Bump version: 0.3.2 → 0.3.4

.travis.yml

@@ -15,9 +15,15 @@ deploy:
     distributions: sdist bdist_wheel
     repo: zalando/expan
 env:
-- TOXENV=py27
+- TOXENV=py
 install:
-- pip install -U tox
+- pip install -U pip setuptools tox
+- pip install -r requirements.txt
 language: python
-python: 2.7
+python:
+#- 2.6
+- 2.7
+#- 3.3
+- 3.4
+- 3.5
 script: tox

MANIFEST.in

@@ -3,6 +3,7 @@ include CONTRIBUTING.rst
 include HISTORY.rst
 include LICENSE
 include README.rst
+include requirements.txt
 
 recursive-include tests *
 recursive-exclude * __pycache__

README.rst

@@ -135,6 +135,11 @@ with the following exceptions:
 
 - Use *tabs instead of spaces* - this allows all individuals to have visual depth of indentation they prefer, without changing the source code at all, and it is simply smaller
 
+Testing
+-------
+
+Easiest way to run tests is by running the command ``tox`` from the terminal. The default Python environments for testing with are py27 and py34, but you can specify your own by running e.g. ``tox -e py35``.
+
 Branching / Release
 -------------------
 
@@ -152,7 +157,7 @@ Versioning
 when doing the analysis!**
 
 We use semantic versioning (http://semver.org), and the current version of
-ExpAn is: v0.3.2.
+ExpAn is: v0.3.4.
 
 The version is maintained in ``setup.cfg``, and propagated from there to various files
 by the ``bumpversion`` program. The most important propagation destination is
@@ -176,7 +181,7 @@ repository.
 
     >>> import core.binning
     >>> core.version()
-    'v0.3.2'
+    'v0.3.4'
     >>> core.version('{major}.{minor}..{commits}')
     '0.0..176'
     >>> core.version('{commit}')
@@ -263,21 +268,3 @@ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
 CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
 TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
 SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
-
-History
-=======
-
-0.2.5
------
-
-* Inclusion of cli in install
-* many other minor changes since open-sourcing...
-
-0.2.0 (2016-05-03)
-------------------
-
-* First opensource release to GitHub
-
-0.1.0 (2016-04-29)
-------------------

expan/__init__.py

@@ -1,6 +1,7 @@
-from core import *
-from core.version import __version__
-from data import *
-from cli import *
+from __future__ import absolute_import
+from expan.core import *
+from expan.core.version import __version__
+from expan.data import *
+from expan.cli import *
 
 __all__ = ["core", "data", "cli"]

expan/cli/cli.py

@@ -45,7 +45,8 @@ def run_analysis(features_file, kpis_file, metadata_file):
 	return (exp.delta(), exp.sga())
 
 
-def run_expan((features_file, kpis_file, metadata_file, output_file)):
+def run_expan(xxx_todo_changeme):
+	(features_file, kpis_file, metadata_file, output_file) = xxx_todo_changeme
 	(delta_result, sga_result) = run_analysis(features_file, kpis_file, metadata_file)
 	print_results(delta_result, sga_result, output_file)
 
@@ -58,12 +59,12 @@ def print_results(delta, sga, output_file):
 		output.write(delta_s)
 		output.write(sga_s)
 	else:
-		print delta_s
-		print sga_s
+		print(delta_s)
+		print(sga_s)
 
 
 def check_input_data(args):
-	print args
+	print(args)
 	if not args.kpis:
 		raise UsageError('Kpis file shall be provided (-k cli parameter)')
 	if not args.metadata:

expan/core/__init__.py

@@ -1,5 +1,7 @@
+from __future__ import absolute_import
+
 __all__ = ["binning", "debugging", "experiment", "experimentdata", "results", "statistics", "util", "version"]
 
-from version import __version__, version
+from expan.core.version import __version__, version
 
-print 'ExpAn core init: {}'.format(version())
+print(('ExpAn core init: {}'.format(version())))

expan/core/binning.py

@@ -13,7 +13,7 @@
 
 def dbg(lvl, msg):
 	if lvl <= dbg_lvl:
-		print
+		print()
 		'D{:d}|{}'.format(lvl, msg)
 
 
@@ -157,9 +157,9 @@ def _labels(self, format_str='{standard}'):
 
 		if '{iter.' in format_str:
 			if '{iter.uppercase}' in format_str:
-				it = iter(string.uppercase)
+				it = iter(string.ascii_uppercase)
 			elif '{iter.lowercase}' in format_str:
-				it = iter(string.lowercase)
+				it = iter(string.ascii_lowercase)
 			elif '{iter.integer}' in format_str:
 				import itertools
 				it = itertools.count()
@@ -184,10 +184,10 @@ def _labels(self, format_str='{standard}'):
 				format_args['set_notation'] = '{unseen}'
 
 			if it is not None:
-				print
+				print()
 				'ii: ' + str(ii)
 				if not is_catchall:
-					format_args['iterator'] = it.next()
+					format_args['iterator'] = next(it)
 				else:
 					format_args['iterator'] = '?'
 
@@ -445,9 +445,9 @@ def _labels(self, format_str='{standard}'):
 
 		if '{iter.' in format_str:
 			if '{iter.uppercase}' in format_str:
-				it = iter(string.uppercase)
+				it = iter(string.ascii_uppercase)
 			elif '{iter.lowercase}' in format_str:
-				it = iter(string.lowercase)
+				it = iter(string.ascii_lowercase)
 			elif '{iter.integer}' in format_str:
 				import itertools
 				it = itertools.count()
@@ -485,7 +485,7 @@ def _labels(self, format_str='{standard}'):
 				'up_bracket': ']' if uc else ')',
 			}
 			if (not is_catchall) and (it is not None):
-				format_args['iterator'] = it.next()
+				format_args['iterator'] = next(it)
 
 			lbl = (format_str if not is_catchall else catchall_format_str).format(**format_args)
 

expan/core/experiment.py

@@ -2,19 +2,20 @@
 # the proper interface is through the Experiment instance functions, surely?
 
 # import numpy as np
-import statistics as statx
+
+import expan.core.statistics as statx
 import warnings
 
-import binning as binmodule  # name conflict with binning...
+import expan.core.binning as binmodule  # name conflict with binning...
 import numpy as np
 import pandas as pd
-from experimentdata import ExperimentData
-from results import Results, delta_to_dataframe_all_variants, feature_check_to_dataframe
+from expan.core.experimentdata import ExperimentData
+from expan.core.results import Results, delta_to_dataframe_all_variants, feature_check_to_dataframe
 
 # raise the same warning multiple times
 warnings.simplefilter('always', UserWarning)
 
-from debugging import Dbg
+from expan.core.debugging import Dbg
 
 def _binned_deltas(df, variants, n_bins=4, binning=None, cumulative=False,
 				   assume_normal=True, percentiles=[2.5, 97.5],
@@ -258,7 +259,7 @@ def time_dependent_deltas(df, variants, time_step=1, cumulative=False,
 
 	# create binning manually, ASSUMING uniform sampling
 	tpoints = np.unique(df.iloc[:,1])
-	binning = binmodule.NumericalBinning(uppers=tpoints, lowers=tpoints, 
+	binning = binmodule.NumericalBinning(uppers=tpoints, lowers=tpoints,
 		up_closed=[True]*len(tpoints), lo_closed=[True]*len(tpoints))
 
 	# Push computation to _binned_deltas() function

expan/core/results.py

@@ -1,15 +1,16 @@
+from __future__ import absolute_import
 import datetime
 from copy import deepcopy
 
 import numpy as np
 import pandas as pd
-import version
-import statistics as statx
+from expan.core.version import __version__
+import expan.core.statistics as statx
 from scipy.stats import norm
 
 # from tests.tests_core.test_data import generate_random_data
 
-from debugging import Dbg
+from expan.core.debugging import Dbg
 from pdb import set_trace
 
 class Results(object):
@@ -42,7 +43,7 @@ def __init__(self, df, metadata={}, dbg=None):
 	    """
 		self.df = df
 		self.metadata = metadata
-		self.metadata['version'] = version.__version__
+		self.metadata['version'] = __version__
 		self.metadata['errors'] = {}
 		self.metadata['warnings'] = {}
 
@@ -275,7 +276,7 @@ def to_hdf(self, fpath):
 		hfile = h5py.File(fpath)
 		md = hfile.require_group('metadata')
 		datetime_conversions = set(md.attrs.get('_datetime_attributes', set()))
-		for k, v in self.metadata.iteritems():
+		for k, v in list(self.metadata.items()):
 			if k == '_datetime_attributes':
 				continue
 			if v is None:
@@ -350,7 +351,7 @@ def from_hdf(fpath, dbg=None):
 	md = hfile['metadata']
 	datetime_conversions = set(md.attrs.get('_datetime_attributes', set()))
 	metadata = {}
-	for k, v in md.attrs.iteritems():
+	for k, v in list(md.attrs.items()):
 		if k == '_datetime_attributes':
 			continue
 		dbg(3, 'from_hdf: retrieving metadata {}'.format(k))
@@ -371,8 +372,8 @@ def delta_to_dataframe(metric, variant, mu, pctiles, samplesize_variant, samples
 		'metric': metric,
 		'variant': variant,
 		'statistic': 'pctile',
-		'pctile': pctiles.keys(),
-		'value': pctiles.values(),
+		'pctile': list(pctiles.keys()),
+		'value': list(pctiles.values()),
 		'subgroup_metric': subgroup_metric,
 		'subgroup': subgroup
 	})
@@ -403,8 +404,8 @@ def delta_to_dataframe_all_variants(metric, mu, pctiles, samplesize_variant,
 	df = pd.DataFrame({
 		'metric': metric,
 		'statistic': 'uplift_pctile',
-		'pctile': pctiles.keys(),
-		'value': pctiles.values(),
+		'pctile': list(pctiles.keys()),
+		'value': list(pctiles.values()),
 		'subgroup_metric': subgroup_metric,
 		'subgroup': subgroup
 	})
@@ -435,8 +436,8 @@ def feature_check_to_dataframe(metric,
 	if pval is None:
 		df = pd.DataFrame({'metric': metric,
 						   'statistic': 'pre_treatment_diff_pctile',
-						   'pctile': pctiles.keys(),
-						   'value': pctiles.values(),
+						   'pctile': list(pctiles.keys()),
+						   'value': list(pctiles.values()),
 						   'subgroup_metric': '-',
 						   'subgroup': None})
 		df = df.append(pd.DataFrame({

expan/core/statistics.py

@@ -70,7 +70,7 @@ def delta(x, y, assume_normal=True, percentiles=[2.5, 97.5],
 		# Set mean to nan
 		mu = np.nan
 		# Create nan dictionary
-		c_i = dict(zip(percentiles, np.empty(len(percentiles)) * np.nan))
+		c_i = dict(list(zip(percentiles, np.empty(len(percentiles)) * np.nan)))
 	else:
 		# Computing the mean
 		mu = _delta_mean(_x, _y)
@@ -239,7 +239,7 @@ def bootstrap(x, y, func=_delta_mean, nruns=10000, percentiles=[2.5, 97.5],
 	# Checking if enough observations are left after dropping NaNs
 	if min(ss_x, ss_y) < min_observations:
 		# Create nan percentile dictionary
-		c_val = dict(zip(percentiles, np.empty(len(percentiles)) * np.nan))
+		c_val = dict(list(zip(percentiles, np.empty(len(percentiles)) * np.nan)))
 		return (c_val, None)
 	else:
 		# Initializing bootstraps array and random sampling for each run
@@ -254,7 +254,7 @@ def bootstrap(x, y, func=_delta_mean, nruns=10000, percentiles=[2.5, 97.5],
 		if relative:
 			bootstraps -= np.nanmean(bootstraps)
 		# Confidence values per given percentile as dictionary
-		c_val = dict(zip(percentiles, np.percentile(bootstraps, q=percentiles)))
+		c_val = dict(list(zip(percentiles, np.percentile(bootstraps, q=percentiles))))
 		return (c_val, None) if not return_bootstraps else (c_val, bootstraps)
 
 

expan/core/version.py

@@ -1,5 +1,5 @@
 #
-__version__ = "0.3.2"
+__version__ = "0.3.4"
 
 
 def version_numbers():
@@ -12,7 +12,7 @@ def git_commit_count():
 
   http://programmers.stackexchange.com/a/151558
   """
-	print
+	print()
 	'dd'
 	commit_count = None
 
@@ -30,7 +30,7 @@ def git_latest_commit():
 
   http://programmers.stackexchange.com/a/151558
   """
-	print
+	print()
 	'ee'
 	latest_commit = None
 	import subprocess
@@ -67,7 +67,7 @@ def version(format_str='{short}'):
 
 
 if __name__ == '__main__':
-	print
+	print()
 	version('Short Version String: "{short}"')
-	print
+	print()
 	version('Full Version String: "{long}"')

expan/data/csv_fetcher.py

@@ -30,21 +30,21 @@ def get_data(folder_path):
 				try:
 					metrics = pd.read_csv(folder_path + '/' + f)
 				except Exception as e:
-					print
+					print()
 					e
 
 			elif 'metadata' in f:
 				try:
 					with open(folder_path + '/' + f, 'r') as input_json:
 						metadata = json.load(input_json)
 				except ValueError as e:
-					print
+					print()
 					e
 					raise
 
 		return ExperimentData(metrics=metrics, metadata=metadata)
 
 	except AssertionError as e:
-		print
+		print()
 		e
 		raise

requirements.txt

@@ -1,3 +1,4 @@
+pip >= 8.1.0
 pandas >= 0.17.1
 scipy >= 0.17.0
 numpy >= 1.10.4

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.3.2
+current_version = 0.3.4
 commit = True
 tag = True