tkdagdelen / gnocchi_old Goto Github PK

• Clean up Gnocchi

getting a FileNotFound error after rebooting the way we grab the input files.

Adding identity by state clustering using Frank's similarity code

The ration of training to test data should be a user-provided flag.

if the -overwriteParquet is fed into Gnocchi, it should check for the parquetFiles directory and delete it before trying to write new parquet files. Right now, for some reason, when the parquetFiles directory is within HDFS, it isn't getting deleted.

@tkdagdelen LoadPhenotypesWithCovariates is handling covariates incorrectly, probably change
val joinedData = data.cogroup(covarData).map(pair => {
need to ignore rows that don't have a covariate in the covariate file

#8