Add option related to agrep match on obigrep and obiannotate

pkg/obitools/obiannotate/obiannotate.go

@@ -25,7 +25,7 @@ func DeleteAttributesWorker(toBeDeleted []string) obiseq.SeqWorker {
 	return f
 }
 
-func MatchPatternWorker(pattern, name string, errormax int, allowsIndel bool) obiseq.SeqWorker {
+func MatchPatternWorker(pattern, name string, errormax int, bothStrand, allowsIndel bool) obiseq.SeqWorker {
 	pat, err := obiapat.MakeApatPattern(pattern, errormax, allowsIndel)
 	if err != nil {
 		log.Fatalf("error in compiling pattern (%s) : %v", pattern, err)
@@ -56,7 +56,7 @@ func MatchPatternWorker(pattern, name string, errormax int, allowsIndel bool) ob
 
 		start, end, nerr, matched := pat.BestMatch(apats, 0, s.Len())
 
-		if matched {
+		if matched && start >= 0 && end <= s.Len() {
 			annot := s.Annotations()
 			annot[slot] = pattern
 
@@ -75,7 +75,7 @@ func MatchPatternWorker(pattern, name string, errormax int, allowsIndel bool) ob
 		} else {
 			start, end, nerr, matched := cpat.BestMatch(apats, 0, s.Len())
 
-			if matched {
+			if matched && start >= 0 && end <= s.Len() {
 				annot := s.Annotations()
 				annot[slot] = pattern
 				match, err := s.Subsequence(start, end, false)
@@ -328,9 +328,10 @@ func CLIAnnotationWorker() obiseq.SeqWorker {
 	}
 
 	if CLIHasPattern() {
-		log.Infof("Match pattern %s with %d error", CLIPattern(), CLIPatternError())
+		log.Infof("Match pattern %s with %d error", CLIPattern(), obigrep.CLIPatternError())
 		w := MatchPatternWorker(CLIPattern(), CLIHasPatternName(),
-			CLIPatternError(), CLIPatternInDels())
+			obigrep.CLIPatternError(), obigrep.CLIPatternBothStrand(),
+			obigrep.CLIPatternInDels())
 
 		annotator = annotator.ChainWorkers(w)
 	}

pkg/obitools/obiannotate/options.go

@@ -24,8 +24,6 @@ var _setSeqLength = false
 var _uniqueID = false
 var _ahoCorazick = ""
 var _pattern = ""
-var _pattern_error = 0
-var _pattern_indel = false
 var _pattern_name = "pattern"
 var _lcaSlot = ""
 var _lcaError = 0.0
@@ -62,14 +60,6 @@ func SequenceAnnotationOptionSet(options *getoptions.GetOpt) {
 		options.Description("specify the name to use as prefix for the slots reporting the match"),
 	)
 
-	options.IntVar(&_pattern_error, "pattern-error", _pattern_error,
-		options.Description("Maximum number of allowed error during pattern matching"),
-	)
-
-	options.BoolVar(&_pattern_indel, "allows-indels", _pattern_indel,
-		options.Description("Allows for indel during pattern matching"),
-	)
-
 	options.StringVar(&_lcaSlot, "add-lca-in", _lcaSlot,
 		options.ArgName("SLOT_NAME"),
 		options.Description("From the taxonomic annotation of the sequence (taxid slot or merged_taxid slot), "+
@@ -304,14 +294,6 @@ func CLIHasPatternName() string {
 	return _pattern_name
 }
 
-func CLIPatternError() int {
-	return _pattern_error
-}
-
-func CLIPatternInDels() bool {
-	return _pattern_indel
-}
-
 func CLISetTaxonomicPath() bool {
 	return _taxonomicPath
 }

pkg/obitools/obiconvert/options.go

@@ -19,6 +19,8 @@ var __input_fastobi_format__ = false
 var __input_ecopcr_format__ = false
 var __input_embl_format__ = false
 var __input_genbank_format__ = false
+var __input_fastq_format__ = false
+var __input_fasta_format__ = false
 
 var __output_in_fasta__ = false
 var __output_in_fastq__ = false
@@ -56,6 +58,12 @@ func InputOptionSet(options *getoptions.GetOpt) {
 	options.BoolVar(&__input_genbank_format__, "genbank", __input_genbank_format__,
 		options.Description("Read data following the Genbank flatfile format."))
 
+	options.BoolVar(&__input_fastq_format__, "fastq", __input_fastq_format__,
+		options.Description("Read data following the fastq format."))
+
+	options.BoolVar(&__input_fasta_format__, "fasta", __input_fasta_format__,
+		options.Description("Read data following the fasta format."))
+
 	options.BoolVar(&__no_ordered_input__, "no-order", __no_ordered_input__,
 		options.Description("When several input files are provided, "+
 			"indicates that there is no order among them."))
@@ -116,6 +124,10 @@ func OptionSet(options *getoptions.GetOpt) {
 // file has to be printed.
 func CLIInputFormat() string {
 	switch {
+	case __input_fasta_format__:
+		return "fasta"
+	case __input_fastq_format__:
+		return "fastq"
 	case __input_ecopcr_format__:
 		return "ecopcr"
 	case __input_embl_format__:

pkg/obitools/obiconvert/sequence_reader.go

@@ -138,6 +138,10 @@ func CLIReadBioSequences(filenames ...string) (obiiter.IBioSequence, error) {
 		}
 
 		switch CLIInputFormat() {
+		case "fastq":
+			reader = obiformats.ReadFastqFromFile
+		case "fasta":
+			reader = obiformats.ReadFastaFromFile
 		case "ecopcr":
 			reader = obiformats.ReadEcoPCRFromFile
 		case "embl":

pkg/obitools/obigrep/options.go

@@ -6,6 +6,7 @@ import (
 
 	log "github.com/sirupsen/logrus"
 
+	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiapat"
 	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiformats/ncbitaxdump"
 	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq"
 	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obitax"
@@ -43,6 +44,11 @@ var _SaveRejected = ""
 
 var _PairedMode = "forward"
 
+var _approx_pattern = make([]string, 0)
+var _pattern_error = 0
+var _pattern_indel = false
+var _pattern_only_forward = false
+
 func TaxonomySelectionOptionSet(options *getoptions.GetOpt) {
 
 	options.StringVar(&_Taxdump, "taxdump", _Taxdump,
@@ -143,6 +149,23 @@ func SequenceSelectionOptionSet(options *getoptions.GetOpt) {
 		options.Description("If paired reads are passed to obibrep, that option determines how the conditions "+
 			"are applied to both reads."),
 	)
+
+	options.StringSliceVar(&_approx_pattern, "approx-pattern", 1, 1,
+		options.ArgName("PATTERN"),
+		options.Description("Regular expression pattern to be tested against the sequence itself. The pattern is case insensitive."))
+
+	options.IntVar(&_pattern_error, "pattern-error", _pattern_error,
+		options.Description("Maximum number of allowed error during pattern matching"),
+	)
+
+	options.BoolVar(&_pattern_indel, "allows-indels", _pattern_indel,
+		options.Description("Allows for indel during pattern matching"),
+	)
+
+	options.BoolVar(&_pattern_only_forward, "only-forward", _pattern_only_forward,
+		options.Description("Look for pattern only on forward strand"),
+	)
+
 }
 
 // OptionSet adds to the basic option set every options declared for
@@ -212,6 +235,18 @@ func CLISelectedNCBITaxDump() string {
 	return _Taxdump
 }
 
+func CLIPatternError() int {
+	return _pattern_error
+}
+
+func CLIPatternInDels() bool {
+	return _pattern_indel
+}
+
+func CLIPatternBothStrand() bool {
+	return !_pattern_only_forward
+}
+
 func CLILoadSelectedTaxonomy() *obitax.Taxonomy {
 	if CLISelectedNCBITaxDump() != "" {
 		if _Taxonomy == nil {
@@ -327,6 +362,30 @@ func CLISequencePatternPredicate() obiseq.SequencePredicate {
 	return nil
 }
 
+func CLISequenceAgrep() obiseq.SequencePredicate {
+	if len(_approx_pattern) > 0 {
+		p := obiapat.IsPatternMatchSequence(
+			_approx_pattern[0],
+			CLIPatternError(),
+			CLIPatternBothStrand(),
+			CLIPatternInDels(),
+		)
+
+		for _, pattern := range _approx_pattern[1:] {
+			p = p.And(obiapat.IsPatternMatchSequence(
+				pattern,
+				CLIPatternError(),
+				CLIPatternBothStrand(),
+				CLIPatternInDels(),
+			))
+		}
+
+		return p
+	}
+
+	return nil
+}
+
 func CLIDefinitionPatternPredicate() obiseq.SequencePredicate {
 
 	if len(_DefinitionPatterns) > 0 {
@@ -419,6 +478,7 @@ func CLISequenceSelectionPredicate() obiseq.SequencePredicate {
 	p = p.And(CLIIdListPredicate())
 	p = p.And(CLIHasAttibutePredicate())
 	p = p.And(CLIIsAttibuteMatchPredicate())
+	p = p.And(CLISequenceAgrep())
 
 	if _InvertMatch {
 		p = p.Not()