How to create a Python library using pylibtemplate
The following instructions assume that the reader understands completely the
purpose of each file in the pylibtemplate GitHub repository, in relation to the corresponding
Python library pylibtemplate.
Create a new empty GitHub repository for your new Python library
Before generating your git repository template for building your new Python
library, several other actions need to be performed, the first of which is to
create a new empty public GitHub repository, which will serve as the remote
repository for your new Python library. The GitHub repository should share the
same name as your new Python library, as it would appear in a pip install
command. We will refer to the name of your new Python library as
<your_lib_name>.
Create a new GitHub gist for your new Python library
We need to create a new GitHub gist that will be used to record the code coverage from your unit tests. To create a new gist, first go to https://gist.github.com. Next, in the field containing the placeholder text “Gist description” write:
To store the code coverage of the ``<your_lib_name>`` library.
Next, in the field containing the placeholder text “Filename including extension…”, write:
<your_lib_name>_coverage_badge.json
Next, in the field directly below — where the gist contents are suppose to be written — write:
{}
Next, click on the green down arrow button in the bottom right corner, select the “Create public gist” option, and then click the “Create public gist” button.
Upon creating the new gist, you will be redirected to a new page that presents
the gist. Copy or save somewhere the string of characters following the last /
in the URL of said page. This string of characters is the ID of the gist that
you just created. We will refer to the gist ID as <your_gist_id>.
Create a new GitHub token for your new Python library
Next, we need to create a new GitHub token with gist scope. To do this, first go to https://github.com/settings/tokens. Next, click on the button “Generate new token” and select the “Generate new token (classic)” option. Next, in the field directly below “Note”, write:
To access the ``<your_lib_name>`` code coverage file.
Next, in the field directly below “Expiration”, select whatever expiration date
that you prefer. Next, select the “gist” box amongst the options of scope. Next,
click on the “Generate token” button. Next, copy or save somewhere the personal
access token that appears on the current page. We will refer to the personal
access token as <your_access_token>.
Next, go to the main page of the <your_lib_name> GitHub repository. Next,
click on “Settings”, then “Secrets and variables” in the side menu that appears,
followed by “Actions” in the sub-menu. Next, click on the “New repository
secret” button. In the field containing the placeholder text “YOUR_SECRET_NAME”,
write:
CODE_COVERAGE_SECRET
In the field directly below “Secret *”, copy and paste <your_access_token>.
Create a new GitHub environment for your new Python library
In order to add your GitHub repository as a pending publisher on PyPI, you need
to create a specific GitHub environment for said repository. To do this, first
go to the main page of the <your_lib_name> GitHub repository. Next, click on
“Settings”, then “Environments” in the side menu that appears. Next, click on
the “New environment” button. In the field directly below “Name *”, write:
release
and then click on the “Configure environment” button.
Configure GitHub Pages for your new Python library
Next, we need to configure the GitHub Pages for your new Python library. To do
this, first go to the main page of the <your_lib_name> GitHub repository.
Next, click on “Settings”, then “Pages” in the side menu that appears. In the
field directly below “Source”, select “GitHub Actions”.
Create local git repository template using pylibtemplate
Next, we need to create a local git repository template using
pylibtemplate. To do this, activate an environment in which
pylibtemplate is installed, and then execute the following Python code
block:
import pylibtemplate
kwargs = {"lib_name_for_imports": \
<your_lib_name>,
"abbreviated_lib_name_for_docs": \
<your_abbreviated_lib_name_for_docs>,
"non_abbreviated_lib_name_for_docs": \
<your_non_abbreviated_lib_name_for_docs>,
"author": \
<author>,
"email": \
<email>,
"gist_id": \
<your_gist_id>,
"path_to_directory_to_contain_new_repo": \
<path_to_directory_to_contain_new_repo>}
pylibtemplate.generate_local_git_repo_template(**kwargs)
This code block will perform several actions: First, it will clone the git
commit of the pylibtemplate GitHub repository corresponding to the version
of pylibtemplate being used currently, in the directory at the path
<path_to_directory_to_contain_new_repo>, i.e. the git clone command is
executed while the working directory is set temporarily to the path
path_to_directory_to_contain_new_repo; Next, it will rename the cloned
repository to <your_lib_name> such that the path to the cloned repository
becomes <path_to_directory_to_contain_new_repo>/<your_lib_name>; Next, all
instances of the string of characters “pylibtemplate” are replaced with
<your_lib_name>, be it in file contents, directory basenames, or file
basenames; Next, all instances of the string of characters “PyLibTemplate” are
replaced with <your_abbreviated_lib_name_for_docs>; Next, all instances of
the string of characters “Python Library Template” are replaced with
<your_non_abbreviated_lib_name_for_docs>; Next, all email address
placeholders (i.e. instances of the string of characters
“matthew.rc.fitzpatrick@gmail.com”) are replaced with <email>; Next, all
instances of the gist ID of pylibtemplate are replaced with
<your_gist_id>; Next, all author placeholders (i.e. instances of the string
of characters “Matthew Fitzpatrick”) are replaced with <author>; Next, all
copyright statements are updated according to the current year; And lastly, the
following file is removed:
* ``<local_repo_root>/docs/how_to_create_a_python_library_using_pylibtemplate.rst``
where <local_repo_root> is the root of the local git repository, as well
as the following directory:
* ``<local_repo_root>/.git``
Instead of executing the above Python code block, one can achieve the same
result by running pylibtemplate as a command line too. To do this, activate
an environment in which pylibtemplate is installed, and then execute the
following command in the terminal:
pylibtemplate --lib_name_for_imports=<your_lib_name> --abbreviated_lib_name_for_docs=<your_abbreviated_lib_name_for_docs> --non_abbreviated_lib_name_for_docs=<your_non_abbreviated_lib_name_for_docs> --author=<author> --email=<email> --gist_id=<your_gist_id> --path_to_directory_to_contain_new_repo=<path_to_directory_to_contain_new_repo>
Add GitHub remote repository to local git repository
Once you have created your local git repository template, you should add to
it the GitHub repository that you created in the very first step above. To do
this, run the following commands in a terminal:
git init
git remote add origin https://github.com/<your_username>/<your_lib_name>.git
git branch -M main
git push -u origin main
where <your_username> is the name of the GitHub user that created the GitHub
repository.
Modify local repository files
Of course, now you must modify the local git repository files in order to
develop your new Python library.
The following files do not need to be modified under any circumstances:
* ``<local_repo_root>/.coveragerc``
* ``<local_repo_root>/tox.ini``
* ``<local_repo_root>/docs/Makefile``
* ``<local_repo_root>/docs/make.bat``
* ``<local_repo_root>/docs/api.rst``
* ``<local_repo_root>/docs/private_members_to_publish_to_docs.rst``
The following files may need to be modified to reflect the appropriate license
should it differ from that of pylibtemplate:
* ``<local_repo_root>/run_tests.sh``
* ``<local_repo_root>/setup.py``
* ``<local_repo_root>/LICENSE``
* ``<local_repo_root>/docs/license.rst``
* ``<local_repo_root>/docs/build_docs.py``
The following files may need to be modified if a custom installation procecdure is required to run your new Python library’s unit tests, that differs from the default installation procedure:
* ``<local_repo_root>/.github/workflows/measure_code_coverage.yml``
* ``<local_repo_root>/.github/workflows/test_library.yml``
* ``<local_repo_root>/.github/workflows/publish_documentation_website.yml``
* ``<local_repo_root>/.github/workflows/publish_release_to_pypi.yml``
The following files need to be modified according to the specifics of your new Python library:
* ``<local_repo_root>/README.md``
* ``<local_repo_root>/pyproject.toml``
* ``<local_repo_root>/.gitignore``
* ``<local_repo_root>/docs/INSTALL.rst``
* ``<local_repo_root>/docs/conf.py``
* ``<local_repo_root>/docs/index.rst``
* ``<local_repo_root>/docs/literature.rst``
* ``<local_repo_root>/docs/examples.rst``
as well as the files stored in the directories:
* ``<local_repo_root>/<your_lib_name>``
* ``<local_repo_root>/tests``
* ``<local_repo_root>/examples``
* ``<local_repo_root>/docs/examples``
After making the necessary modifications, you can proceed to test and debug your new Python library.