From 97378283a55404b39426515033948fc9ff37547f Mon Sep 17 00:00:00 2001
From: Geo Pertea <geo.pertea@gmail.com>
Date: Fri, 5 Feb 2021 17:33:52 -0500
Subject: [PATCH] GRangeParser updated to support single coordinate

---
 GBase.cpp | 76 ++++++++++++++++++++++++++++++++-----------------------
 GBase.h   |  3 +--
 gff.cpp   |  7 +++--
 3 files changed, 51 insertions(+), 35 deletions(-)

diff --git a/GBase.cpp b/GBase.cpp
index 94ca5a9..f5148e7 100644
--- a/GBase.cpp
+++ b/GBase.cpp
@@ -254,14 +254,15 @@ FILE* Gfopen(const char *path, char *mode) {
 		GMessage("Error opening file '%s':  %s\n", path, strerror(errno));
 	return f;
 }
-
+#define IS_CHR_DELIM(c) ( c == ' ' || c == '\t' || c == ':' )
 void GRangeParser::parse(char* s) {
-	 //parses general range format: [[+/-]refID[+/-][ ]:]start[-/..]end[ :][+/-]
-	// if ref ID has ':' characters a space delimited format is accepted to separate
-	//    the ref ID from the coordinate range: [[+/-]ref<space>start[[-/..][end]]
-	//the safest way is to parse from the end in case the ref ID has ':' characters
+	 //parses general range format: [+\-|.]refID[+\-\.][ |:][start][-|.\s]end[\s:][+\-\.]
+	// if ref ID has ':' a space delimited format is preferred
+	//   or just separate the ref ID from the coordinate range: [[+/-]ref<space>start[[-..][end]]
+	//the safest way would be to parse from the end in case the ref ID has ':' characters
+	//if the whole chromosome is intended (no coordinates to speficy)
 	   this->start=0;
-	   this->end=MAX_UINT;
+	   this->end=0;
 	   this->strand=0;
 	   int slen=strlen(s);
 	   if (slen==0) return;
@@ -272,49 +273,62 @@ void GRangeParser::parse(char* s) {
 	     strand=c;
 	     ++s;slen--;
 	   }
-	   if (*s==':' || *s==' ') //ignore
+	   if (strand && (*s==':' || *s==' ')) //ignore
 		   { s++;slen--; }
 	   char* p=s; //parsing position for coordinate string
 	   char* isep=strpbrk(s, " \t");
 	   if (isep==NULL) isep=strchr(s, ':');
-	   if (isep) { //chr (ref) ID was given
+	   if (isep) { //chr (ref) ID ending found
 		  p=isep+1;
 		  *isep=0;
-		  char c=*(isep-1);
-		  if (strand==0 && (c=='+' || c=='-')) {
-		  //strand given after the ref ID
-		    isep--;
-		    strand=c;
-		    *isep=0; //ref is now parsable
+	       //character after the delimiter can only be a strand if it's followed by another delimiter
+		   //e.g. chr1 + 134551-204326 or chr1:+:134551-204326
+		  c=*(isep+1);
+		  if (strand==0 && (c=='+' || c=='-' || c=='.') && IS_CHR_DELIM(*(isep+2))) {
+			  strand=c;
+			  p=isep+3;
+		  }
+		  if (strand==0) {
+			  c=*(isep-1); //character before the delimiter could be the strand
+			  if (c=='+' || c=='-') { //not '.', sorry
+				isep--;
+				strand=c;
+				*isep=0; //ref is now parsable
+			  }
 		  }
 		  this->refName=Gstrdup(s,isep-1);
 	   }
-	   c=s[slen-1];
-	   if (c=='+' || c=='-' || c=='.') {
-		   strand=c;
-		   slen--;s[slen]=0;
-	   }
+	   //here we are after ref ID (and possibly strand) delimiter
 	   char* pend=p;
-	   while (isdigit(*pend)) pend++;
-	   c=*pend;
-	   *pend=0;
-	   this->start=atoi(p);
-	   p=pend;
-	   *p=c;
-	   while (*p=='-' || *p=='.' || isspace(*p)) ++p;
+	   if (isdigit(*pend)) {
+		 //parse the start coordinate then
+		 do { pend++; } while (isdigit(*pend));
+		 c=*pend;
+		 *pend=0;
+		 this->start=atoi(p);
+		 p=pend;
+		 *p=c;
+	   }
+	   while (*p=='-' || *p=='.' || *p==' ' || *p=='\t') ++p;
 	   pend=p;
 	   while (isdigit(*pend)) pend++;
-	   if (pend>=p) { //parse the 2nd coordinate
+	   if (pend>p) { //parse the 2nd coordinate
+		   c=*pend;
 		   *pend=0;
 		   this->end=atoi(p);
-		   if (this->end==0) this->end=MAX_UINT;
+		   *pend=c;
+	   }
+	   if (start && end && end<start) Gswap(this->start, this->end);
+	   //if (strand==0) { ?
+	   c=s[slen-1]; //peek at the end of the string for strand
+	   if (c=='+' || c=='-' || c=='.') {
+		   if (end || IS_CHR_DELIM(s[slen-2]))
+		   strand=c;
+		   //slen--;s[slen]=0;
 	   }
-	   if (this->end<this->start) Gswap(this->start, this->end);
 }
 
 
-
-
 bool GstrEq(const char* a, const char* b) {
 	 if (a==NULL || b==NULL) return false;
 	 return (strcmp(a, b)==0);
diff --git a/GBase.h b/GBase.h
index ea399c3..1ea4307 100644
--- a/GBase.h
+++ b/GBase.h
@@ -431,10 +431,9 @@ struct GSeg {
 
 struct GRangeParser: GSeg {
 	char* refName=NULL;
-	int gseq_id=-1;
 	char strand=0;
 	void parse(char* s);
-	GRangeParser(char* s=NULL):GSeg(0, INT_MAX) {
+	GRangeParser(char* s=NULL):GSeg(0, 0) {
 		if (s) parse(s);
 	}
 	~GRangeParser() {
diff --git a/gff.cpp b/gff.cpp
index 48d4f7e..df167d2 100644
--- a/gff.cpp
+++ b/gff.cpp
@@ -32,7 +32,7 @@ void gffnames_unref(GffNames* &n) {
 }
 
 const byte CLASSCODE_OVL_RANK = 14; //rank value just above 'o' class code
-//rank value < this means exon overlap 
+//rank value < this means exon overlap
 
 const byte CLASSCODE_J_RANK = 6; // all junctional based overlaps
 
@@ -415,13 +415,16 @@ GffLine::GffLine(GffReader* reader, const char* l): _parents(NULL), _parents_len
    line[i]=0;
    t[tidx]=line+i+1;
    tidx++;
-   if (tidx>8) break;
+   //if (tidx>8) break;
    }
   i++;
   }
  if (tidx<8) { // ignore non-GFF lines
   return;
  }
+ if (tidx>9) {
+	 GMessage("Warning: unexpected tab character in last column, line truncated:\n\%s\n",l);
+ }
  gffWarnings=reader->gff_warns;
  gseqname=t[0];
  track=t[1];