feat: Support for additional template variables #73
Conversation
Additional template variables for the Jinja template can be passed via CLI or can be defined in a configuration file. To pass variables via CLI, the option "--template-var KEY VALUE" can be given multiple times. To define variables in a configuration file, the variables can be defined in a "[template_vars]" table. [template_vars] KEY = "VALUE" All values are accessible in the template through the "template_vars" (dict) variable.
There was a problem hiding this comment.
Amazing feature once again! Thank you so much 😊
This all looks good, but I think I would prefer the following syntax on the CLI: --template-var key=value. We can do key_value.split("=", 1) so that values can have the = character too. What do you think?
Also, I don't think I like "template var" for the name and variable/parameter names in the code. What do you think of "jinja context" instead? This way we can also add a short option with -j which is not yet taken 🙂 I would then pass this context as context or user_context when rendering the template ("jinja" would be a bit redundant there).
Finally, it will need some documentation too! I can help with that if needed.
tests/test_cli.py
Outdated
| k1 = "ignored" | ||
| k2 = "v2" | ||
| k3 = "v3" | ||
| """.lstrip(), |
There was a problem hiding this comment.
lstrip does nothing here, since the string starts with [. You probably meant to use something like textwrap.dedent, to dedent every line?
I first thought, that having key and value as separate parameters, would make it a bit easier, when calling it from a shell script (if the value has to be escaped). I am fine with
Naming is not something I am especially good at 😁, so whatever you prefer is fine for me.
I can add some documentation (though help on that would be appreciated). |
|
I've pushed some docs 😄 Feel free to tweak them. |
src/git_changelog/cli.py
Outdated
| @@ -487,6 +521,7 @@ def build_and_render( | |||
| bump: str | None = None, | |||
| zerover: bool = True, # noqa: FBT001,FBT002 | |||
| filter_commits: str | None = None, | |||
| jinja_context: dict[str, str] | None = None, | |||
There was a problem hiding this comment.
It's true that values passed from the CLI will always be strings, but when using the API directly from Python, values could be anything, so I think it should be dict[str, Any].
tests/test_cli.py
Outdated
|
|
||
| assert contents == "k1 = v1\nk2 = v2\nk3 = v3\n" |
There was a problem hiding this comment.
| assert contents == "k1 = v1\nk2 = v2\nk3 = v3\n" |
Awesome, thanks! I only, did a small change in the config file example. I changed it now to use KEY=VALUE and I hope, I addressed all other comments as well. |
When using a custom Jinja template to render the changelog, it would be great, if additional context information could be accessed in that template. This feature request should be related to #17.
With the changes in this PR, additional context can be passed as key/value pairs either via CLI or via the toml config file.
To pass variables via CLI, the option
--template-var KEY VALUEcan be given multiple times.To define variables in a configuration file, the variables can be defined in a "[template_vars]" table.
When template vars are given via CLI and via config file, they are merged with the CLI vars taking precedence.
All values are accessible in the template through the "template_vars" (dict) variable.
The parsing of the CLI option into a dict is based on this gist: https://gist.github.com/vadimkantorov/37518ff88808af840884355c845049ea
(Unfortunately argparse does not support it out of the box, or at least I could not find a better way to achieve this)