Hey, this addresses a few things I ran into while trying to create a Dockerfile to package this all up.
1) When predicting, treat ambiguity codes in the reference as N's. This prevents an error that was thrown when encountering these.
2) Add a --seed argument to ftd-learn-dispersion-model and ftd-compute-deviation, to ensure that we get reproducible results.
3) Remove "scripts/ftd-update-dispersion-model" from setup.py, as the referenced script isn't in the repository.
Hey, this addresses a few things I ran into while trying to create a Dockerfile to package this all up.
1) When predicting, treat ambiguity codes in the reference as N's. This prevents an error that was thrown when encountering these. 2) Add a
--seed
argument to ftd-learn-dispersion-model and ftd-compute-deviation, to ensure that we get reproducible results. 3) Remove "scripts/ftd-update-dispersion-model" from setup.py, as the referenced script isn't in the repository.