Skip to main content

Reversible Data Transforms

Project description


This repository is part of The Synthetic Data Vault Project, a project from DataCebo.

Development Status PyPi Shield Unit Tests Downloads Coverage Status

Overview

RDT is a Python library used to transform data for data science libraries and preserve the transformations in order to revert them as needed.

Important Links
:computer: Website Check out the SDV Website for more information about the project.
:orange_book: SDV Blog Regular publshing of useful content about Synthetic Data Generation.
:book: Documentation Quickstarts, User and Development Guides, and API Reference.
:octocat: Repository The link to the Github Repository of this library.
:scroll: License The entire ecosystem is published under the MIT License.
:keyboard: Development Status This software is in its Alpha stage.
Community Join our Slack Workspace for announcements and discussions.
Tutorials Run the RDT Tutorials in a notebook.

Install

RDT is part of the SDV project and is automatically installed alongside it. For details about this process please visit the SDV Installation Guide

Optionally, RDT can also be installed as a standalone library using the following commands:

Using pip:

pip install rdt

Using conda:

conda install -c conda-forge rdt

For more installation options please visit the RDT installation Guide

Quickstart

In this short series of tutorials we will guide you through a series of steps that will help you getting started using RDT to transform columns, tables and datasets.

Load the demo data

After you have installed RDT, you can get started using the demo dataset.

from rdt import get_demo

customers = get_demo()

This dataset contains some randomly generated values that describes the customers of an online marketplace.

  last_login email_optin credit_card  age  dollars_spent
0 2021-06-26       False        VISA   29          99.99
1 2021-02-10       False        VISA   18            NaN
2        NaT       False        AMEX   21           2.50
3 2020-09-26        True         NaN   45          25.00
4 2020-12-22         NaN    DISCOVER   32          19.99

Let's transform this data so that each column is converted to full, numerical data ready for data science.

Creating the HyperTransformer & config

The HyperTransformer is capable of transforming multi-column datasets.

from rdt import HyperTransformer

ht = HyperTransformer()

The HyperTransformer needs to know about the columns in your dataset and which transformers to apply to each. These are described by a config. We can ask the HyperTransformer to automatically detect it based on the data we plan to use.

ht.detect_initial_config(data=customers)

This will create and set the config.

Config:
{
    "sdtypes": {
        "last_login": "datetime",
        "email_optin": "boolean",
        "credit_card": "categorical",
        "age": "numerical",
        "dollars_spent": "numerical"
    },
    "transformers": {
        "last_login": "UnixTimestampEncoder(missing_value_replacement='mean')",
        "email_optin": "BinaryEncoder(missing_value_replacement='mode')",
        "credit_card": "FrequencyEncoder()",
        "age": "FloatFormatter(missing_value_replacement='mean')",
        "dollars_spent": "FloatFormatter(missing_value_replacement='mean')"
    }
}

The sdtypes dictionary describes the semantic data types of each of your columns and the transformers dictionary describes which transformer to use for each column.

Fitting & using the HyperTransformer

The HyperTransformer references the config while learning the data during the fit stage.

ht.fit(customers)

Once the transformer is fit, it's ready to use. Use the transform method to transform all columns of your dataset at once.

transformed_data = ht.transform(customers)
   last_login.value  email_optin.value  credit_card.value  age.value  dollars_spent.value
0      1.624666e+18                0.0                0.2         29                99.99
1      1.612915e+18                0.0                0.2         18                36.87
2      1.611814e+18                0.0                0.5         21                 2.50
3      1.601078e+18                1.0                0.7         45                25.00
4      1.608595e+18                0.0                0.9         32                19.99

The HyperTransformer applied the assigned transformer to each individual column. Each column now contains fully numerical data that you can use for your project!

When you're done with your project, you can also transform the data back to the original format using the reverse_transform method.

original_format_data = ht.reverse_transform(transformed_data)
  last_login email_optin credit_card  age  dollars_spent
0        NaT       False        VISA   29          99.99
1 2021-02-10       False        VISA   18            NaN
2        NaT       False        AMEX   21            NaN
3 2020-09-26        True         NaN   45          25.00
4 2020-12-22       False    DISCOVER   32          19.99

Transforming a single column

It is also possible to transform a single column of a pandas.DataFrame. To do this, follow the following steps.

Load the transformer

In this example we will use the datetime column, so let's load a UnixTimestampEncoder.

from rdt.transformers import UnixTimestampEncoder

transformer = UnixTimestampEncoder()

Fit the Transformer

Before being able to transform the data, we need the transformer to learn from it.

We will do this by calling its fit method passing the column that we want to transform.

transformer.fit(customers, column='last_login')

Transform the data

Once the transformer is fitted, we can pass the data again to its transform method in order to get the transformed version of the data.

transformed = transformer.transform(customers)

The output will be a pandas.DataFrame similar to the input data, except with the original datetime column replaced with last_login.value.

  email_optin credit_card  age  dollars_spent  last_login.value
0       False        VISA   29          99.99      1.624666e+18
1       False        VISA   18            NaN      1.612915e+18
2       False        AMEX   21           2.50               NaN
3        True         NaN   45          25.00      1.601078e+18
4         NaN    DISCOVER   32          19.99      1.608595e+18

Revert the column transformation

In order to revert the previous transformation, the transformed data can be passed to the reverse_transform method of the transformer:

reversed_data = transformer.reverse_transform(transformed)

The output will be a pandas.DataFrame containing the reverted values, which should be exactly like the original ones, except for the order of the columns.

  email_optin credit_card  age  dollars_spent last_login
0       False        VISA   29          99.99 2021-06-26
1       False        VISA   18            NaN 2021-02-10
2       False        AMEX   21           2.50        NaT
3        True         NaN   45          25.00 2020-09-26
4         NaN    DISCOVER   32          19.99 2020-12-22



The Synthetic Data Vault Project was first created at MIT's Data to AI Lab in 2016. After 4 years of research and traction with enterprise, we created DataCebo in 2020 with the goal of growing the project. Today, DataCebo is the proud developer of SDV, the largest ecosystem for synthetic data generation & evaluation. It is home to multiple libraries that support synthetic data, including:

  • 🔄 Data discovery & transformation. Reverse the transforms to reproduce realistic data.
  • 🧠 Multiple machine learning models -- ranging from Copulas to Deep Learning -- to create tabular, multi table and time series data.
  • 📊 Measuring quality and privacy of synthetic data, and comparing different synthetic data generation models.

Get started using the SDV package -- a fully integrated solution and your one-stop shop for synthetic data. Or, use the standalone libraries for specific needs.

History

0.6.4 - 2022-3-7

This release fixes multiple bugs concerning the HyperTransformer. One is that the get_transformer_tree_yaml method no longer crashes on every call. Another is that calling the update_field_data_types and update_default_data_type_transformers after fitting no longer breaks the transform method.

The HyperTransformer now sorts its outputs for both transform and reverse_transform based on the order of the input's columns. It is also now possible to create transformers that simply drops columns during transform and don't return any new columns.

New Features

  • Support dropping a column trough a transformer - Issue #393 by @pvk-developer
  • HyperTransformer should sort columns after transform and reverse_transform - Issue #405 by @fealho

Bugs

  • get_transformer_tree_yaml fails - Issue #389 by @amontanez24
  • HyperTransformer _unfit method not working correctly - Issue #390 by @amontanez24
  • Blank dataframe after updating the data types - Issue #401 by @amontanez24

0.6.3 - 2022-2-4

This release adds a new module to the RDT library called performance. This module can be used to evaluate the speed and peak memory usage of any transformer in RDT. This release also increases the maximum acceptable version of scikit-learn to make it more compatible with other libraries in the SDV ecosystem. On top of that, it fixes a bug related to a new version of pandas.

New Features

  • Move profiling functions into RDT library - Issue #353 by @amontanez24

Housekeeping

  • Increase scikit-learn dependency range - Issue #351 by @amontanez24
  • pandas 1.4.0 release causes a small error - Issue #358 by @fealho

Bugs

  • Performance tests get stuck on Unix if multiprocessing is involved - Issue #337 by @amontanez24

0.6.2 - 2021-12-28

This release adds a new BayesGMMTransformer. This transformer can be used to convert a numerical column into two columns: a discrete column indicating the selected component of the GMM for each row, and a continuous column containing the normalized value of each row based on the mean and std of the selected component. It is useful when the column being transformed came from multiple distributions.

This release also adds multiple new methods to the HyperTransformer API. These allow for users to access the specfic transformers used on each input field, as well as view the entire tree of transformers that are used when running transform. The exact methods are:

  • BaseTransformer.get_input_columns() - Return list of input columns for a transformer.
  • BaseTransformer.get_output_columns() - Return list of output columns for a transformer.
  • HyperTransformer.get_transformer(field) - Return the transformer instance used for a field.
  • HyperTransformer.get_output_transformers(field) - Return dictionary mapping output columns of a field to the transformers used on them.
  • HyperTransformer.get_final_output_columns(field) - Return list of all final output columns related to a field.
  • HyperTransformer.get_transformer_tree_yaml() - Return YAML representation of transformers tree.

Additionally, this release fixes a bug where the HyperTransformer was incorrectly raising a NotFittedError. It also improved the DatetimeTransformer by autonomously detecting if a column needs to be converted from dtype object to dtype datetime.

New Features

  • Cast column to datetime if specified in field transformers - Issue #321 by @amontanez24
  • Add a BayesianGMM Transformer - Issue #183 by @fealho
  • Add transformer tree structure and traversal methods - Issue #330 by @amontanez24

Bugs fixed

  • HyperTransformer raises NotFittedError after fitting - Issue #332 by @amontanez24

0.6.1 - 2021-11-10

This release adds support for Python 3.9! It also removes unused document files.

Internal Improvements

  • Add support for Python 3.9 - Issue #323 by @amontanez24
  • Remove docs - PR #322 by @pvk-developer

0.6.0 - 2021-10-29

This release makes major changes to the underlying code for RDT as well as the API for both the HyperTransformer and BaseTransformer. The changes enable the following functionality:

  • The HyperTransformer can now apply a sequence of transformers to a column.
  • Transformers can now take multiple columns as an input.
  • RDT has been expanded to allow for infinite data types to be added instead of being restricted to pandas.dtypes.
  • Users can define acceptable output types for running HyperTransformer.transform.
  • The HyperTransformer will continuously apply transformations to the input fields until only acceptable data types are in the output.
  • Transformers can return data of any data type.
  • Transformers now have named outputs and output types.
  • Transformers can suggest which transformer to use on any of their outputs.

To take advantage of this functionality, the following API changes were made:

  • The HyperTransformer has new initialization parameters that allow users to specify data types for any field in their data as well as specify which transformer to use for a field or data type. The parameters are:
    • field_transformers - A dictionary allowing users to specify which transformer to use for a field or derived field. Derived fields are fields created by running transform on the input data.
    • field_data_types - A dictionary allowing users to specify the data type of a field.
    • default_data_type_transformers - A dictionary allowing users to specify the default transformer to use for a data type.
    • transform_output_types - A dictionary allowing users to specify which data types are acceptable for the output of transform. This is a result of the fact that transformers can now be applied in a sequence, and not every transformer will return numeric data.
  • Methods were also added to the HyperTransformer to allow these parameters to be modified. These include get_field_data_types, update_field_data_types, get_default_data_type_transformers, update_default_data_type_transformers and set_first_transformers_for_fields.
  • The BaseTransformer now requires the column names it will transform to be provided to fit, transform and reverse_transform.
  • The BaseTransformer added the following method to allow for users to see its output fields and output types: get_output_types.
  • The BaseTransformer added the following method to allow for users to see the next suggested transformer for each output field: get_next_transformers.

On top of the changes to the API and the capabilities of RDT, many automated checks and tests were also added to ensure that contributions to the library abide by the current code style, stay performant and result in data of a high quality. These tests run on every push to the repository. They can also be run locally via the following functions:

  • validate_transformer_code_style - Checks that new code follows the code style.
  • validate_transformer_quality - Tests that new transformers yield data that maintains relationships between columns.
  • validate_transformer_performance - Tests that new transformers don't take too much time or memory.
  • validate_transformer_unit_tests - Checks that the unit tests cover all new code, follow naming conventions and pass.
  • validate_transformer_integration - Checks that the integration tests follow naming conventions and pass.

New Features

  • Update HyperTransformer API - Issue #298 by @amontanez24
  • Create validate_pull_request function - Issue #254 by @pvk-developer
  • Create validate_transformer_unit_tests function - Issue #249 by @pvk-developer
  • Create validate_transformer_performance function - Issue #251 by @katxiao
  • Create validate_transformer_quality function - Issue #253 by @amontanez24
  • Create validate_transformer_code_style function - Issue #248 by @pvk-developer
  • Create validate_transformer_integration function - Issue #250 by @katxiao
  • Enable users to specify transformers to use in HyperTransformer - Issue #233 by @amontanez24 and @csala
  • Addons implementation - Issue #225 by @pvk-developer
  • Create ways for HyperTransformer to know which transformers to apply to each data type - Issue #232 by @amontanez24 and @csala
  • Update categorical transformers - PR #231 by @fealho
  • Update numerical transformer - PR #227 by @fealho
  • Update datetime transformer - PR #230 by @fealho
  • Update boolean transformer - PR #228 by @fealho
  • Update null transformer - PR #229 by @fealho
  • Update the baseclass - PR #224 by @fealho

Bugs fixed

  • If the input data has a different index, the reverse transformed data may be out of order - Issue #277 by @amontanez24

Documentation changes

  • RDT contributing guide - Issue #301 by @katxiao and @amontanez24

Internal improvements

  • Add PR template for new transformers - Issue #307 by @katxiao
  • Implement Quality Tests for Transformers - Issue #252 by @amontanez24
  • Update performance test structure - Issue #257 by @katxiao
  • Automated integration test for transformers - Issue #223 by @katxiao
  • Move datasets to its own module - Issue #235 by @katxiao
  • Fix missing coverage in rdt unit tests - Issue #219 by @fealho
  • Add repo-wide automation - Issue #309 by @katxiao

Other issues closed

  • DeprecationWarning: np.float is a deprecated alias for the builtin float - Issue #304 by @csala
  • Add pip check to CI workflows - Issue #290 by @csala
  • Should Transformers subclasses exist for specific configurations? - Issue #243 by @fealho

0.5.3 - 2021-10-07

This release fixes a bug with learning rounding digits in the NumericalTransformer, and includes a few housekeeping improvements.

Issues closed

  • Update learn rounding digits to handle all nan data - Issue #244 by @katxiao
  • Adapt to latest PyLint housekeeping - Issue #216 by @fealho

0.5.2 - 2021-08-16

This release fixes a couple of bugs introduced by the previous release regarding the OneHotEncodingTransformer and the BooleanTransformer.

Issues closed

  • BooleanTransformer.reverse_transform sometimes crashes with TypeError - Issue #210 by @katxiao
  • OneHotEncodingTransformer causing shape misalignment in CopulaGAN, CTGAN, and TVAE - Issue #208 by @sarahmish
  • Boolean.transformer.reverse_transform modifies the input data - Issue #211 by @katxiao

0.5.1 - 2021-08-11

This release improves the overall performance of the library, both in terms of memory and time consumption. More specifically, it makes the following modules more efficient: NullTransformer, DatetimeTransformer, LabelEncodingTransformer, NumericalTransformer, CategoricalTransformer, BooleanTransformer and OneHotEncodingTransformer.

It also adds performance-based testing and a script for profiling the performance.

Issues closed

  • Add performance-based testing - Issue #194 by @amontanez24
  • Audit the NullTransformer - Issue #192 by @amontanez24
  • Audit DatetimeTransformer - Issue #189 by @sarahmish
  • Audit the LabelEncodingTransformer - Issue #184 by @amontanez24
  • Audit the NumericalTransformer - Issue #181 by @fealho
  • Audit CategoricalTransformer - Issue #180 by @katxiao
  • Audit BooleanTransformer - Issue #179 by @katxiao
  • Auditing OneHotEncodingTransformer - Issue #178 by @sarahmish
  • Create script for profiling - Issue #176 by @amontanez24
  • Create folder structure for performance testing - Issue #174 by @amontanez24

0.5.0 - 2021-07-12

This release updates the NumericalTransformer by adding a new rounding argument. Users can now obtain numerical values with precision, either pre-specified or automatically computed from the given data.

Issues closed

  • Add rounding argument to NumericalTransformer - Issue #166 by @amontanez24 and @csala
  • NumericalTransformer rounding error with infinity - Issue #169 by @amontanez24
  • Add min and max arguments to NumericalTransformer - Issue #106 by @amontanez24

0.4.2 - 2021-06-08

This release adds a new method to the CategoricalTransformer to solve a bug where the transformer becomes unusable after being pickled and unpickled if it had NaN values in the data which it was fit on.

It also fixes some grammar mistakes in the documentation.

Issues closed

  • CategoricalTransformer with NaN values cannot be pickled bug - Issue #164 by @pvk-developer and @csala

Documentation changes

  • docs: fix typo - PR #163 by @sbrugman

0.4.1 - 2021-03-29

This release improves the HyperTransformer memory usage when working with a high number of columns or a high number of categorical values when using one hot encoding.

Issues closed

  • Boolean, Datetime and LabelEncoding transformers fail with 2D ndarray - Issue #160 by @pvk-developer
  • HyperTransformer: Memory usage increase when reverse_transform is called - Issue #156 by @pvk-developer and @AnupamaGangadhar

0.4.0 - 2021-02-24

In this release a change in the HyperTransformer allows using it to transform and reverse transform a subset of the columns seen during training.

The anonymization functionality which was deprecated and not being used has also been removed along with the Faker dependency.

Issues closed

  • Allow the HyperTransformer to be used on a subset of the columns - Issue #152 by @csala
  • Remove faker - Issue #150 by @csala

0.3.0 - 2021-01-27

This release changes the behavior of the HyperTransformer to prevent it from modifying any column in the given DataFrame if the transformers dictionary is passed empty.

Issues closed

  • If transformers is an empty dict, do nothing - Issue #149 by @csala

0.2.10 - 2020-12-18

This release adds a new argument to the HyperTransformer which gives control over which transformers to use by default for each dtype if no specific transformer has been specified for the field.

This is also the first version to be officially released on conda.

Issues closed

  • Add dtype_transformers argument to HyperTransformer - Issue #148 by @csala
  • Makes Copulas an optional dependency - Issue #144 by @fealho

0.2.9 - 2020-11-27

This release fixes a bug that prevented the CategoricalTransformer from working properly when being passed data that contained numerical data only, without any strings, but also contained None or NaN values.

Issues closed

  • KeyError: nan - CategoricalTransformer fails on numerical + nan data only - Issue #142 by @csala

0.2.8 - 2020-11-20

This release fixes a few minor bugs, including some which prevented RDT from fully working on Windows systems.

Thanks to this fixes, as well as a new testing infrastructure that has been set up, from now on RDT is officially supported on Windows systems, as well as on the Linux and macOS systems which were previously supported.

Issues closed

  • TypeError: unsupported operand type(s) for: 'NoneType' and 'int' - Issue #132 by @csala
  • Example does not work on Windows - Issue #114 by @csala
  • OneHotEncodingTransformer producing all zeros - Issue #135 by @fealho
  • OneHotEncodingTransformer support for lists and lists of lists - Issue #137 by @fealho

0.2.7 - 2020-10-16

In this release we drop the support for the now officially dead Python 3.5 and introduce a new feature in the DatetimeTransformer which reduces the dimensionality of the generated numerical values while also ensuring that the reverted datetimes maintain the same level as time unit precision as the original ones.

  • Drop Py35 support - Issue #129 by @csala
  • Add option to drop constant parts of the datetimes - Issue #130 by @csala

0.2.6 - 2020-10-05

  • Add GaussianCopulaTransformer - Issue #125 by @csala
  • dtype category error - Issue #124 by @csala

0.2.5 - 2020-09-18

Miunor bugfixing release.

Bugs Fixed

  • Handle NaNs in OneHotEncodingTransformer - Issue #118 by @csala
  • OneHotEncodingTransformer fails if there is only one category - Issue #119 by @csala
  • All NaN column produces NaN values enhancement - Issue #121 by @csala
  • Make the CategoricalTransformer learn the column dtype and restore it back - Issue #122 by @csala

0.2.4 - 2020-08-08

General Improvements

  • Support Python 3.8 - Issue #117 by @csala
  • Support pandas >1 - Issue #116 by @csala

0.2.3 - 2020-07-09

  • Implement OneHot and Label encoding as transformers - Issue #112 by @csala

0.2.2 - 2020-06-26

Bugs Fixed

  • Escape column_name in hypertransformer - Issue #110 by @csala

0.2.1 - 2020-01-17

Bugs Fixed

  • Boolean Transformer fails to revert when there are NO nulls - Issue #103 by @JDTheRipperPC

0.2.0 - 2019-10-15

This version comes with a brand new API and internal implementation, removing the old metadata JSON from the user provided arguments, and making each transformer work only with pandas.Series of their corresponding data type.

As part of this change, several transformer names have been changed and a new BooleanTransformer and a feature to automatically decide which transformers to use based on dtypes have been added.

Unit test coverage has also been increased to 100%.

Special thanks to @JDTheRipperPC and @csala for the big efforts put in making this release possible.

Issues

  • Drop the usage of meta - Issue #72 by @JDTheRipperPC
  • Make CatTransformer.probability_map deterministic - Issue #25 by @csala

0.1.3 - 2019-09-24

New Features

  • Add attributes NullTransformer and col_meta - Issue #30 by @ManuelAlvarezC

General Improvements

  • Integrate with CodeCov - Issue #89 by @csala
  • Remake Sphinx Documentation - Issue #96 by @JDTheRipperPC
  • Improve README - Issue #92 by @JDTheRipperPC
  • Document RELEASE workflow - Issue #93 by @JDTheRipperPC
  • Add support to Python 3.7 - Issue #38 by @ManuelAlvarezC
  • Create way to pass HyperTransformer table dict - Issue #45 by @ManuelAlvarezC

0.1.2

  • Add a numerical transformer for positive numbers.
  • Add option to anonymize data on categorical transformer.
  • Move the col_meta argument from method-level to class-level.
  • Move the logic for missing values from the transformers into the HyperTransformer.
  • Removed unreacheble lines in NullTransformer.
  • Numbertransfomer to set default value to 0 when the column is null.
  • Add a CLA for collaborators.
  • Refactor performance-wise the transformers.

0.1.1

  • Improve handling of NaN in NumberTransformer and CatTransformer.
  • Add unittests for HyperTransformer.
  • Remove unused methods get_types and impute_table from HyperTransformer.
  • Make NumberTransformer enforce dtype int on integer data.
  • Make DTTransformer check data format before transforming.
  • Add minimal API Reference.
  • Merge rdt.utils into HyperTransformer class.

0.1.0

  • First release on PyPI.

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

rdt-1.0.0.dev0.tar.gz (143.4 kB view hashes)

Uploaded Source

Built Distribution

rdt-1.0.0.dev0-py2.py3-none-any.whl (52.6 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page