Support of incremental builds?

[arwedus]

I have a sphinx project with ~ 600 python files, and use autoapi to parse them.
On every build, AutoAPI spends around 1 minute in the reading phase, even though not a single file has changed.

What I'd hope for is that the reading of documents can be skipped if the cached data is newer than all files that AutoAPI has to parse.

Is this a known issue, or is this feature not yet available? Could this possibly work?

/edit: I was able to remove something else that seems to have caused sphinx incremental build to be compromised. Now AutoAPI reads through the documents, but much faster. So does it actually already check for every python file if it needs to be re-read?

[AWhetter]

AutoAPI doesn't check if Python files have already changed.
It would also need to check if an object that's in a changed file gets imported into any other files, and update those as well. But the import resolution is in need of an overhaul.