From ec65f00cf2caaaebe0538e6ff9064bd631b71c61 Mon Sep 17 00:00:00 2001
From: Eric Coissac <eric.coissac@metabarcoding.org>
Date: Thu, 27 Jul 2017 16:05:30 +0200
Subject: [PATCH] Complete the fasta iterator to manage new input options

---
 python/obitools3/parsers/fasta.pyx | 87 +++++++++++++++++++++---------
 1 file changed, 63 insertions(+), 24 deletions(-)

diff --git a/python/obitools3/parsers/fasta.pyx b/python/obitools3/parsers/fasta.pyx
index 38ac738..b2f0b43 100644
--- a/python/obitools3/parsers/fasta.pyx
+++ b/python/obitools3/parsers/fasta.pyx
@@ -6,12 +6,14 @@ Created on 30 mars 2016
 @author: coissac
 '''
 
-#from obitools3.dms._obiseq cimport OBI_Seq
+from obitools3.dms.obiseq import Nuc_Seq
 
 
 def fastaIterator(lineiterator, 
-                  int buffersize=100000000,
-                  int skip=0
+                  int skip=0,
+                  only=None,
+                  firstline=None,
+                  int buffersize=100000000
                  ):
     cdef LineBuffer lb
     cdef str        ident
@@ -19,10 +21,14 @@ def fastaIterator(lineiterator,
     cdef dict       tags
     cdef list       s
     cdef bytes      sequence
-    cdef bytes      quality
-    cdef int        skipped
+    cdef int        skipped, ionly, read
 #    cdef OBI_Seq    seq
     
+    if only is None:
+        ionly=-1
+    else:
+        ionly=int(only)
+
     if isinstance(lineiterator,(str,bytes)):
         lineiterator=uopen(lineiterator)
     
@@ -33,9 +39,17 @@ def fastaIterator(lineiterator,
     
     skipped = 0
     i = iter(lb)
-    line = next(i)
+    
+    if firstline is None:
+        line = next(i)
+    else:
+        line = firstline
+        
     
     while True:
+        
+        if read >= ionly:
+            break
                 
         while skipped < skip :
             line = next(i)
@@ -59,7 +73,6 @@ def fastaIterator(lineiterator,
             pass
         
         sequence  = b"".join(s)
-        quality   = None
                 
     #         seq = OBI_Seq(id,
     #                       sequence,
@@ -70,26 +83,35 @@ def fastaIterator(lineiterator,
         yield { "id"         : ident,
                 "definition" : definition,
                 "sequence"   : sequence,
-                "quality"    : quality,
+                "quality"    : None,
+                "offset"     : None,
                 "tags"       : tags,
                 "annotation" : {}
               }
+        
+        read+=1
 
     
 def fastaNucIterator(lineiterator, 
-                     int buffersize=100000000,
-                     int skip=0
-                     ):
+                     int skip=0,
+                     only=None,
+                     firstline=None,
+                     int buffersize=100000000
+                    ):
     cdef LineBuffer lb
     cdef str        ident
     cdef str        definition
     cdef dict       tags
     cdef list       s
     cdef bytes      sequence
-    cdef bytes      quality
-    cdef int        skipped
+    cdef int        lines_to_skip, ionly, read
 #    cdef OBI_Seq    seq
     
+    if only is None:
+        ionly=-1
+    else:
+        ionly=int(only)
+
     if isinstance(lineiterator,(str,bytes)):
         lineiterator=uopen(lineiterator)
     
@@ -100,11 +122,19 @@ def fastaNucIterator(lineiterator,
 
 
     skipped = 0
+    read = 0
     i = iter(lb)
-    line = next(i)
     
+    if firstline is None:
+        line = next(i)
+    else:
+        line = firstline
+        
     while True:
         
+        if read >= ionly:
+            break
+       
         while skipped < skip :
             line = next(i)
             try:
@@ -125,18 +155,27 @@ def fastaNucIterator(lineiterator,
         except StopIteration:
             pass
         
-        sequence  = b"".join(s)
-        quality   = None
-        
+        sequence  = b"".join(s)        
         
 #        seq = 
-        yield { "id"         : ident,
-                "definition" : definition,
-                "sequence"   : sequence,
-                "quality"    : quality,
-                "tags"       : tags,
-                "annotation" : {}
-              }
+        seq = Nuc_Seq(ident,
+                      sequence,
+                      definition,
+                      None,-1,
+                      tags)
+        
+        yield seq
+        
+#         yield { "id"         : ident,
+#                 "definition" : definition,
+#                 "sequence"   : sequence,
+#                 "quality"    : None,
+#                 "offset"     : None,
+#                 "tags"       : tags,
+#                 "annotation" : {}
+#               }
+        
+        read+=1