obi clean: fixed a memory bug, fixed the behaviour when no sample info,

and added checks warnings and error handling when sample info not
dereplicated

python/obitools3/commands/clean.pyx

@@ -36,8 +36,7 @@ def addOptions(parser):
                        dest="clean:sample-tag-name",
                        metavar="<SAMPLE TAG NAME>",
                        type=str,
-                       default="merged_sample",
+                       help="Name of the tag where merged sample count informations are kept (typically generated by obi uniq, usually MERGED_sample, default: None).")
-                       help="Name of the tag where sample counts are kept.")
     
     group.add_argument('--ratio', '-r',
                        action="store", dest="clean:ratio",
@@ -107,6 +106,9 @@ def run(config):
     command_line = " ".join(sys.argv[1:])
     comments = View.print_config(config, "clean", command_line, input_dms_name=[i_dms_name], input_view_name=[i_view_name])
 
+    if 'sample-tag-name' not in config['clean']:
+        config['clean']['sample-tag-name'] = ""
+        
     if obi_clean(i_dms.name_with_full_path, tobytes(i_view_name), tobytes(config['clean']['sample-tag-name']), tobytes(o_view_name), comments, \
               config['clean']['distance'], config['clean']['ratio'], config['clean']['heads-only'], config['clean']['thread-count']) < 0:
         raise Exception("Error running obiclean")

python/obitools3/version.py

@@ -1,5 +1,5 @@
 major = 3
 minor = 0
-serial= '0b26'
+serial= '0b26a'
 
 version ="%d.%d.%s" % (major,minor,serial)

src/obi_clean.c

@@ -246,7 +246,16 @@ int obi_clean(const char* dms_name,
 
 	// Open the sample column if there is one
 	if ((strcmp(sample_column_name, "") == 0) || (sample_column_name == NULL))
-		sample_column = NULL;
+	{
+		fprintf(stderr, "Info: No sample information provided, assuming one sample.\n");
+		sample_column = obi_view_get_column(i_view, COUNT_COLUMN);
+		if (sample_column == NULL)
+		{
+			obidebug(1, "\nError getting the COUNT column");
+			return -1;
+		}
+		sample_count = 1;
+	}
 	else
 	{
 		sample_column = obi_view_get_column(i_view, sample_column_name);
@@ -255,6 +264,13 @@ int obi_clean(const char* dms_name,
 			obidebug(1, "\nError getting the sample column");
 			return -1;
 		}
+		sample_count = (sample_column->header)->nb_elements_per_line;
+		// Check that the sample column is a merged column with all sample informations
+		if (sample_count == 1)
+		{
+			obidebug(1, "\n\nError: If a sample column is provided, it must contain 'merged' sample counts as built by obi uniq with the -m option\n");
+			return -1;
+		}
 	}
 
 	// Create the output view, or a temporary one if heads only
@@ -279,8 +295,6 @@ int obi_clean(const char* dms_name,
 		return -1;
 	}
 
-	sample_count = (sample_column->header)->nb_elements_per_line;
-
 	// Create the output columns
 	if (create_output_columns(o_view, sample_column, sample_count) < 0)
 	{
@@ -549,7 +563,7 @@ int obi_clean(const char* dms_name,
 
 	if (heads_only)
 	{
-		line_selection = malloc((o_view->infos)->line_count * sizeof(index_t));
+		line_selection = malloc((((o_view->infos)->line_count) + 1) * sizeof(index_t));
 		if (line_selection == NULL)
 		{
 			obi_set_errno(OBI_MALLOC_ERROR);

src/obi_clean.h

@@ -52,7 +52,8 @@
  *
  * @param dms A pointer on an OBIDMS.
  * @param i_view_name The name of the input view.
- * @param sample_column_name The name of the OBI_STR column in the input view where the sample information is kept.
+ * @param sample_column_name The name of the column in the input view where the sample information is kept.
+ *                           Must be merged informations as built by the obi uniq tool (checked by the function).
  *                           NULL or "" (empty string) if there is no sample information.
  * @param o_view_name The name of the output view where the results should be written (should not already exist).
  * @param o_view_comments The comments that should be associated with the output view.