRADING.md) for detailed information about how to upgrade from earlier versions. ## Usage 99% of the time you're going to want to run ```python from praatio import textgrid tg = textgrid.openTextgrid(r"C:\Users\tim\Documents\transcript.TextGrid", False) ``` Or if you want to work with KlattGrid files ```python from praatio import klattgrid kg = klattgrid.openKlattGrid(r"C:\Users\tim\Documents\transcript.KlattGrid") ``` See /test for example usages ## Common Use Cases What can you do with this library? - query a textgrid to get information about the tiers or intervals contained within ```python tg = textgrid.openTextgrid("path_to_textgrid", False) entries = tg.tierDict["speaker_1_tier"].entries # Get all intervals entries = tg.tierDict["phone_tier"].find("a") # Get the indicies of all occurrences of 'a' ``` - create or augment textgrids using data from other sources - found that you clipped your audio file five seconds early and have added it back to your wavefile but now your textgrid is misaligned? Add five seconds to every interval in the textgrid ```python tg = textgrid.openTextgrid("path_to_textgrid", False) moddedTG = tg.editTimestamps(5) moddedTG.save('output_path_to_textgrid', 'long_textgrid', True) ``` - utilize the klattgrid interface to raise all speech formants by 20% ```python kg = klattgrid.openKlattGrid("path_to_klattgrid") incrTwenty = lambda x: x * 1.2 kg.tierDict["oral_formants"].modifySubtiers("formants",incrTwenty) kg.save(join(outputPath, "bobby_twenty_percent_less.KlattGrid")) ``` - replace labeled segments in a recording with silence or delete them - see /examples/deleteVowels.py - use set operations (union, intersection, difference) on textgrid tiers - see /examples/textgrid_set_operations.py - see /praatio/praatio_scripts.py for various ready-to-use functions such as - `splitAudioOnTier()`: split an audio file into chunks specified by intervals in one tier - `spellCheckEntries()`: spellcheck a textgrid tier - `tgBoundariesToZeroCrossings()`: adjust all boundaries and points to fall at the nearest zero crossing in the corresponding audio file - `alignBoundariesAcrossTiers()`: for handmade textgrids, sometimes entries may look as if they are aligned at the same time but actually are off by a small amount, this will correct them ## Output types PraatIO supports 4 textgrid output file types: short textgrid, long textgrid, json, and textgrid-like json. Short textgrids and long textgrids are both formats that are natively supported by praat. Short textgrids are meant to be more concise while long textgrids are meant to be more human-readable. For more information on these file formats, please see [praat's official documentation](https://www.fon.hum.uva.nl/praat/manual/TextGrid_file_formats.html) JSON and textgrid-like JSON are more developer-friendly formats, but they are not supported by praat. The default JSON format is more minimal while the textgrid-like JSON is formatted with information similar to a textgrid file. The default JSON format does not support one use-case: a textgrid has a specified minimum and maximum timestamp. The textgrid's tiers also have a specified minimum and maximum timestamp. Under most circumstances, they are the same, but the user can specify them to be different and praat will respect this. If you have such textgrids, you should use the textgrid-like JSON. Here is the schema for the JSON output file: ``` { "start": 0.0, "end": 1.8, "tiers": { "phone": { "type": "IntervalTier", "entries": [[0.0, 0.3, ""], [0.3, 0.38, "m"]] }, "pitch": { "type": "TextTier", "entries": [[0.32, "120"], [0.37, "85"]] } } } ``` Here is the schema for the Textgrid-like JSON output file. Notably, `tiers` is a list of hashes, rather than a hash of hashes. Also, each tier specifies it's name, and a min and max time. ``` { "xmin": 0.0, "xmax": 1.8, "tiers": [ { "class": "IntervalTier", "name": "phone", "xmin": 0.0, "xmax": 1.8, "entries": [[0.0, 0.3, ""], [0.3, 0.38, "m"]] }, { "class": "TextTier", "name": "pitch", "xmin": 0.0, "xmax": 1.8, "entries": [[0.32, "120"], [0.37, "85"]] } ] } ``` ## Citing praatIO PraatIO is general purpose coding and doesn't need to be cited but if you would like to, it can be cited like so: Tim Mahrt. PraatIO. https://github.com/timmahrt/praatIO, 2016. ## Acknowledgements Development of PraatIO was possible thanks to NSF grant **BCS 12-51343** to Jennifer Cole, José I. Hualde, and Caroline Smith and to the A\*MIDEX project (n° **ANR-11-IDEX-0001-02**) to James Sneed German funded by the Investissements d'Avenir French Government program, managed by the French National Research Agency (ANR). %prep %autosetup -n praatio-6.0.0 %build %py3_build %install %py3_install install -d -m755 %{buildroot}/%{_pkgdocdir} if [ -d doc ]; then cp -arf doc %{buildroot}/%{_pkgdocdir}; fi if [ -d docs ]; then cp -arf docs %{buildroot}/%{_pkgdocdir}; fi if [ -d example ]; then cp -arf example %{buildroot}/%{_pkgdocdir}; fi if [ -d examples ]; then cp -arf examples %{buildroot}/%{_pkgdocdir}; fi pushd %{buildroot} if [ -d usr/lib ]; then find usr/lib -type f -printf "/%h/%f\n" >> filelist.lst fi if [ -d usr/lib64 ]; then find usr/lib64 -type f -printf "/%h/%f\n" >> filelist.lst fi if [ -d usr/bin ]; then find usr/bin -type f -printf "/%h/%f\n" >> filelist.lst fi if [ -d usr/sbin ]; then find usr/sbin -type f -printf "/%h/%f\n" >> filelist.lst fi touch doclist.lst if [ -d usr/share/man ]; then find usr/share/man -type f -printf "/%h/%f.gz\n" >> doclist.lst fi popd mv %{buildroot}/filelist.lst . mv %{buildroot}/doclist.lst . %files -n python3-praatio -f filelist.lst %dir %{python3_sitelib}/* %files help -f doclist.lst %{_docdir}/* %changelog * Wed May 10 2023 Python_Bot - 6.0.0-1 - Package Spec generated