small fixes to sd (#231)

* small fixes

* more small fixes

Author: brandondutra

solutionbox/structured_data/datalab_solutions/structured_data/_package.py

@@ -50,9 +50,6 @@
 from . import predict
 
 
-_TF_GS_URL = 'gs://cloud-datalab/deploy/tf/tensorflow-0.12.0rc1-cp27-none-linux_x86_64.whl'
-
-
 def _default_project():
   import datalab.context
   context = datalab.context.Context.default()
@@ -171,7 +168,7 @@ def cloud_preprocess(output_dir, dataset, project_id=None):
 
     print('Starting cloud preprocessing.')
     print('Track BigQuery status at')
-    print('https://bigquery.cloud.google.com/queries/%s' % datalab_project_id())
+    print('https://bigquery.cloud.google.com/queries/%s' % _default_project())
     preprocess.cloud_preprocess.main(args)
     print('Cloud preprocessing done.')
   finally:

solutionbox/structured_data/datalab_solutions/structured_data/predict/predict.py

@@ -363,7 +363,7 @@ def make_prediction_pipeline(pipeline, args):
   predicted_values, errors = (
       pipeline
       | 'Read CSV Files'
-      >> beam.io.ReadFromText(args.predict_data,
+      >> beam.io.ReadFromText(str(args.predict_data),  # DF bug: DF does not work with unicode strings
                               strip_trailing_newlines=True)
       | 'Batch Input'
       >> beam.ParDo(EmitAsBatchDoFn(args.batch_size))

solutionbox/structured_data/datalab_solutions/structured_data/trainer/util.py

@@ -65,7 +65,7 @@ def _copy_all(src_files, dest_dir):
   for src_file in src_files:
     file_name = os.path.basename(src_file)
     new_file_location = os.path.join(dest_dir, file_name)
-    file_io.copy(src_file, new_file_location)
+    file_io.copy(src_file, new_file_location, overwrite=True)
 
 
 def _recursive_copy(src_dir, dest_dir):
@@ -83,7 +83,7 @@ def _recursive_copy(src_dir, dest_dir):
     if file_io.is_directory(old_path):
       _recursive_copy(old_path, new_path)
     else:
-      file_io.copy(old_path, new_path)
+      file_io.copy(old_path, new_path, overwrite=True)
 
 def serving_from_csv_input(train_config, args, keep_target):
   """Read the input features from a placeholder csv string tensor."""