obi cat: fixed open file descriptor leak and switch to version 3.0.0b34

3.0.0b33

python/obitools3/commands/cat.pyx

@@ -4,7 +4,7 @@ from obitools3.apps.progress cimport ProgressBar  # @UnresolvedImport
 from obitools3.dms import DMS
 from obitools3.dms.view.view cimport View
 from obitools3.uri.decode import open_uri
-from obitools3.apps.optiongroups import addMinimalOutputOption
+from obitools3.apps.optiongroups import addMinimalOutputOption, addNoProgressBarOption
 from obitools3.dms.view import RollbackException
 from obitools3.apps.config import logger
 from obitools3.utils cimport str2bytes
@@ -28,6 +28,7 @@ __title__="Concatenate views."
 def addOptions(parser):
     
     addMinimalOutputOption(parser)
+    addNoProgressBarOption(parser)
 
     group=parser.add_argument_group('obi cat specific options')
 
@@ -47,9 +48,9 @@ def run(config):
     
     logger("info", "obi cat")
 
-    # Open the views to concatenate
+    # Check the views to concatenate
-    iview_list = []
     idms_list = []
+    iview_list = []
     total_len = 0
     remove_qual = False
     remove_rev_qual = False
@@ -67,8 +68,9 @@ def run(config):
         if REVERSE_QUALITY_COLUMN not in i_view: # same as above for reverse quality
             remove_rev_qual = True
         total_len += len(i_view)
-        iview_list.append(i_view)
         idms_list.append(i_dms)
+        iview_list.append(i_view.name)
+        i_view.close()
 
     # Open the output: only the DMS
     output = open_uri(config['obi']['outputURI'],
@@ -97,8 +99,10 @@ def run(config):
     # Initialize multiple elements columns
     if type(output_0)==BufferedWriter: 
         dict_cols = {}
-        for v in iview_list:
+        for v_uri in config["cat"]["views_to_cat"]:
+            v = open_uri(v_uri)[1]
             for coln in v.keys():
+                col = v[coln]
                 if v[coln].nb_elements_per_line > 1:
                     if coln not in dict_cols:
                         dict_cols[coln] = {}
@@ -108,6 +112,7 @@ def run(config):
                     else:
                         dict_cols[coln]['eltnames'] = set(v[coln].elements_names + list(dict_cols[coln]['eltnames']))
                         dict_cols[coln]['nbelts'] = len(dict_cols[coln]['eltnames'])
+            v.close()            
         for coln in dict_cols:
             Column.new_column(o_view, coln, dict_cols[coln]['obitype'], 
                               nb_elements_per_line=dict_cols[coln]['nbelts'], elements_names=list(dict_cols[coln]['eltnames']))
@@ -119,7 +124,8 @@ def run(config):
         pb = None
     
     i = 0
-    for v in iview_list:
+    for v_uri in config["cat"]["views_to_cat"]:
+        v = open_uri(v_uri)[1]
         for entry in v:
             PyErr_CheckSignals()
             if pb is not None:
@@ -130,6 +136,7 @@ def run(config):
             else:
                 o_view[i] = entry
             i+=1
+        v.close()
 
     # Deletes quality columns if needed
     if type(output_0)!=BufferedWriter: 
@@ -144,7 +151,7 @@ def run(config):
     
     # Save command config in DMS comments
     command_line = " ".join(sys.argv[1:])
-    o_view.write_config(config, "cat", command_line, input_dms_name=[d.name for d in idms_list], input_view_name=[v.name for v in iview_list])
+    o_view.write_config(config, "cat", command_line, input_dms_name=[d.name for d in idms_list], input_view_name=[vname for vname in iview_list])
     o_dms.record_command_line(command_line)
 
     #print("\n\nOutput view:\n````````````", file=sys.stderr)

python/obitools3/commands/uniq.pyx

@@ -354,6 +354,9 @@ cdef uniq_sequences(View_NUC_SEQS view, View_NUC_SEQS o_view, ProgressBar pb, di
         key = mergedKeys[k]
         merged_col_name = mergedKeys_m[k]
         
+        if merged_infos[merged_col_name]['nb_elts'] == 1:
+            raise Exception("Can't merge information from a tag with only one element (e.g. one sample ; don't use -m option)")
+        
         if merged_col_name in view:
             i_col = view[merged_col_name]
         else:

python/obitools3/format/tab.pyx

@@ -5,6 +5,7 @@ from obitools3.dms.view.view cimport Line
 from obitools3.utils cimport bytes2str_object, str2bytes, tobytes
 from obitools3.dms.column.column cimport Column_line, Column_multi_elts
 
+import sys
 
 cdef class TabFormat:
     
@@ -26,18 +27,22 @@ cdef class TabFormat:
             
             if self.header and self.first_line:
                 if isinstance(data.view[k], Column_multi_elts):
-                    for k2 in data.view[k].keys():
+                    keys = data.view[k].keys()
+                    keys.sort()
+                    for k2 in keys:
                         line.append(tobytes(k)+b':'+tobytes(k2))
                 else:
                     line.append(tobytes(k))
             else:
                 value = data[k]
                 if isinstance(data.view[k], Column_multi_elts):
+                    keys = data.view[k].keys()
+                    keys.sort()
                     if value is None:  # all keys at None
-                        for k2 in data.view[k].keys(): # TODO could be much more efficient
+                        for k2 in keys: # TODO could be much more efficient
                             line.append(self.NAString)
                     else:
-                        for k2 in data.view[k].keys(): # TODO could be much more efficient
+                        for k2 in keys: # TODO could be much more efficient
                             if value[k2] is not None:
                                 line.append(str2bytes(str(bytes2str_object(value[k2]))))  # genius programming
                             else:

python/obitools3/utils.pxd

@@ -2,7 +2,7 @@
 
 from obitools3.dms.capi.obitypes cimport obitype_t, index_t
 
-cpdef bytes format_separator(bytes format)
+cpdef bytes format_uniq_pattern(bytes format)
 cpdef int count_entries(file, bytes format)
 
 cdef obi_errno_to_exception(index_t line_nb=*, object elt_id=*, str error_message=*)

python/obitools3/utils.pyx

@@ -24,11 +24,11 @@ import glob
 import gzip
 
 
-cpdef bytes format_separator(bytes format):
+cpdef bytes format_uniq_pattern(bytes format):
     if format == b"fasta":
         return b"\n>"
     elif format == b"fastq":
-        return b"\n@"
+        return b"\n\+\n"
     elif format == b"ngsfilter" or format == b"tabular":
         return b"\n"
     elif format == b"genbank" or format == b"embl":
@@ -42,7 +42,7 @@ cpdef bytes format_separator(bytes format):
 cpdef int count_entries(file, bytes format):
     
     try:
-        sep = format_separator(format)
+        sep = format_uniq_pattern(format)
         if sep is None:
             return -1
         sep = re.compile(sep)
@@ -72,7 +72,7 @@ cpdef int count_entries(file, bytes format):
                 return -1
             mmapped_file = mmap.mmap(f.fileno(), 0, access=mmap.ACCESS_READ)
             total_count += len(re.findall(sep, mmapped_file))
-            if format != b"ngsfilter" and format != b"tabular" and format != b"embl" and format != b"genbank":
+            if format != b"ngsfilter" and format != b"tabular" and format != b"embl" and format != b"genbank" and format != b"fastq":
                 total_count += 1 # adding +1 for 1st entry because separators include \n (ngsfilter and tabular already count one more because of last \n)
             
     except:

python/obitools3/version.py

@@ -1,5 +1,5 @@
 major = 3
 minor = 0
-serial= '0b31'
+serial= '0b34'
 
 version ="%d.%d.%s" % (major,minor,serial)

src/obidmscolumn.c

@@ -1725,17 +1725,33 @@ int obi_close_column(OBIDMS_column_p column)
 int obi_clone_column_indexer(OBIDMS_column_p column)
 {
 	char* new_indexer_name;
+	int i;
 
-	new_indexer_name = obi_build_indexer_name((column->header)->name, (column->header)->version);
+	i=0;
+	while (true) // find avl name not already used
+	{
+		new_indexer_name = obi_build_indexer_name((column->header)->name, ((column->header)->version)+i);
 		if (new_indexer_name == NULL)
 			return -1;
 
 		column->indexer = obi_clone_indexer(column->indexer, new_indexer_name);	// TODO Need to lock this somehow?
 		if (column->indexer == NULL)
 		{
+			if (errno == EEXIST)
+			{
+				free(new_indexer_name);
+				i++;
+			}
+			else
+			{
+				free(new_indexer_name);
 				obidebug(1, "\nError cloning a column's indexer to make it writable");
 				return -1;
 			}
+		}
+		else
+			break;
+	}
 
 	strcpy((column->header)->indexer_name, new_indexer_name);
 
@@ -2415,16 +2431,20 @@ char* obi_get_formatted_elements_names(OBIDMS_column_p column)
 }
 
 
-char* obi_column_formatted_infos(OBIDMS_column_p column)
+char* obi_column_formatted_infos(OBIDMS_column_p column, bool detailed)
 {
-	char* column_infos;
+	char* column_infos = NULL;
-	char* elt_names;
+	char* elt_names = NULL;
-
+	char* column_name = NULL;
-	column_infos = malloc(1024 * sizeof(char));
+	 // should be in view.c because alias exists in the context of view
+	column_infos = malloc(2048 * sizeof(char)); // TODO
 
 	elt_names = obi_get_formatted_elements_names(column);
 
 
+//  "column_name, data type: OBI_TYPE, element names: [formatted element names](, all comments)"
+
+
 	free(elt_names);
 	return column_infos;
 }