Changeset 1480:b343ce6be869 in orange-bioinformatics

Timestamp:

09/29/11 19:46:44 (20 months ago)

Author:

markotoplak

Branch:

default

Convert:

0f6898ace32a64c80554a9ca782e7c03201e12c5

Message:

obiDicty: when joining replicates sort replicates by their replicate id

Files:

• obiDicty.py (modified) (2 diffs)
• widgets/OWPIPA.py (modified) (1 diff)

obiDicty.py

@@ -1350 +1350 @@
     return od
 
-def join_replicates(data, ignorenames=["id", "replicate", "name", "map_stop1"], namefn=None, avg=median):
+def join_replicates(data, ignorenames=["replicate", "id", "name", "map_stop1"], namefn=None, avg=median):
     """ Join replicates by median. 
-    Default parameters work for PIPA data."""
+    Default parameters work for PIPA data.
+    Sort elements in the same order as ignorenames!
+    """
+
     d = defaultdict(list)
 
@@ -1396 +1399 @@
             return None
 
+    def data_type(vals): #data type converter
+        try:
+            _ = [ int(a) for a in vals ]
+            return int
+        except:
+            try:
+                _ = [ float(a) for a in vals ]
+                return float
+            except:
+                return lambda x: x
+
+    all_values = defaultdict(set)
+    for a in [ at.attributes for at in data.domain.attributes ]:
+        for k,v in a.iteritems():
+            all_values[k].add(v)
+
+    types = {}
+    for k,vals in all_values.iteritems():
+        types[k] = data_type(vals)
+  
     for group, elements in d.items():
         a = orange.FloatVariable()
+        #here sort elements
+    
+        def sk(x):
+            return ( types[n](x[n]) for n in ignorenames if n in all_values )
+
+        elements = sorted(elements, key=lambda x: sk(data.domain.attributes[x].attributes))
+
         a.attributes.update(join_ats([data.domain.attributes[i].attributes for i in elements]))
         a.name = namefn(a.attributes)

widgets/OWPIPA.py

@@ -680 +680 @@
 
             if self.joinreplicates:
-                table = obiDicty.join_replicates(table, ignorenames=["id", "replicate", "name", "map_stop1"], namefn=None, avg=obiDicty.median)
+                table = obiDicty.join_replicates(table, ignorenames=["replicate", "id", "name", "map_stop1"], namefn=None, avg=obiDicty.median)
 
             end = int(time.time()-start)