This PR, used along with libROM PR#72, allows specifying an arbitrary number MV, ME of samples to be used in hyperreduction. It assumes a csv file format with the filenames specified by initSamples_basename + "V.csv" and initSamples_basename + "E.csv" placed in outputPath, where each row specifies a sample. The actual total number of samples NV, NE used is bounded by the options -sfacv, -sface or -nsamv, -nsame. Several cases in the usage (* stands for V or E):
initSamples_basename + "*.csv" is absent: use the greedy algorithm to sample the N* indices as usual
initSamples_basename + "*.csv" is present: read all the M* specified samples in the file. If N* > M*, use all the M* specified samples and use the greedy algorithm to sample the remainingN*-M* indices. Otherwise, use only the first N* specified samples and no greedy algorithm will be casted.
This PR, used along with libROM PR#72, allows specifying an arbitrary number
MV, MEof samples to be used in hyperreduction. It assumes a csv file format with the filenames specified byinitSamples_basename + "V.csv"andinitSamples_basename + "E.csv"placed inoutputPath, where each row specifies a sample. The actual total number of samplesNV, NEused is bounded by the options-sfacv, -sfaceor-nsamv, -nsame. Several cases in the usage (*stands forVorE):initSamples_basename + "*.csv"is absent: use the greedy algorithm to sample theN*indices as usualinitSamples_basename + "*.csv"is present: read all theM*specified samples in the file. IfN* > M*, use all theM*specified samples and use the greedy algorithm to sample the remainingN*-M*indices. Otherwise, use only the firstN*specified samples and no greedy algorithm will be casted.