Implemented PhenotypeDataset with tests.

workflows/affy_workflow.py

@@ -1,6 +1,6 @@
 import gtk
 from system import dataset, logger, plots, workflow
-from scipy import randn
+from scipy import randn, array, transpose, zeros
 import cPickle
 
 
@@ -185,3 +185,53 @@ class PCAFunction(workflow.Function):
         
         return [T, P, loading_plot, score_plot]
 
+
+class PhenotypeDataset(dataset.Dataset):
+    def __init__(self, string):
+        self._table = rows = [line.split("\t") for line in string.splitlines()]
+        columns = zip(*rows[1:])
+        cel_names = columns[0]
+        col_names = rows[0][1:]
+        phenotypes = []
+        categories = {}
+
+        for col_name, column in zip(col_names, columns[1:]):
+            try:
+                categories[col_name] = map(int, column)
+                phenotypes.append(col_name)
+            except ValueError:
+                # category-data
+                keys = []
+                entries = {}
+                for i, entry in enumerate(column):
+                    if entry not in entries:
+                        keys.append(entry)
+                        entries[entry] = []
+
+                    entries[entry].append(i)
+                    
+
+                for key in keys:
+                    z = zeros(len(column))
+                    for i in entries[key]:
+                        z[i] = 1
+                    key = "%s-%s" % (col_name, key)
+                    phenotypes.append(key)
+                    categories[key] = z
+
+        matrix_data = []
+        for key in phenotypes:
+            matrix_data.append(categories[key])
+
+        if matrix_data:
+            a = transpose(array(matrix_data))
+        else:
+            a = None
+
+        dataset.Dataset.__init__(self, a, identifiers=[('CEL', cel_names),
+                                                       ('phenotypes', phenotypes)],
+                                 shape=(len(cel_names),len(phenotypes)))
+            
+
+    def get_phenotype_table(self):
+        return self._table