[question] Are large files only required for tests, or also required at build and run-time ?

[yurivict]

At one point the README says:

[...] large files used to build GATK, and test files required to run the test suite [...]

Does this mean that all production builds should contain large files?
Are they used by some gatk commands?

The README is just not clear about this.

[droazen]

@yurivict The large files under src/main/resources/large/ are required to build GATK, since they are packaged inside the GATK jar and used by tools at runtime. These are things like ML models and native C/C++ libraries used for acceleration of certain tools.

The large files under src/test/resources/large/, on the other hand, are only required by the test suite when running tests.

Hope this clears things up!

[yurivict]

This clarifies it, thank you!

Could you please also add this information to the README so that anybody would be able to read this in the future?

[droazen]

@yurivict There is a note in the README about this in the "Building GATK4" section:

Note that you must have a full git clone in order to build GATK, including the git-lfs files in src/main/resources. The zipped source code alone is not buildable.

However, I will edit it to make it clearer.

[ghost]

.