Use an ignore file to avoid regenerating pieces instead of a separate update command.

[dbanty]

Is your feature request related to a problem? Please describe.
Several folks have indicated issues that could be solved by instructing openapi-python-client to ignore files. This could also greatly simplify the interface since the separate update command wouldn't be necessary, you'd just tell it to not update the files you don't want it to (e.g. README.md).

Describe the solution you'd like
Similar behavior to the openapitools generator where a .gitignore-style file will be automatically read from an existing directory and respected.

[dbanty]

The update command is gone, replaced by some options on generate. The only piece to this puzzle left is to give a good experience for skipping regeneration of certain pieces (probably, a .ignore file of some sort)

[sternj]

I think that the easiest thing to do would be to add in a preserve_files field in the config, seems like a naive implementation would create a temporary directory, move each of the preserve_files into it, do the regeneration, then move them back. It'd break files generated and then modified, but I'm not sure that's all that big a deal.

Another option could be to create a .manifest file in the generated directory and do the above for all files that don't match the gitignore and aren't in the manifest.

Any opinions one way or the other? I'm gonna make a PR when I have a few spare minutes

[rcoops]

Hi there.
We've previously been using this library to provide auto-generated artifacts within our own project structure. i.e. There are a few more files and modules that we want to be there.
Using the update command left anything of our own (that the generator didn't have knowledge of) intact, but the suggested alternative (generate and --overwrite) destroys anything that's not generated by the library.
Had a read through the docs but nothing leapt out as able to handle this - is there something we're missing?

[slovdahl]

In addition to version in pyproject.toml, it would be nice to be able to preserve other fields as well, e.g. authors.

[slovdahl]

And also e.g. tool.poetry.source, we use a proxy for pypi.org.