obitools/obitools4
2
0
Fork 0
mirror of https://github.com/metabarcoding/obitools4.git synced 2026-03-26 05:50:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add max frequency filtering and top-kmer saving capabilities

Browse Source 
This commit introduces max frequency filtering to limit k-mer occurrences and adds functionality to save the N most frequent k-mers per set to CSV files. It also includes the ability to output k-mer frequency spectra as CSV and updates the CLI options accordingly.
This commit is contained in:
Eric Coissac
2026-02-10 09:26:46 +01:00
parent 56c1f4180c
commit f2937af1ad
7 changed files with 530 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
pkg/obitools/obik/index.go
View File

@@ -31,11 +31,19 @@ func runIndex(ctx context.Context, opt *getoptions.GetOpt, args []string) error
return fmt.Errorf("invalid min-occurrence: %d (must be >= 1)", minOcc)
}
maxOcc := CLIMaxOccurrence()
// Build options
var opts []obikmer.BuilderOption
if minOcc > 1 {
opts = append(opts, obikmer.WithMinFrequency(minOcc))
}
if maxOcc > 0 {
opts = append(opts, obikmer.WithMaxFrequency(maxOcc))
}
if topN := CLISaveFreqKmer(); topN > 0 {
opts = append(opts, obikmer.WithSaveFreqKmers(topN))
}
// Determine whether to append to existing group or create new
var builder *obikmer.KmerSetGroupBuilder
@@ -50,7 +58,11 @@ func runIndex(ctx context.Context, opt *getoptions.GetOpt, args []string) error
}
} else {
// New group
log.Infof("Creating new kmer index: k=%d, m=%d, min-occurrence=%d", k, m, minOcc)
if maxOcc > 0 {
log.Infof("Creating new kmer index: k=%d, m=%d, occurrence=[%d,%d]", k, m, minOcc, maxOcc)
} else {
log.Infof("Creating new kmer index: k=%d, m=%d, min-occurrence=%d", k, m, minOcc)
}
builder, err = obikmer.NewKmerSetGroupBuilder(outDir, k, m, 1, -1, opts...)
if err != nil {
return fmt.Errorf("failed to create kmer index builder: %w", err)
@@ -99,6 +111,9 @@ func runIndex(ctx context.Context, opt *getoptions.GetOpt, args []string) error
if minOcc > 1 {
ksg.SetAttribute("min_occurrence", minOcc)
}
if maxOcc > 0 {
ksg.SetAttribute("max_occurrence", maxOcc)
}
if err := ksg.SaveMetadata(); err != nil {
return fmt.Errorf("failed to save metadata: %w", err)