version 0.1

Author: semio

ddf_utils/chef/cook.py

@@ -212,9 +212,14 @@ def dish_to_csv(dishes, outpath):
                 if t == 'datapoints':
                     df = df.set_index(by)
                     if not np.issubdtype(df[k].dtype, np.number):
-                        df[k] = df[k].astype(float)
-                    # TODO: better handle the float format.
-                    df[k] = df[k].map(lambda x: format_float_digits(x, 5))
+                        try:
+                            df[k] = df[k].astype(float)
+                            # TODO: better handle the float format.
+                            df[k] = df[k].map(lambda x: format_float_digits(x, 5))
+                        except ValueError:
+                            logging.warning("data not numeric: " + k)
+                    else:
+                        df[k] = df[k].map(lambda x: format_float_digits(x, 5))
                     df[[k]].to_csv(path)
                 else:
                     df.to_csv(path, index=False)

ddf_utils/chef/ingredient.py

@@ -51,7 +51,7 @@ def dtype(self):
         """
         returns the type of ddf data, i.e. concepts/entities/datapoints.
 
-        It will be inferred from the key porperity of ingredient.
+        It will be inferred from the key property of ingredient.
         TODO: what if key == '*'? Is it possible?
         """
         if self.key == 'concept':

ddf_utils/str.py

@@ -72,6 +72,7 @@ def format_float_digits(number, digits=5, threshold=None):
     #     d = decimal.Decimal(number)
     # except TypeError:
     #     d = _float_to_decimal(float(number))
+    #FIXME: deal with nan
 
     d = float(number)
 

setup.py

@@ -5,7 +5,7 @@
 
 setup(
     name='ddf_utils',
-    version='0.0.4-dev',
+    version='0.1.0-dev',
     description='Commonly used functions/utilities for DDF file model.',
     url='https://github.com/semio/ddf_utils',
     author='Semio Zheng',