Switch to version 3.0.0b27

values in an int64 would mess trees up resulting in failed data
dereplication

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 sample counts are kept.")
+                       help="Name of the tag where merged sample count informations are kept (typically generated by obi uniq, usually MERGED_sample, default: None).")
     
     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/commands/count.pyx

@@ -22,7 +22,7 @@ def addOptions(parser):
     group.add_argument('-s','--sequence',
                         action="store_true", dest="count:sequence",
                         default=False,
-                        help="Prints only the number of sequence records.")
+                        help="Prints only the number of sequence records (much faster, default: False).")
  
     group.add_argument('-a','--all',
                         action="store_true", dest="count:all",

python/obitools3/commands/grep.pyx

@@ -161,8 +161,7 @@ def obi_eval(compiled_expr, loc_env, line):
     return obi_eval_result
     
 
-def Filter_generator(options, tax_filter):
-    #taxfilter = taxonomyFilterGenerator(options)
+def Filter_generator(options, tax_filter, i_view):
 
     # Initialize conditions
     predicates = None
@@ -171,6 +170,9 @@ def Filter_generator(options, tax_filter):
     attributes = None
     if "attributes" in options and len(options["attributes"]) > 0:
         attributes = options["attributes"]
+        for attribute in attributes:
+            if attribute not in i_view:
+                return None
     lmax = None
     if "lmax" in options:
         lmax = options["lmax"]
@@ -196,6 +198,8 @@ def Filter_generator(options, tax_filter):
     if "attribute_patterns" in options and len(options["attribute_patterns"]) > 0:
         for p in options["attribute_patterns"]:
             attribute, pattern = p.split(":", 1)
+            if attribute not in i_view:
+                return None
             attribute_patterns[tobytes(attribute)] = re.compile(tobytes(pattern))
     
     def filter(line, loc_env):
@@ -324,21 +328,29 @@ def run(config):
  
     # Apply filter
     tax_filter = Taxonomy_filter_generator(taxo, config["grep"])
-    filter = Filter_generator(config["grep"], tax_filter)
+    filter = Filter_generator(config["grep"], tax_filter, i_view)        
     selection = Line_selection(i_view)
-    for i in range(len(i_view)):
-        PyErr_CheckSignals()
-        pb(i)
-        line = i_view[i]
-                 
-        loc_env = {"sequence": line, "line": line, "taxonomy": taxo, "obi_eval_result": False}
-        
-        good = filter(line, loc_env)
- 
-        if good :
-            selection.append(i)
-
-    pb(i, force=True)
+    
+    if filter is None and config["grep"]["invert_selection"]: # all sequences are selected: filter is None if no line will be selected because some columns don't exist
+        for i in range(len(i_view)):
+            PyErr_CheckSignals()
+            pb(i)
+            selection.append(i)        
+    
+    elif filter is not None :   # filter is None if no line will be selected because some columns don't exist
+        for i in range(len(i_view)):
+            PyErr_CheckSignals()
+            pb(i)
+            line = i_view[i]
+                     
+            loc_env = {"sequence": line, "line": line, "taxonomy": taxo, "obi_eval_result": False}
+            
+            good = filter(line, loc_env)
+     
+            if good :
+                selection.append(i)
+    
+    pb(len(i_view), force=True)
     print("", file=sys.stderr)
 
     # Create output view with the line selection

python/obitools3/commands/ls.pyx

@@ -34,9 +34,10 @@ def run(config):
     if input[2] == DMS and not config['ls']['longformat']:
         dms = input[0]
         l = []
-        for view in input[0]:
-            l.append(tostr(view) + "\t(Date created: " + str(bytes2str_object(dms[view].comments["Date created"]))+")")
-            dms[view].close()
+        for viewname in input[0]:
+            view = dms[viewname]
+            l.append(tostr(viewname) + "\t(Date created: " + str(bytes2str_object(view.comments["Date created"]))+")")
+            view.close()
         l.sort()
         for v in l:
             print(v)

python/obitools3/commands/ngsfilter.pyx

@@ -479,6 +479,8 @@ cdef tuple annotate(sequences, infos, no_tags, verbose=False):
     if not directmatch[0].forward:
         sequences[0] = sequences[0].reverse_complement
         sequences[0][b'reversed'] = True   # used by the alignpairedend tool (in kmer_similarity.c)
+    else:
+        sequences[0][b'reversed'] = False   # used by the alignpairedend tool (in kmer_similarity.c)
 
     sample=None
     if not no_tags:
@@ -506,7 +508,7 @@ cdef tuple annotate(sequences, infos, no_tags, verbose=False):
                     sample=None
         
         if sample is None:
-            sequences[0][b'error']=b"No tags found"
+            sequences[0][b'error']=b"No sample with that tag combination"
             return False, sequences[0]
     
         sequences[0].update(sample)

python/obitools3/commands/uniq.pyx

@@ -307,6 +307,7 @@ cdef uniq_sequences(View_NUC_SEQS view, View_NUC_SEQS o_view, ProgressBar pb, di
         for x in categories :
             catl.append(i_seq[x])    
           
+        #unique_id = tuple(catl) + (i_seq_col[i],)
         unique_id = tuple(catl) + (i_seq_col.get_line_idx(i),)
         #unique_id = tuple(i_seq[x] for x in categories) + (seq_col.get_line_idx(i),)  # The line that cython can't read properly
          
@@ -453,6 +454,7 @@ cdef uniq_sequences(View_NUC_SEQS view, View_NUC_SEQS o_view, ProgressBar pb, di
             merged_dict[mkey] = {}
 
         for i_idx in merged_sequences:
+            PyErr_CheckSignals()
             pb(total_treated)
  
             i_id = i_id_col[i_idx]
@@ -591,10 +593,11 @@ def run(config):
     # Initialize the progress bar
     pb = ProgressBar(len(entries), config, seconde=5)
     
-    try:
-        uniq_sequences(entries, o_view, pb, config, mergedKeys_list=config['uniq']['merge'], taxonomy=taxo, mergeIds=config['uniq']['mergeids'], categories=config['uniq']['categories'], max_elts=config['obi']['maxelts'])       
-    except Exception, e:
-        raise RollbackException("obi uniq error, rollbacking view: "+str(e), o_view)
+    if len(entries) > 0:
+        try:
+            uniq_sequences(entries, o_view, pb, config, mergedKeys_list=config['uniq']['merge'], taxonomy=taxo, mergeIds=config['uniq']['mergeids'], categories=config['uniq']['categories'], max_elts=config['obi']['maxelts'])       
+        except Exception, e:
+            raise RollbackException("obi uniq error, rollbacking view: "+str(e), o_view)
     
     print("", file=sys.stderr)
 

python/obitools3/dms/column/column.pxd

@@ -22,6 +22,7 @@ cdef class Column(OBIWrapper) :
     
     cdef inline OBIDMS_column_p pointer(self)
     cdef read_elements_names(self)
+    cpdef list keys(self)
     
     @staticmethod
     cdef type get_column_class(obitype_t obitype, bint multi_elts, bint tuples)

python/obitools3/dms/column/column.pyx

@@ -323,7 +323,10 @@ cdef class Column(OBIWrapper) :
         free(elts_names_b)
         return elts_names_list
     
-    
+    cpdef list keys(self):
+        return self._elements_names
+
+
     # Column alias property getter and setter
     @property
     def name(self):
@@ -340,7 +343,7 @@ cdef class Column(OBIWrapper) :
     @property
     def elements_names(self):
         return self._elements_names
- 
+       
     # nb_elements_per_line property getter
     @property
     def nb_elements_per_line(self):

python/obitools3/dms/dms.pyx

@@ -227,7 +227,9 @@ cdef class DMS(OBIWrapper):
         cdef str s 
         s=""
         for view_name in self.keys():
-            s = s + repr(self.get_view(view_name)) + "\n"
+            view = self.get_view(view_name)
+            s = s + repr(view) + "\n"
+            view.close()
         return s
     
 

python/obitools3/format/tab.pyx

@@ -3,7 +3,7 @@
 cimport cython
 from obitools3.dms.view.view cimport Line
 from obitools3.utils cimport bytes2str_object, str2bytes, tobytes
-from obitools3.dms.column.column cimport Column_line
+from obitools3.dms.column.column cimport Column_line, Column_multi_elts
 
 
 cdef class TabFormat:
@@ -25,19 +25,29 @@ cdef class TabFormat:
         for k in self.tags:
             
             if self.header and self.first_line:
-                value = tobytes(k)
+                if isinstance(data.view[k], Column_multi_elts):
+                    for k2 in data.view[k].keys():
+                        line.append(tobytes(k)+b':'+tobytes(k2))
+                else:
+                    line.append(tobytes(k))
             else:
                 value = data[k]
-                if value is not None:
-                    if type(value) == Column_line:
-                        value = value.bytes()
+                if isinstance(data.view[k], Column_multi_elts):
+                    if value is None:  # all keys at None
+                        for k2 in data.view[k].keys(): # TODO could be much more efficient
+                            line.append(self.NAString)
                     else:
-                        value = str2bytes(str(bytes2str_object(value))) # genius programming
-                if value is None:
-                    value = self.NAString
-            
-            line.append(value)
-      	
+                        for k2 in data.view[k].keys(): # TODO could be much more efficient
+                            if value[k2] is not None:
+                                line.append(str2bytes(str(bytes2str_object(value[k2]))))  # genius programming
+                            else:
+                                line.append(self.NAString)
+                else:
+                    if value is not None:
+                        line.append(str2bytes(str(bytes2str_object(value))))
+                    else:
+                        line.append(self.NAString)
+                  	
         if self.first_line:
             self.first_line = False
       		

python/obitools3/utils.pyx

@@ -166,7 +166,9 @@ cdef object bytes2str_object(object value):  # Only works if complex types are d
                     value[k] = bytes2str(v)
             if type(k) == bytes:
                 value[bytes2str(k)] = value.pop(k)
-    elif isinstance(value, list):
+    elif isinstance(value, list) or isinstance(value, tuple):
+        if isinstance(value, tuple):
+            value = list(value)
         for i in range(len(value)):
             if isinstance(value[i], list) or isinstance(value[i], dict):
                 value[i] = bytes2str_object(value[i])

python/obitools3/version.py

@@ -1,5 +1,5 @@
 major = 3
 minor = 0
-serial= '0b24'
+serial= '0b27'
 
 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.

src/obiavl.c

@@ -2259,7 +2259,13 @@ index_t obi_avl_add(OBIDMS_avl_p avl, Obi_blob_p value)
 		parent = next;
 
 		// Compare the crc of the value with the crc of the current node
-		comp = (current_node->crc64) - crc;
+		//comp = (current_node->crc64) - crc;
+		if ((current_node->crc64) == crc)
+			comp = 0;
+		else if ((current_node->crc64) > crc)
+			comp = 1;
+		else
+			comp = -1;
 
 		if (comp == 0)
 		{ // check if really same value
@@ -2354,7 +2360,13 @@ index_t obi_avl_find(OBIDMS_avl_p avl, Obi_blob_p value)
 		current_node = (avl->tree)+next;
 
 		// Compare the crc of the value with the crc of the current node
-		comp = (current_node->crc64) - crc;
+		//comp = (current_node->crc64) - crc;
+		if ((current_node->crc64) == crc)
+			comp = 0;
+		else if ((current_node->crc64) > crc)
+			comp = 1;
+		else
+			comp = -1;
 
 		if (comp == 0)
 		{ // Check if really same value

src/obidmscolumn.c

@@ -1350,6 +1350,8 @@ OBIDMS_column_p obi_create_column(OBIDMS_p     dms,
 		}
 		strncpy(header->indexer_name, final_indexer_name, INDEXER_MAX_NAME);
 	}
+	else
+		new_column->indexer = NULL;
 
 	// Fill the data with NA values
 	obi_ini_to_NA_values(new_column, 0, nb_lines);
@@ -1558,6 +1560,8 @@ OBIDMS_column_p obi_open_column(OBIDMS_p     dms,
 			return NULL;
 		}
 	}
+	else
+		column->indexer = NULL;
 
 	if (close(column_file_descriptor) < 0)
 	{
@@ -1694,7 +1698,7 @@ int obi_close_column(OBIDMS_column_p column)
 			ret_val = -1;
 
 		// If it's a tuple column or the data type is OBI_STR, OBI_SEQ or OBI_QUAL, the associated indexer is closed
-		if (((column->header)->tuples) || (((column->header)->returned_data_type == OBI_STR) || ((column->header)->returned_data_type == OBI_SEQ) || ((column->header)->returned_data_type == OBI_QUAL)))
+		if ((column->indexer) != NULL)
 			if (obi_close_indexer(column->indexer) < 0)
 				ret_val = -1;