Feature: Num Quantiles into dev #982
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* add code for reservoir sampling and insert sample_nrows options * pre commit fix * add tests for reservoir sampling * fixed mypy issues * fix import to relative path --------- Co-authored-by: Taylor Turner <[email protected]> Co-authored-by: Richard Bann <[email protected]>
* New preset implementation and test (#867) * memory optimization preset ttrying again ttrying again 3 ttrying again 4 accidentally pushed my updated makefile * Wrote catch for invalid presets, wrote test for catch for invalid presets, debugged new optimization preset * Forgot to run pre-commit, fixed those issues * black doing weird things * made preset validation more maintainable by moving it to the constructor and getting rid of preset list * RowStatisticsOptions: Add option (#865) * RowStatisticsOptions: Add null row count Added null_row_count as an option in RowStatisticsOptions. It toggles the functionality for row_has_null_ratio and row_is_null_ratio in _update_row_statistics. * Unit test for RowStatisticOptions: * Black formatting * RowStatisticsOptions: Add null row count Added null_row_count as an option in RowStatisticsOptions. It toggles the functionality for row_has_null_ratio and row_is_null_ratio in _update_row_statistics. * Unit test for RowStatisticOptions: * Black formatting * added a unit test for RowStatisticsOptions * Deleted test cases that were written in the wrong file * updated testing for null_count toggle in _update_row_statistics * removed the RowStatisticsOptions from test_profiler_options imports * add line * Created toggle option for null_count * RowStatisticsOptions: Add implementation * Revert "RowStatisticsOptions: Add implementation" This reverts commit 2da6a93. * RowStatsticsOptions: Create option * fixed pre-commit error * Update dataprofiler/profilers/profiler_options.py Co-authored-by: Taylor Turner <[email protected]> * Update dataprofiler/profilers/profiler_options.py Co-authored-by: Taylor Turner <[email protected]> * fixed documentation --------- Co-authored-by: Taylor Turner <[email protected]> * Preset test updated w new names and different toggles (#880) * memory optimization preset ttrying again ttrying again 3 ttrying again 4 accidentally pushed my updated makefile * trying * trying * black doing weird things * trying * made preset validation more maintainable by moving it to the constructor and getting rid of preset list * Update to open-source in prep for wrapper changes for mem op preset * updated preset toggles and preset name (mem op -> large data) * updated tests to match * continued name and test and toggle updates * fix comments * RowStatisticsOptions: Implementing option (#871) * Implementing option * Implementing option * took out redundant if statement. added test case for when null_count is disabled. * attempt to check for conflicts between profile merges * added test to check if two profilers have null_count enabled before merging them together * fixed typo and added a trycatch to prevent failing test * No mocks needed. Fixed assertRaisesRegex error * Changed variables names and added a new test to check for check the null_count when null_count is disabled. * Changed name of test, moved tests to TestStructuredProfilerRowStatistics. Fixed position of if statement to prevent unnecessary code from running. * added null_count test cases * fixed indentation mistake * fixed typo * removed a useless commented a line * Updated test name * update --------- Co-authored-by: Liz Smith <[email protected]> Co-authored-by: Richard Bann <[email protected]>
* WIP cms implementation * add heavy hitters implementation * add heavy hitters implementation * WIP: mypy issue * WIP: mypy issue * add cms bool and refactor options handler * WIP: testing for CMS * WIP: testing for CMS * use new heavy_hitters_threshold, add test for it
* refactored all but tests * removed some superfluous tests * moved variables around
* initial changes to categoricalColumn decoder (#818) * Implemented decoding for numerical stats mixin and integer profiles (#844) * hot fixes for encode and decode of numeric stats mixin and intcol profiler (#852) * Float column profiler encode decode (#854) * hot fixes for encode and decode of numeric stats mixin and intcol profiler * cleaned up type checking and updated numericstatsmixin readin helper to give type conversions to more attributes * Added docstring to the _load_stats_helper function * Update dataprofiler/profilers/numerical_column_stats.py Co-authored-by: Taylor Turner <[email protected]> * Update dataprofiler/profilers/numerical_column_stats.py * fix for nan values issue in pytesting * Implementation of float profiler encode and decode process --------- Co-authored-by: Taylor Turner <[email protected]> * Json decode date time column (#861) * more verbose error log with types for easy debug * add load_from_dict to handle tiimestamps * add json decode tests * include DateTimeColumn class * Added decoding for encoding of ordered column profiles (#864) * Added ordered col test to ensure correct response to update when different ordering of values is introduced (#868) * added decode text_column_profiler functionality and tests (#870) * Created encoder for the datalabelercolumn (#869) * feat: add test and compiler serialization (#884) * [WIP] Adds tests validating serialization with Primitive type for compiler (#885) * feat: add test and compiler serialization * fix: move primitive tests to own class * feat: add primitive col compiler save tests * fix: float serializers asserts * Adds deserialization for compilers and validates tests for Primitive; fixes numerical deserialization (#886) * feat: add test and compiler serialization * fix: move primitive tests to own class * feat: add primitive col compiler save tests * fix: float serializers asserts * feat: add tests and allow primitive compiler to deserialize * fix: bug in numeric stats deserial * fix: missing `)` after conflict resolution * Add Serialization and Deserialization Tests for Stats Compiler, plus refactors for order Typing (#887) * fix: organize categorical and add get function * refactor: reorganize tests and add stats test * feat: order typing * feat: add serial and deserial for stats compiler * fix: bug when sample_size == 0 * ready datalabeler for deserialization and improvement on serialization for datalabeler (#879) * Deserialization of datalabeler (#891) * Added initial profiler decoding for datalabeler column (WIP) * Intialial implementation for deserialization of datalabelercolumn * Fix LSP violations (#840) * Make profiler superclasses generic Makes the superclasses BaseColumnProfiler, NumericStatsMixin, and BaseCompiler generic, to avoid casting in subclass diff() methods and violating LSP in principle. * Add needed cast import --------- Co-authored-by: Junho Lee <[email protected]> * Encode Options (#875) * encode testing * encode dataLabeler testing * encode structuredOptions testing * cleaned up datalabeler test * added text options * [WIP] ColumnDataLabelerCompiler: serialize / deserialize (#888) * formatting * update formatting * setting up full test suite for DataLabelerCompiler * update isort * updates to test -- still failing * update * Quick Test update (#893) * update * string in list * formatting * Decode options (#894) * refactored options encode testing * updated test name * updated class names * fixing test * initial base option decode * inital tests * refactor: allow options to go through all (#902) * refactor: allow options to go through all * fix: bug * StructuredColProfiler Encode / Decode (#901) * refactor: allow options to go through all * fix: bug * update * update * update * updates * update * Fixes for taylors StructuredCol Issue * update * update * remove try/except --------- Co-authored-by: Jeremy Goodsitt <[email protected]> Co-authored-by: ksneab7 <[email protected]> * fix: bug and add tests for structuredcolprofiler (#904) * fix: bug and add tests * fix: limit scipy requirements till problem understood and fixed * Stuctured profiler encode decode (#903) * refactor: allow options to go through all * fix: bug in loading options * update * update * Fixes for taylors StructuredCol Issue * Created load and save code from structuredprofiler * intermidiate commit for fixing structured profile --------- Co-authored-by: Jeremy Goodsitt <[email protected]> Co-authored-by: taylorfturner <[email protected]> * [WIP] Added NoImplementationError for UnstructuredProfiler (#907) * refactor: allow options to go through all * fix: bug in loading options * update * update * Fixes for taylors StructuredCol Issue * Created load and save code from structuredprofiler * intermidiate commit for fixing structured profile * test fix * mypy fixes for typing issues * fix for none case of the datalabler in options * Added mock of datalabeler to structured profile test * Added tests for encoding of the Structured profiler * Update dataprofiler/profilers/json_decoder.py Co-authored-by: Michael Davis <[email protected]> * Update dataprofiler/profilers/profile_builder.py Co-authored-by: Michael Davis <[email protected]> * Update dataprofiler/profilers/profiler_options.py Co-authored-by: Michael Davis <[email protected]> * Pr fixes * Fixed typo in test * Update dataprofiler/profilers/json_decoder.py Co-authored-by: Taylor Turner <[email protected]> * Update dataprofiler/profilers/profile_builder.py Co-authored-by: Michael Davis <[email protected]> * Update dataprofiler/tests/profilers/utils.py Co-authored-by: Taylor Turner <[email protected]> * Update dataprofiler/profilers/profile_builder.py Co-authored-by: Michael Davis <[email protected]> * Fixes for unneeeded callout for _profile check * small change --------- Co-authored-by: Jeremy Goodsitt <[email protected]> Co-authored-by: taylorfturner <[email protected]> Co-authored-by: ksneab7 <[email protected]> Co-authored-by: ksneab7 <[email protected]> * Added testing for values for test_json_decode_after_update (#915) * Reuse passed labeler (#924) * refactor: loading labeler for reuse and abstract loading * refactor: use for DataLabelerColumn as well * fix: don't error if doesn't exist * refactor: allow for config dict to be passed entire way * fix: compiler tests * fix: structCol tests * fix: test * BaseProfiler save() for json (#923) * added save for top level and tests * small refactor * small fix * refactor: use seed for sample for consistency (#927) * refactor: use seed for sample for consistency * fix: formatting and variables * WIP top level load (#925) * quick hot fix for input validation on save() save_metho (#931) * BaseProfiler: `load_method` hotfix (#932) * added load_method * updated tests * fix: null_rep mat should calculate even if datetime (#933) * Notebook Example save/load Profile (#930) * update example data profiler demo save/load * update notebook cells * Update examples/data_profiler_demo.ipynb * Update examples/data_profiler_demo.ipynb * fix: order bug (#939) * fix: typo on rebase * fix: typing and bugs from rebase * fix: options tests due to merge and loading new options --------- Co-authored-by: Michael Davis <[email protected]> Co-authored-by: ksneab7 <[email protected]> Co-authored-by: Taylor Turner <[email protected]> Co-authored-by: Tyler <[email protected]> Co-authored-by: Junho Lee <[email protected]> Co-authored-by: ksneab7 <[email protected]>
* fix: to use config instead of options * fix: comment * fix: maxdiff
Inside the BaseDataProcessor class definition, references to __subclasses are automatically replaced with _BaseDataProcessor__subclasses. This remains the case even in static methods _register_subclass() and get_class(). Same with BaseModel and its __subclasses field. So we do not have to write out the full name mangled identifiers inside the class definitions. Also, mypy doesn't seem to be able to handle the return type of BaseDataProcessor.get_class() being a typevar, so that was changed to type[BaseDataProcessor]. This does not affect the functionality of get_class() since it always returns a subclass of BaseDataProcessor.
taylorfturner
added
Medium Priority
taylorfturner
requested review from
ksneab7,
micdavis and
tyfarnan
as code owners
* Fixed F1Score Import * Linted example file with Black Linter
* update * renamed var and removed from for loops * refactored var
) A method signature that uses *args: Any, **kwargs: Any is compatible with any set of arguments in mypy, despite being an LSP violation. This lets us assert that subclasses of BaseDataProcessor should have some process() method with an arbitrary signature. We also add to the return type of BaseDataPreprocessor so that it is inclusive of all of its subclasses. Co-authored-by: JGSweets <[email protected]>
…nt `num_quantiles` option to `HistogramAndQuantilesOptions` (#956) * Implemented options for `num_quantiles` feature in `NumericStatsMixin` * Implemented options for `num_quantiles.py`. * Added tests for `num_quantiles`. * Fixed tests for num_quantiles option in column profiles. * Added tests for num_quantiles option. * Update dataprofiler/tests/profilers/profiler_options/test_num_quantiles_option.py Co-authored-by: Taylor Turner <[email protected]> * Minor edits to tests * Empty commit * Precommits * Minor change to Makefile. * Got rid of print statements. * "empty commit" * Added deprecation for `HistogramOption`. * Cleaner deprecation: removed `HistogramOption` class. * json_decoder, test_h_a_q changes. * Fixed string bug for testing in test_histogram_and_quantiles * Explicitfied deprecation warning, validated profile conversion from histogramoption to histogramandquantilesoption in decoding phase. * Minor deprecation warning wording change. * Update dataprofiler/tests/profilers/profiler_options/test_histogram_and_quantiles_option.py Co-authored-by: Taylor Turner <[email protected]> * Changed deprecation warning. * Changed quantiles option to initialize to None. * Minor change * Minor change again D: * Minor change again again D: sadness --------- Co-authored-by: Taylor Turner <[email protected]>
taylorfturner
force-pushed
the
staging/dev/num-quantiles
branch
from
5bd8a9e
to
01113a2
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.