Allow user to pass in desired feature return types by RogerTangos · Pull Request #372 · alteryx/featuretools

RogerTangos · 2019-01-15T17:31:59Z

dfs() can be passed a list of allowed_variable_types, None or 'all'. This filters returned features based on their output types.
None defaults to [Numeric, Discrete, Boolean]
Docstrings updated accordingly
Completes work begun (but unexposed in the api) in ff70cbe, but
Tests changes

…d_features

…_synthesis.build_features

codecov · 2019-01-15T18:00:06Z

Codecov Report

Merging #372 into master will increase coverage by 0.09%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #372      +/-   ##
==========================================
+ Coverage   95.99%   96.09%   +0.09%     
==========================================
  Files          92       92              
  Lines        7996     8021      +25     
==========================================
+ Hits         7676     7708      +32     
+ Misses        320      313       -7

Impacted Files	Coverage Δ
featuretools/synthesis/deep_feature_synthesis.py	`96.57% <100%> (+2%)`	⬆️
...ols/tests/dfs_tests/test_deep_feature_synthesis.py	`98.49% <100%> (+0.12%)`	⬆️
featuretools/synthesis/dfs.py	`100% <100%> (ø)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update de2ff0e...c10c4b3. Read the comment docs.

RogerTangos · 2019-01-29T17:15:20Z

There were some initial strange linting errors here, which are now fixed. (strange, because the lines in question were from the master branch.)

Now this seems to be failing on pandas 0.24.0. Once #383 fixes that, I'll merge master back in.

RogerTangos · 2019-01-30T11:03:24Z

I fixed a bunch of indentation errors that were being picked up by the linter update, and now codecov/patch is knocking the coverage score for it. Many of the fixes it's penalizing were on commented lines.

Would you like me to fix these, or can we go ahead and merge? On the original PR, the coverage was unchanged.

kmax12 · 2019-01-30T14:59:31Z

@RogerTangos thanks for investigating the linting issues and resolving. we still need to review this before it's ready to merge. we're currently preparing the next release of featuretools (likely to go out this week), but after that we will return to review this.

kmax12

Thank for working on this. It is looking good overall. The main thing missing is a test case so we can verify behavior works as intended.

kmax12 · 2019-02-04T16:20:19Z

-                                          variable_type=[Numeric,
-                                                         Categorical,
-                                                         Ordinal],
+                                          variable_type=allowed_variable_types,


I think we actually shouldn't be filtering any features based on type here since happens on line 216.

That means the correct change here would be to actually to update _features_by_type to not filter based on variable_type if variable_type == None.

Good point. I'll update this.

kmax12 · 2019-02-04T16:21:47Z


-    def _run_dfs(self, entity, entity_path, all_features, max_depth):
+    def _run_dfs(self, entity, entity_path, all_features, max_depth,
+                 allowed_variable_types=None):


once we remove allowed_variable_types from _build_forward_features, we no longer need this parameter to _run_dfs

kmax12 · 2019-02-04T16:22:36Z

        dask_kwargs=None,
-        verbose=False):
+        verbose=False,
+        allowed_variable_types=None):


rename to return_variable_types per comment in deep_feature_synthesis.py

kmax12 · 2019-02-04T16:22:56Z

        dask_kwargs=None,
-        verbose=False):
+        verbose=False,
+        allowed_variable_types=None):


please add test cases for this change before we can merge

RogerTangos · 2019-02-04T17:04:17Z

Thanks @kmax12. It's the end of the day here, so I'll get to these updates and the tests tomorrow. Thank you for the review.

…is._run_dfs()

RogerTangos · 2019-02-05T14:11:02Z

@kmax12 - requested changes made.

kmax12 · 2019-02-05T14:49:03Z

                    self._handle_new_feature(new_f, all_features)

-    def _features_by_type(self, all_features, entity, variable_type, max_depth):
+    def _features_by_type(


to keep the style of our other code - let's keep this as one line or break up the arguments over two lines

RogerTangos · 2019-02-05T18:49:12Z

Updates to everything but line 212. I did actually write some code for this, but IMO, it made the file harder to understand, not easier. If you have an opinion on how that should be done, feel free to copy past and I'll add it.

CLAassistant · 2019-02-05T19:04:04Z

All committers have signed the CLA.

kmax12 · 2019-02-06T00:03:33Z

Looks good. Merging!

* Make S3 dependencies optional * Allow user to pass in desired feature return types (#372) * change method var variable_types to prevent overloading * add and pass allowed_variable_types var to _run_dfs and _build_forward_features * allowed_variable_types can be passed to dfs * correct docstrings for allowed_variable_types in dfs and deep_feature_synthesis.build_features * fix linting issues. * fix linting errors * fix another linting error * fix F632 linter error * add test. rename build_features return_variable_types arg. * remove redundant return_variable_types attr from deep_feature_synthesis._run_dfs() * add datetime tests. minor formatting change. * remove bad line break * Update deep_feature_synthesis.py * Make S3 dependencies optional * Move boto3 back to regular dependencies * Move s3fs usage into NoCredentialsError * Make S3 dependencies optional * Move boto3 back to regular dependencies * Move s3fs usage into NoCredentialsError

RogerTangos added 5 commits January 15, 2019 16:39

change method var variable_types to prevent overloading

9874abb

add and pass allowed_variable_types var to _run_dfs and _build_forwar…

30e317d

…d_features

allowed_variable_types can be passed to dfs

7b9215d

correct docstrings for allowed_variable_types in dfs and deep_feature…

04cdfea

…_synthesis.build_features

fix linting issues.

fcb677d

RogerTangos added 3 commits January 29, 2019 17:42

Merge branch 'master' into feature-return-types

1b2ca7f

fix linting errors

2914c11

fix another linting error

cbdccc9

Merge branch 'master' into feature-return-types

1d61a9f

RogerTangos added 2 commits January 31, 2019 09:57

Merge branch 'master' into feature-return-types

0476cd7

fix F632 linter error

ecae68b

RogerTangos changed the title ~~Feature return types~~ Allow user to pass in desired feature return types Feb 4, 2019

kmax12 reviewed Feb 4, 2019

View reviewed changes

RogerTangos added 4 commits February 5, 2019 12:45

add test. rename build_features return_variable_types arg.

11c4f5a

remove redundant return_variable_types attr from deep_feature_synthes…

c55dfa3

…is._run_dfs()

Merge branch 'master' into feature-return-types

215cddd

merge master

8876983

kmax12 reviewed Feb 5, 2019

View reviewed changes

RogerTangos added 2 commits February 5, 2019 19:38

add datetime tests. minor formatting change.

20f5ead

remove bad line break

493d9ff

kmax12 added 2 commits February 5, 2019 14:02

Merge branch 'master' into feature-return-types

cee4c8b

Update deep_feature_synthesis.py

c10c4b3

kmax12 merged commit d0bad62 into alteryx:master Feb 6, 2019

RogerTangos deleted the feature-return-types branch February 6, 2019 08:43

rwedge mentioned this pull request Feb 15, 2019

v0.6.1 #436

Merged

Conversation

RogerTangos commented Jan 15, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov Bot commented Jan 15, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

RogerTangos commented Jan 29, 2019

Uh oh!

RogerTangos commented Jan 30, 2019

Uh oh!

kmax12 commented Jan 30, 2019

Uh oh!

kmax12 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

kmax12 Feb 4, 2019

Choose a reason for hiding this comment

Uh oh!

RogerTangos Feb 4, 2019

Choose a reason for hiding this comment

Uh oh!

kmax12 Feb 4, 2019

Choose a reason for hiding this comment

Uh oh!

kmax12 Feb 4, 2019

Choose a reason for hiding this comment

Uh oh!

kmax12 Feb 4, 2019

Choose a reason for hiding this comment

Uh oh!

RogerTangos commented Feb 4, 2019

Uh oh!

RogerTangos commented Feb 5, 2019

Uh oh!

Uh oh!

Uh oh!

kmax12 Feb 5, 2019

Choose a reason for hiding this comment

Uh oh!

RogerTangos commented Feb 5, 2019

Uh oh!

CLAassistant commented Feb 5, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

kmax12 commented Feb 6, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

RogerTangos commented Jan 15, 2019 •

edited

Loading

codecov Bot commented Jan 15, 2019 •

edited

Loading

CLAassistant commented Feb 5, 2019 •

edited

Loading