Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Data Labeler Compiler Diff #336

Merged
merged 6 commits into from
Jul 16, 2021
Merged

Data Labeler Compiler Diff #336

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
eefa567
Woah. NICE.
grant-eden Jul 14, 2021
b093627
Small refactor for the good of the universe
grant-eden Jul 15, 2021
1bf6c8c
Merge branch 'main' into datalabelercompilerdiff
grant-eden Jul 15, 2021
93d8eaf
Here we goooooo
grant-eden Jul 15, 2021
776bac6
Merge branch 'main' of https://github.com/capitalone/DataProfiler int…
grant-eden Jul 15, 2021
6d503f3
Merge branch 'datalabelercompilerdiff' of https://github.com/gme5078/…
grant-eden Jul 15, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 29 additions & 2 deletions dataprofiler/profilers/column_profile_compilers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -325,9 +325,36 @@ def profile(self):
}
# TODO: Only works for last profiler. Abstracted for now.
for _, profiler in self._profiles.items():
profile["data_label"] = profiler.data_label
profile["statistics"].update(profiler.profile)
col_profile = profiler.profile
profile["data_label"] = col_profile.pop("data_label")
profile["statistics"].update(col_profile)
return profile

def diff(self, other, options=None):
"""
Finds the difference between 2 compilers and returns the report

:param other: profile compiler finding the difference with this one.
:type other: ColumnDataLabelerCompiler
:return: difference of the profiles
:rtype: dict
"""
# Call super for compiler instance check
diff_profile = super().diff(other, options)
diff_profile["statistics"] = dict()

# Iterate through profile(s)
all_profiles = set(self._profiles.keys()) | set(other._profiles.keys())
for key in all_profiles:
if key in self._profiles and key in other._profiles:
diff = self._profiles[key].diff(other._profiles[key], options)
diff_profile["data_label"] = diff.pop("data_label")
diff_profile["statistics"].update(diff)

if not diff_profile["statistics"]:
diff_profile.pop("statistics")

return diff_profile


class UnstructuredCompiler(BaseCompiler):
Expand Down
1 change: 1 addition & 0 deletions dataprofiler/profilers/data_labeler_column_profile.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -233,6 +233,7 @@ def profile(self):
Property for profile. Returns the profile of the column.
"""
profile = {
"data_label": self.data_label,
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@JGSweets @AnhTruong I've made a small refactor to put the data_label in the profile instead of having it extracted from the compiler level. Was there a specific reason the data_label was left out of the profile? Is this refactor inappropriate?

Copy link
Contributor

@JGSweets JGSweets Jul 15, 2021 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this is appropriate.

The update above: profile["data_label"] = col_profile.pop("data_label") completes the refactor.

"avg_predictions": self.avg_predictions,
"data_label_representation": self.label_representation,
"times": self.times
Expand Down
73 changes: 73 additions & 0 deletions dataprofiler/tests/profilers/test_column_profile_compilers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -269,6 +269,79 @@ def test_compiler_stats_diff(self):
expected_diff = {}
self.assertDictEqual(expected_diff, compiler1.diff(compiler2))

@mock.patch(
'dataprofiler.profilers.data_labeler_column_profile.DataLabeler')
@mock.patch("dataprofiler.profilers.data_labeler_column_profile."
"DataLabelerColumn.update")
def test_compiler_data_labeler_diff(self, *mocked_datalabeler):
# Initialize dummy data
data = pd.Series([])

# Test normal diff
compiler1 = col_pro_compilers.ColumnDataLabelerCompiler(data)
compiler2 = col_pro_compilers.ColumnDataLabelerCompiler(data)

# Mock out the data_label, avg_predictions, and label_representation
# properties
with mock.patch("dataprofiler.profilers.data_labeler_column_profile"
".DataLabelerColumn.data_label"), \
mock.patch("dataprofiler.profilers.data_labeler_column_profile."
"DataLabelerColumn.avg_predictions"), \
mock.patch("dataprofiler.profilers.data_labeler_column_profile."
"DataLabelerColumn.label_representation"):
compiler1._profiles["data_labeler"].data_label = "a"
compiler1._profiles["data_labeler"].avg_predictions = {
"a": 0.25,
"b": 0.0,
"c": 0.75
}
compiler1._profiles["data_labeler"].label_representation = {
"a": 0.15,
"b": 0.01,
"c": 0.84
}

compiler2._profiles["data_labeler"].data_label = "b"
compiler2._profiles["data_labeler"].avg_predictions = {
"a": 0.25,
"b": 0.70,
"c": 0.05
}
compiler2._profiles["data_labeler"].label_representation = {
"a": 0.99,
"b": 0.01,
"c": 0.0
}

expected_diff = {
'statistics': {
'avg_predictions': {
'a': 'unchanged',
'b': -0.7,
'c': 0.7
},
'label_representation': {
'a': -0.84,
'b': 'unchanged',
'c': 0.84
}
},
'data_label': [['a'], [], ['b']]
}
self.assertDictEqual(expected_diff, compiler1.diff(compiler2))

# Test disabling one datalabeler profile for compiler diff
options = StructuredOptions()
options.data_labeler.is_enabled = False
compiler1 = col_pro_compilers.ColumnDataLabelerCompiler(data, options)
expected_diff = {}
self.assertDictEqual(expected_diff, compiler1.diff(compiler2))

# Test disabling both datalabeler profiles for compiler diff
compiler2 = col_pro_compilers.ColumnDataLabelerCompiler(data, options)
expected_diff = {}
self.assertDictEqual(expected_diff, compiler1.diff(compiler2))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can we add a test with one profile is empty, then the result should be the other profile?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The result wont be the other profile, it would be empty, since you cant have a difference unless you have 2 profiles to compare


@mock.patch.multiple(
col_pro_compilers.BaseCompiler, __abstractmethods__=set())
def test_no_profilers_error(self):
Expand Down
6 changes: 5 additions & 1 deletion dataprofiler/tests/profilers/test_data_labeler_column_profile.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,10 +51,11 @@ def test_base_case(self, mock_instance):
self.assertEqual(None, profiler.avg_predictions)
six.assertCountEqual(
self,
["avg_predictions", "data_label_representation", "times"],
["data_label", "avg_predictions", "data_label_representation", "times"],
list(profiler.profile.keys())
)
self.assertEqual({
"data_label": None,
"avg_predictions": None,
"data_label_representation": None,
"times": defaultdict()
Expand Down Expand Up @@ -140,6 +141,7 @@ def test_profile(self, mock_instance):
profiler = DataLabelerColumn(data.name)

expected_profile = {
"data_label": 'a',
"avg_predictions": dict(a=2/3, b=1/3),
"data_label_representation": dict(a=2/3, b=1/3),
"times": defaultdict(float, {'data_labeler_predict': 1.0})
Expand Down Expand Up @@ -187,6 +189,7 @@ def test_profile_merge(self, mock_instance):
data2 = pd.Series(['4', '5', '6', '7', '9', '10', '12'])

expected_profile = {
"data_label": "a|b",
"avg_predictions": dict(a=54 / 99, b=45 / 99),
"data_label_representation": dict(a=54 / 99, b=45 / 99),
"times": defaultdict(float, {'data_labeler_predict': 2.0})
Expand Down Expand Up @@ -229,6 +232,7 @@ def test_profile_merge(self, mock_instance):
# Check adding even more profiles together
profiler3 = profiler + profiler3
expected_profile = {
"data_label": "a|b",
"avg_predictions": dict(a=8 / 15, b=7 / 15),
"data_label_representation": dict(a=8 / 15, b=7 / 15),
"times": defaultdict(float, {'data_labeler_predict': 3.0})
Expand Down