Merge pull request #188 from bnaul/featureset_df_save

Update calls to save/load featureset to newer cesium API

Author: bnaul

cesium_app/handlers/prediction.py

@@ -115,15 +115,15 @@ def post(self):
                                      model_or_gridcv)
         preds = executor.submit(lambda fset, model: model.predict(fset),
                                 imputed_fset, model_data)
-        pred_probs = executor.submit(lambda fset, model: model.predict_proba(fset)
+        pred_probs = executor.submit(lambda fset, model:
+                                     pd.DataFrame(model.predict_proba(fset),
+                                                  index=fset.index,
+                                                  columns=model.classes_)
                                      if hasattr(model, 'predict_proba') else [],
                                      imputed_fset, model_data)
-        all_classes = executor.submit(lambda model: model.classes_
-                                      if hasattr(model, 'classes_') else [],
-                                      model_data)
         future = executor.submit(featurize.save_featureset, imputed_fset,
                                  pred_path, labels=all_labels, preds=preds,
-                                 pred_probs=pred_probs, all_classes=all_classes)
+                                 pred_probs=pred_probs)
 
         prediction.task_id = future.key
         prediction.save()
@@ -182,9 +182,12 @@ def post(self):
                                                features_to_use=features_to_use,
                                                meta_features=meta_feats)
         fset = featurize.impute_featureset(fset, **impute_kwargs)
-        data = {'preds': model_data.predict(fset),
-                'all_classes': model_data.classes_}
+        data = {'preds': model_data.predict(fset)}
         if hasattr(model_data, 'predict_proba'):
-            data['pred_probs'] = model_data.predict_proba(fset)
+            data['pred_probs'] = pd.DataFrame(model_data.predict_proba(fset),
+                                              index=fset.index,
+                                              columns=model_data.classes_)
+        else:
+            data['pred_probs'] = []
         pred_info = Prediction.format_pred_data(fset, data)
         return self.success(pred_info)

cesium_app/models.py

@@ -3,7 +3,7 @@
 import os
 import sys
 import time
-import numpy as np
+import pandas as pd
 
 import peewee as pw
 from playhouse.postgres_ext import ArrayField, BinaryJSONField
@@ -73,6 +73,7 @@ class File(BaseModel):
     name = pw.CharField(null=True)
     created = pw.DateTimeField(default=datetime.datetime.now)
 
+
 @signals.post_delete(sender=File)
 def remove_file_after_delete(sender, instance):
     try:
@@ -135,6 +136,7 @@ class Meta:
             (('dataset', 'file'), True),
         )
 
+
 @signals.pre_delete(sender=Dataset)
 def remove_related_files(sender, instance):
     for f in instance.files:
@@ -148,7 +150,7 @@ class Featureset(BaseModel):
     name = pw.CharField()
     created = pw.DateTimeField(default=datetime.datetime.now)
     features_list = ArrayField(pw.CharField)
-    custom_features_script = pw.CharField(null=True) # move to fset file?
+    custom_features_script = pw.CharField(null=True)  # move to fset file?
     file = pw.ForeignKeyField(File, on_delete='CASCADE')
     task_id = pw.CharField(null=True)
     finished = pw.DateTimeField(null=True)
@@ -194,16 +196,15 @@ def is_owned_by(self, username):
     def format_pred_data(fset, data):
         fset.columns = fset.columns.droplevel('channel')
         fset.index = fset.index.astype(str)  # can't use ints as JSON keys
-        result = {}
-        for i, name in enumerate(fset.index):
-            result[name] = {'features': fset.loc[name].to_dict()}
-            if 'labels' in data:
-                result[name]['label'] = data['labels'][i]
-            if len(data['pred_probs']) > 0:
-                result[name]['prediction'] = dict(zip(data['all_classes'],
-                                                      data['pred_probs'][i]))
-            else:
-                result[name]['prediction'] = data['preds'][i]
+        labels = pd.Series(data.get('labels'), index=fset.index)
+        if len(data.get('pred_probs', [])) > 0:
+            preds = pd.DataFrame(data.get('pred_probs', []),
+                                 index=fset.index).to_dict(orient='index')
+        else:
+            preds = pd.Series(data['preds'], index=fset.index).to_dict()
+        result = {name: {'features': feats, 'label': labels.loc[name],
+                         'prediction': preds[name]}
+                  for name, feats in fset.to_dict(orient='index').items()}
         return result
 
     def display_info(self):
@@ -238,6 +239,7 @@ def create_tables(retry=5):
                 print('Could not connect to database...sleeping 5')
                 time.sleep(5)
 
+
 def drop_tables():
     db.drop_tables(models, safe=True, cascade=True)
 

cesium_app/tests/fixtures.py

@@ -14,6 +14,7 @@
 import peewee
 import datetime
 import joblib
+import pandas as pd
 
 
 @contextmanager
@@ -160,14 +161,14 @@ def create_test_prediction(dataset, model):
     if hasattr(model_data, 'best_estimator_'):
         model_data = model_data.best_estimator_
     preds = model_data.predict(fset)
-    pred_probs = (model_data.predict_proba(fset)
+    pred_probs = (pd.DataFrame(model_data.predict_proba(fset),
+                               index=fset.index, columns=model_data.classes_)
                   if hasattr(model_data, 'predict_proba') else [])
     all_classes = model_data.classes_ if hasattr(model_data, 'classes_') else []
     pred_path = pjoin(cfg['paths']['predictions_folder'],
                       '{}.npz'.format(str(uuid.uuid4())))
     featurize.save_featureset(fset, pred_path, labels=data['labels'],
-                              preds=preds, pred_probs=pred_probs,
-                              all_classes=all_classes)
+                              preds=preds, pred_probs=pred_probs)
     f, created = m.File.get_or_create(uri=pred_path)
     pred = m.Prediction.create(file=f, dataset=dataset, project=dataset.project,
                                model=model, finished=datetime.datetime.now())