obitools/obitools4
2
0
Fork 0
mirror of https://github.com/metabarcoding/obitools4.git synced 2025-12-12 18:30:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Cleaning of obiminion

Browse Source 
Former-commit-id: 75148afd70e5006cc6855bcddc86506b099761a1
This commit is contained in:
Eric Coissac
2024-05-14 11:45:46 +02:00
parent 017030bcce
commit 7fcb0538a3
5 changed files with 11 additions and 219 deletions
Download Patch File Download Diff File Expand all files Collapse all files

128
pkg/obitools/obiconsensus/obiconsensus.go
View File

@@ -17,9 +17,7 @@ import (
func BuildConsensus(seqs obiseq.BioSequenceSlice,
consensus_id string,
kmer_size int, quorum float64,
min_depth float64,
max_length int,
kmer_size int,
save_graph bool, dirname string) (*obiseq.BioSequence, error) {
if save_graph {
@@ -37,7 +35,7 @@ func BuildConsensus(seqs obiseq.BioSequenceSlice,
}
}
fasta, err := os.Create(path.Join(dirname, fmt.Sprintf("%s.fasta", consensus_id)))
fasta, err := os.Create(path.Join(dirname, fmt.Sprintf("%s_consensus.fasta", consensus_id)))
if err == nil {
defer fasta.Close()
@@ -58,16 +56,7 @@ func BuildConsensus(seqs obiseq.BioSequenceSlice,
longest[i] = slices.Max(sa.CommonSuffix())
}
// o := obiutils.Order(sort.IntSlice(longest))
// i := int(float64(len(seqs)) * quorum)
// if i >= len(o) {
// i = len(o) - 1
// }
kmer_size = slices.Max(longest) + 1
// kmer_size = longest[o[i]] + 1
log.Printf("estimated kmer size : %d", kmer_size)
}
@@ -90,7 +79,7 @@ func BuildConsensus(seqs obiseq.BioSequenceSlice,
if save_graph {
file, err := os.Create(path.Join(dirname,
fmt.Sprintf("%s_raw_consensus.gml", consensus_id)))
fmt.Sprintf("%s_consensus.gml", consensus_id)))
if err != nil {
fmt.Println(err)
@@ -103,65 +92,7 @@ func BuildConsensus(seqs obiseq.BioSequenceSlice,
log.Printf("Graph size : %d\n", graph.Len())
total_kmer := graph.Len()
// threshold := 0
// switch {
// case min_depth < 0:
// spectrum := graph.WeightSpectrum()
// cum := make(map[int]int)
// spectrum[1] = 0
// for i := 2; i < len(spectrum); i++ {
// spectrum[i] += spectrum[i-1]
// cum[spectrum[i]]++
// }
// max := 0
// kmax := 0
// for k, obs := range cum {
// if obs > max {
// max = obs
// kmax = k
// }
// }
// for i, total := range spectrum {
// if total == kmax {
// threshold = i
// break
// }
// }
// threshold /= 2
// if threshold < 1 {
// threshold = 1
// }
// log.Info("Estimated kmer_min_occur = ", threshold)
// case min_depth >= 1:
// threshold = int(min_depth)
// default:
// threshold = int(float64(len(seqs)) * min_depth)
// }
// graph.FilterMinWeight(threshold)
// log.Printf("Graph size : %d\n", graph.Len())
// if save_graph {
// file, err := os.Create(path.Join(dirname,
// fmt.Sprintf("%s_consensus.gml", consensus_id)))
// if err != nil {
// fmt.Println(err)
// } else {
// file.WriteString(graph.Gml())
// file.Close()
// }
// }
seq, err := graph.LongestConsensus(consensus_id, max_length)
seq, err := graph.LongestConsensus(consensus_id)
sumCount := 0
@@ -173,7 +104,6 @@ func BuildConsensus(seqs obiseq.BioSequenceSlice,
seq.SetCount(sumCount)
seq.SetAttribute("seq_length", seq.Len())
seq.SetAttribute("kmer_size", kmer_size)
//seq.SetAttribute("kmer_min_occur", threshold)
seq.SetAttribute("kmer_max_occur", graph.MaxWeight())
seq.SetAttribute("filtered_graph_size", graph.Len())
seq.SetAttribute("full_graph_size", total_kmer)
@@ -181,52 +111,6 @@ func BuildConsensus(seqs obiseq.BioSequenceSlice,
return seq, err
}
// func BuildConsensusWithTimeout(seqs obiseq.BioSequenceSlice,
// kmer_size int, quorum float64,
// min_depth float64,
// save_graph bool, dirname string, timeout time.Duration) (*obiseq.BioSequence, error) {
// ctx, cancel := context.WithTimeout(context.Background(), timeout)
// defer cancel()
// consensus := func() *obiseq.BioSequence {
// cons, err := BuildConsensus(seqs, kmer_size, quorum, min_depth, save_graph, dirname,)
// if err != nil {
// cons = nil
// }
// return cons
// }
// computation := func() <-chan *obiseq.BioSequence {
// result := make(chan *obiseq.BioSequence)
// go func() {
// select {
// case <-ctx.Done():
// result <- nil
// default:
// result <- consensus()
// }
// }()
// return result
// }
// calcResult := computation()
// select {
// case result := <-calcResult:
// if result == nil {
// return nil, fmt.Errorf("cannot compute consensus")
// }
// return result, nil
// case <-ctx.Done():
// return nil, fmt.Errorf("compute consensus timeout, exiting")
// }
// }
func Consensus(iterator obiiter.IBioSequence) obiiter.IBioSequence {
newIter := obiiter.MakeIBioSequence()
size := 10
@@ -266,9 +150,7 @@ func Consensus(iterator obiiter.IBioSequence) obiiter.IBioSequence {
}
consensus, err := BuildConsensus(sequences,
id,
CLIKmerSize(), CLIThreshold(),
CLIKmerDepth(),
CLIMaxConsensusLength(),
CLIKmerSize(),
CLISaveGraphToFiles(),
CLIGraphFilesDirectory(),
)

35
pkg/obitools/obiconsensus/options.go
View File

@@ -7,9 +7,6 @@ import (
var _saveGraph = "__@@NOSAVE@@__"
var _kmerSize = -1
var _threshold = 0.99
var _mindepth = -1.0
var _consensus_max_length = -1
func ObiconsensusOptionSet(options *getoptions.GetOpt) {
@@ -25,26 +22,6 @@ func ObiconsensusOptionSet(options *getoptions.GetOpt) {
"Default value = -1, which means that the kmer size is estimated from the data"),
)
options.Float64Var(&_threshold, "threshold", _threshold,
options.ArgName("RATIO"),
options.Description("A threshold between O and 1 used to determine the optimal "+
"kmer size"),
)
options.Float64Var(&_mindepth, "min-depth", _mindepth,
options.ArgName("DEPTH"),
options.Description("if DEPTH is between 0 and 1, it corresponds to fraction of the "+
"reads in which a kmer must occurs to be conserved in the graph. If DEPTH is greater "+
"than 1, indicate the minimum count of occurrence for a kmer to be kept. "+
"Default value = -1, which means that the DEPTH is estimated from the data"),
)
options.IntVar(&_consensus_max_length, "consensus-max-length", _consensus_max_length,
options.ArgName("LENGTH"),
options.Description("Maximum length of the consensus sequence. "+
"Default value = -1, which means that no limit is applied"),
)
}
func OptionSet(options *getoptions.GetOpt) {
@@ -66,15 +43,3 @@ func CLIGraphFilesDirectory() string {
func CLIKmerSize() int {
return _kmerSize
}
func CLIKmerDepth() float64 {
return _mindepth
}
func CLIThreshold() float64 {
return _threshold
}
func CLIMaxConsensusLength() int {
return _consensus_max_length
}