Retrocookie updates Cookiecutter templates with changes from generated projects.
When developing Cookiecutter templates, you often need to work in a generated project rather than the template itself. Reasons for this include the following:
- You need to run the Continuous Integration suite for the generated project
- Your development tools choke when running on the templated project
Any changes you make in the generated project
need to be backported into the template,
carefully replacing expanded variables from cookiecutter.json
by templating tags,
and escaping any use of {{
and }}
or other tokens with special meaning in Jinja.
✨ Retrocookie helps you in this situation. ✨
It is designed to fetch commits from the repository of a generated project, and import them into your Cookiecutter repository, rewriting them on the fly to insert templating tags, escape Jinja-special constructs, and place files in the template directory.
Under the hood, Retrocookie rewrites the selected commits using git-filter-repo, saving them to a temporary repository. It then fetches and cherry-picks the rewritten commits from the temporary repository into the Cookiecutter template, using pygit2.
Maybe you're thinking, how can this possibly work? One cannot reconstruct a Jinja template from its rendered output. However, simple replacements of template variables work well in practice when you're only importing a handful of commits at a time.
Important:
Retrocookie relies on a .cookiecutter.json
file in the generated project
to work out how to rewrite commits.
This file is similar to the cookiecutter.json
file in the template,
but contains the specific values chosen during project generation.
You can generate this file by putting it into the template directory in the Cookiecutter,
with the following contents:
{{ cookiecutter | jsonify }}
- Python 3.8+
- git >= 2.22.0
You can install Retrocookie via pip from PyPI:
$ pip install retrocookie
Optionally, install the pr
extra for the retrocookie-pr command:
$ pip install retrocookie[pr]
Here's an example to demonstrate the general workflow.
To start with, we clone the repository of your Cookiecutter template. For this example, I'll use my own Hypermodern Python Cookiecutter.
$ git clone https://github.com/cjolowicz/cookiecutter-hypermodern-python
Next, we'll create a project from the template, and set up a git repository for it:
$ cookiecutter --no-input cookiecutter-hypermodern-python
$ cd hypermodern-python
$ git init
$ git add .
$ git commit --message="Initial commit"
Let's open a feature branch in the project repository, and make a fictitious change involving the default project name hypermodern-python:
$ git switch --create add-example
$ echo '# hypermodern-python' > EXAMPLE.md
$ git add EXAMPLE.md
$ git commit --message="Add example"
Back in the Cookiecutter repository, we can now invoke retrocookie to import the changes from the feature branch:
$ cd ../cookiecutter-hypermodern-python
$ retrocookie --branch add-example --create ../hypermodern-python
A git show
in the Cookiecutter shows the file under the template directory,
on a branch named as in the original repository,
with the project name replaced by a Jinja tag:
commit abb4f823b9f1760e3a678c927ec9797c0a40a9b6 (HEAD -> add-example)
Author: Your Name <your.name@example.com>
Date: Fri Dec 4 23:40:41 2020 +0100
Add example
diff --git a/{{cookiecutter.project_name}}/EXAMPLE.md b/{{cookiecutter.project_name}}/EXAMPLE.md
new file mode 100644
index 0000000..a158618
--- /dev/null
+++ b/{{cookiecutter.project_name}}/EXAMPLE.md
@@ -0,0 +1 @@
+# {{cookiecutter.project_name}}
The basic form:
$ retrocookie <repository> [<commits>...] $ retrocookie <repository> -b <branch> [--create]
The <repository>
is a filesystem path to the source repository.
For <commits>
, see gitrevisions(7).
Import HEAD
from <repository>
:
$ retrocookie <repository>
Import the last two commits:
$ retrocookie <repository> HEAD~2..
Import by commit hash:
$ retrocookie <repository> 53268f7 6a3368a c0b4c6c
Import commits from branch topic
:
$ retrocookie <repository> --branch=topic
Equivalently:
$ retrocookie <repository> master..topic
Import commits from topic
into a branch with the same name:
$ retrocookie <repository> --branch=topic --create
Equivalently, using short options:
$ retrocookie <repository> -cb topic
Import commits from branch topic
, which was branched off 1.0
:
$ retrocookie <repository> --branch=topic --upstream=1.0
Equivalently:
$ retrocookie <repository> 1.0..topic
Import HEAD
into a new branch topic
:
$ retrocookie <repository> --create-branch=topic
Please see the Command-line Reference for further details.
You can import pull requests from a generated project to the project template,
assuming their repositories are on GitHub.
This requires activating the pr
extra when installing with pip:
$ pip install retrocookie[pr]
The command retrocookie-pr
has the basic form:
$ retrocookie-pr [-R <repository>] [<pr>...] $ retrocookie-pr [-R <repository>] --user=<user> $ retrocookie-pr [-R <repository>] --all
Command-line arguments specify pull requests to import, by number or by branch. Pull requests from forks are currently not supported.
Use the -R <repository>
option to specify the GitHub repository of the generated project
from which the pull requests should be imported.
Provide the full name of the repository on GitHub in the form owner/name
.
The owner can be omitted if the repository is owned by the authenticated user.
This option can be omitted when the command is invoked from a local clone.
You can also select pull requests by specifying the user that opened them, via the --user
option.
This is handy for importing automated pull requests, such as dependency updates from Dependabot.
Use the --all
option to import all open pull requests in the generated project.
You can update previously imported pull requests by specifying --force
.
By default, retrocookie-pr
refuses to overwrite existing pull requests.
The command needs a personal access token to access the GitHub API.
(This token is also used to push to the GitHub repository of the project template.)
You will be prompted for the token when you invoke the command for the first time.
On subsequent invocations, the token is read from the application cache.
Alternatively, you can specify the token using the --token
option or the GITHUB_TOKEN
environment variable;
both of these methods bypass the cache.
Use the --open
option to open each imported pull request in a web browser.
Please see the Command-line Reference for further details.
Contributions are very welcome. To learn more, see the Contributor Guide.
Distributed under the terms of the MIT license, Retrocookie is free and open source software.
If you encounter any problems, please file an issue along with a detailed description.
This project was generated from @cjolowicz's Hypermodern Python Cookiecutter template.