whylogs: A Data and Machine Learning Logging Standard

Last update: Jan 06, 2023

Overview

whylogs: A Data and Machine Learning Logging Standard

whylogs is an open source standard for data and ML logging

whylogs logging agent is the easiest way to enable logging, testing, and monitoring in an ML/AI application. The lightweight agent profiles data in real time, collecting thousands of metrics from structured data, unstructured data, and ML model predictions with zero configuration.

whylogs can be installed in any Python, Java or Spark environment; it can be deployed as a container and run as a sidecar; or invoked through various ML tools (see integrations).

whylogs is designed by data scientists, ML engineers and distributed systems engineers to log data in the most cost-effective, scalable and accurate manner. No sampling. No post-processing. No manual configurations.

whylogs is released under the Apache 2.0 open source license. It supports many languages and is easy to extend. This repo contains the whylogs CLI, language SDKs, and individual libraries are in their own repos.

This repository contains both a Python implementation and a Java implementation.

If you have any questions, comments, or just want to hang out with us, please join our Slack channel.

Getting started
Features
Data Types
Integrations
Examples
Community
Roadmap
Contribute

Getting started

Using pip

Install whylogs using the pip package manager by running

pip install whylogs

From source

Download the source code by cloning the repository or by pressing Download ZIP on this page.
You'll need to install poetry in order to install dependencies using the lock file in this project. Follow their docs to get it set up.
Run the following command at the root of the source code:

make install # installs dependencies
make         # builds the wheel

Quickly Logging Data

whylogs is easy to get up and runnings

from whylogs import get_or_create_session
import pandas as pd

session = get_or_create_session()

df = pd.read_csv("path/to/file.csv")

with session.logger(dataset_name="my_dataset") as logger:
    
    #dataframe
    logger.log_dataframe(df)

    #dict
    logger.log({"name": 1})

    #images
    logger.log_image("path/to/image.png")

whylogs collects approximate statistics and sketches of data on a column-basis into a statistical profile. These metrics include:

Simple counters: boolean, null values, data types.
Summary statistics: sum, min, max, median, variance.
Unique value counter or cardinality: tracks an approximate unique value of your feature using HyperLogLog algorithm.
Histograms for numerical features. whyLogs binary output can be queried to with dynamic binning based on the shape of your data.
Top frequent items (default is 128). Note that this configuration affects the memory footprint, especially for text features.

Multiple Profile Plots

To view your logger profiles you can use, methods within whylogs.viz:

") figure.savefig("/my/image/path.png")">
vizualization = ProfileVisualizer()
vizualization.set_profiles([profile_day_1, profile_day_2])
figure= vizualization.plot_distribution("
    
     "
    )
figure.savefig("/my/image/path.png")

Individual profiles are saved to disk, AWS S3, or WhyLabs API, automatically when loggers are closed, per the configuration found in the Session configuration.

Current profiles from active loggers can be loaded from memory with:

profile = logger.profile()

Profile Viewer

You can also load a local profile viewer, where you upload the json summary file. The default path for the json files is set as output/{dataset_name}/{session_id}/json/dataset_profile.json.

from whylogs.viz import profile_viewer
profile_viewer()

This will open a viewer on your default browser where you can load a profile json summary, using the Select JSON profile button: Once the json is selected you can view your profile's features and associated and statistics.

Documentation

The documentation of this package is generated automatically.

Features

Accurate data profiling: whylogs calculates statistics from 100% of the data, never requiring sampling, ensuring an accurate representation of data distributions
Lightweight runtime: whylogs utilizes approximate statistical methods to achieve minimal memory footprint that scales with the number of features in the data
Any architecture: whylogs scales with your system, from local development mode to live production systems in multi-node clusters, and works well with batch and streaming architectures
Configuration-free: whylogs infers the schema of the data, requiring zero manual configuration to get started
Tiny storage footprint: whylogs turns data batches and streams into statistical fingerprints, 10-100MB uncompressed
Unlimited metrics: whylogs collects all possible statistical metrics about structured or unstructured data

Data Types

Whylogs supports both structured and unstructured data, specifically:

Data type	Features	Notebook Example
Structured Data	Distribution, cardinality, schema, counts, missing values	Getting started with structured data
Images	exif metadata, derived pixels features, bounding boxes	Getting started with images
Video	In development	Github Issue #214
Tensors	derived 1d features (more in developement)	Github Issue #216
Text	top k values, counts, cardinality	String Features
Audio	In developement	Github Issue #212

Integrations

current integration

Integration	Features	Resources
Spark	Run whylogs in Apache Spark environment	Code Example
Pandas	Log and monitor any pandas dataframe	Notebook Example whylogs: Embrace Data Logging
Kafka	Log and monitor Kafka topics with whylogs	Notebook Example Integrating whylogs into your Kafka ML Pipeline
MLflow	Enhance MLflow metrics with whylogs:	Notebook Example Streamlining data monitoring with whylogs and MLflow
Github actions	Unit test data with whylogs and github actions	Notebook Example
RAPIDS	Use whylogs in RAPIDS environment	Notebook Example Monitoring High-Performance Machine Learning Models with RAPIDS and whylogs
Java	Run whylogs in Java environment	Notebook Example
Docker	Run whylogs as in Docker	Rest Container
AWS S3	Store whylogs profiles in S3	S3 example

Examples

For a full set of our examples, please check out whylogs-examples.

Check out our example notebooks with Binder:

Roadmap

whylogs is maintained by WhyLabs.

Community

If you have any questions, comments, or just want to hang out with us, please join our Slack channel.

Contribute

We welcome contributions to whylogs. Please see our contribution guide and our development guide for details.

Comments

Some Example Notebooks are out date or have unclear UX

Problem

Some of the sample note books have gotten outdated and need modification. Some may need to be updated, others are duplicated, and some may not serve a clear purpose.

This issue is to track through the comments those that should be looked into so the knowledge is not lost or siloed.

Fixes

Update through comments with PR or decisions made.

Related to #508
bug maintenance stale :zzz:

opened by TheMellyBee 24
Link to our doc/nbviewer for example notebooks
Description

The constraint violation report is not being rendered (end of notebook, cells 35+) when viewing the example notebook in Github:

Compare this with the collab (or local) render:

I Googled around but the causes of render issues are so legion it's hard to know what might be causing this. The underling cell has quite a bit of javascript - maybe there's some cross site library issue (totally out of my subject area here)?

This is suboptimal because folks won't be able to see what the constraint report looks like, without going to the collab site or running the notebook themselves.

If we can't find a fix, maybe at least a note pointing readers to the collab site?

[x] I have reviewed the Guidelines for Contributing and the Code of Conduct.

wontfix
opened by jghoman 9

int values are being stringified with floating points

Description

import pandas as pd

df = pd.DataFrame({'int-only': [1, 1, 1]})

print(why.log(df).view().to_pandas()["frequent_items/frequent_strings"])

Expected to see only "1", not "1.0000"

The actual output:

column
int-only    [FrequentItem(value='1.000000', est=3, upper=3...
Name: frequent_items/frequent_strings, dtype: object

I suspect somehow the Python code is passing the numpy array as float type, but I'm not sure

bug stale :zzz:

opened by andyndang 8

[Windows/conda/python3.8] Unable to import whylogs : No matching distribution found for whylabs-datasketches>=2.2.0b1

Hi, i was successfully able to install whylogs using : pip install whylogs. but when i try to import it in my notebooks, i am facing this error Error: No module named 'google.protobuf.pyext._message'

OS : windows 10, anaconda env python : 3.8 whylogs version : 0.1.5

i tried uninstalling protobuf and re installing it again. current version of protobuf = 3.17.3

Please help me resolve this error

opened by dalavayi 8
updated glossary for v1
Description

updated glossary for v1

General Checklist

[ ] Tests added for this feature/bug if it was a bug, test must cover it.

[ ] Conform by the style guides, by using formatter

[ ] Documentation updated

[ ] (optional) Please add a label to your PR
opened by FelipeAdachi 7
Support custom mergeable metrics in whylogs
There are three kinds of metrics that whylogs users track:

Tracking derived metrics from customers. Typically this is numerical data. You can use the approach above to track these metrics because they will show up as a “whylogs” column

Custom metrics that are mergeable: basically if you have metrics that can be “summed” or “aggregated” across different profiles, this is a feature request that we are tracking from other customers as well.

One-off metrics: sometimes users have one-off metrics that they want to piggy back on top of whylogs. These metrics are not aggregatable, but they want to use whylogs object to store these metrics.

feature v1
opened by andyndang 7
The code repository for `whylogs-sketches` as mentioned on PyPI is NONACCESSIBLE
Description

whylogs has a dependency: whylogs-sketches.

When you try accessing the code repository of whylogs-sketches on PyPI (https://pypi.org/project/whylogs-sketching), it points to the following repository, which DOES NOT EXIST or IS NOT ACCESSIBLE.

https://github.com/whylabs/whylogs-sketching

Please update the metadata to point to the correct repository. (Or, is the repository a private one?)

[x] I have reviewed the Guidelines for Contributing and the Code of Conduct.

question integrations stale :zzz:
opened by sugatoray 6
Break metrics.py <-> schema.py circular dependency
Description

General Checklist

[ ] Tests added for this feature/bug if it was a bug, test must cover it.

[ ] Conform by the style guides, by using formatter

[ ] Documentation updated

[ ] (optional) Please add a label to your PR
opened by richard-rogers 6
Create requirements.txt
The github CI for the docs keeps running into issues due to the lack of a doc/requirements.txt. This hopefully will fix that

General Checklist

[X] Conform by the style guides, by using formatter

[X] (optional) Please add a label to your PR

documentation
opened by TheMellyBee 6
Unable to import whylogs
Description

Unable to import whylogs

>>> import whylogs as why gives AttributeError: module 'site' has no attribute 'getsitepackages'

Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/home/monk/why/lib/python3.7/site-packages/whylogs/__init__.py", line 29, in <module> from .api.usage_stats import emit_usage as __emit_usage_stats File "/home/monk/why/lib/python3.7/site-packages/whylogs/api/usage_stats/__init__.py", line 30, in <module> _SITE_PACKAGES = site.getsitepackages() AttributeError: module 'site' has no attribute 'getsitepackages'

System info

Operating system and version: Ubuntu 18.04

Tried on Python3.7, Python3.8, Python3.9

virtualenv

Steps to reproduce:

Create a virtualenv -- virtualenv why --python=python3.7

activate the virtualenv -- source why/bin/activate

pip install whylogs

python

import whylogs

Although I was able to install this on colab. I don't know why it's not working on my local machine.

[x] I have reviewed the Guidelines for Contributing and the Code of Conduct.
opened by mayankjobanputra 5
mypy is not catching signature issues in core classes in our CI

Description

mypy is enabled in our pre-commit hooks but seems to be missing issues in the signatures of classes such as: DatasetProfile's get_default_path method has no type annotation for the input parameter. Why is this not showing up in pre-commit checks?
stale :zzz: workflow

opened by jamie256 5
make lint-fix doesn't fix linting issue
Description

This repo: https://github.com/whylabs/whylogs/tree/2dd782a3e284aa5063c5a4b0738407b7244ec6e2

make lint-fix is a no op locally

make format-fix doesn'tchange anything

However the CI is faiilng

- files were modified by this hook reformatted python/tests/core/view/test_dataset_profile.py All done! ✨ 🍰 ✨
opened by andyndang 1
Multi-column constraint proposal
Description

Refactor to support multi-column constraints

Changes

Adds DatasetConstraint and some other refactoring

A DatasetConstraint implements constraints that aren't attached to a specific column. It has a Callable that takes a DatasetProfileView. It returns a bool and a dictionary mapping column_name/metric_name to the metrics used in evaluating the constraint. The latter is necessary for generating the report.

It looks like there's a similar case of a MetricConstraint that applies to all metrics of a specified type across all columns. ~I've changed the implementation of this to find the set of metrics the constraint applies to by using the MetricSelector, then attaching the constraint to each specific column. This might not be the best approach... maybe converting it to a DatasetConstraint would be better.~ MetricConstraintWrapper turns such a MetricConstraint into the Callable for an equivalent DatasetConstraint.

I'm implementing a PrefixCondition that implements the prefix expression syntax used to serialize the Predicate expressions with the addition of a few arithmetic operators. This is probably a reasonable serialization format, but might not be the ultimate user interface for specifying constraints... It also might want a better name :)

Related

Relates to organization/repo#number

Closes organization/repo#number

[ ] I have reviewed the Guidelines for Contributing and the Code of Conduct.
opened by richard-rogers 0
add a print_debug method to help with reported issues
Description

We should add a print_debug() method that outputs the whylogs version, OS, python version, and which extras are installed to help with debugging issues (especially for identifying already fixed issues being reported in older builds).

[x] I have reviewed the Guidelines for Contributing and the Code of Conduct.

feature
opened by jamie256 0
Need example notebook of single value metrics
Description

We have complex custom value examples, but it would help to have a simple single value metric that show WhyLabs integration.

[x] I have reviewed the Guidelines for Contributing and the Code of Conduct.

documentation
opened by jamie256 0

Releases(v1.1.20)

v1.1.20(Jan 3, 2023)
whylogs release v1.1.20

Hi everyone! We’ve now released whylogs 1.1.20 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Data validation docs #1040 [@FelipeAdachi]

Bump version to release version #1038 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.19(Dec 27, 2022)
whylogs release v1.1.19

Hi everyone! We’ve now released whylogs 1.1.19 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Support computing segmented performance metrics in top level API #1037 [@jamie256]

Update performance metrics example to show log_full_data behavior #1028 [@FelipeAdachi]

Source code(tar.gz)
Source code(zip)
v1.1.18(Dec 20, 2022)
whylogs release v1.1.18

Hi everyone! We’ve now released whylogs 1.1.18 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

hellinger calculation edge cases #1023 [@FelipeAdachi]

add_resolver to DeclarativeResolver and DeclarativeSchema #1016 [@FelipeAdachi]

updating blog link in KS benchmarks/KS_Profiling.ipynb and adding it … #1025 [@FelipeAdachi]

Fix read_v0_to_view cardinality metric #1020 [@jamie256]

Fix conversion of counters between v0 and v1 format #1026 [@jamie256]

Adding warning for vector dictionary on why.log #1022 [@murilommen]

Source code(tar.gz)
Source code(zip)
v1.1.17(Dec 13, 2022)
whylogs release v1.1.17

Hi everyone! We’ve now released whylogs 1.1.17 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

performance improvement for low latency row logging #1014 [@jamie256]

Fix reversed frequent item upper/lower bounds #1013 [@richard-rogers]

[fix] make dataset_id param default on ProfileStore #1012 [@murilommen]

Fluent condition predicate API #967 [@richard-rogers]

Declarative dataset schema #997 [@richard-rogers]

Source code(tar.gz)
Source code(zip)
v1.1.16(Dec 6, 2022)
whylogs release v1.1.16

Hi everyone! We’ve now released whylogs 1.1.16 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Register non-standard metrics for deserialization #1010 [@richard-rogers]

Add whylogs benchmark from docs to README.md #1006 [@jamie256]

Add DatasetSchema to Fugue API, add pytest-spark #1008 [@goodwanghan]

[feature] Implementing an SQLite Store #958 [@murilommen]

Bump version to release version #1001 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.15(Nov 29, 2022)
whylogs release v1.1.15

Hi everyone! We’ve now released whylogs 1.1.15 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Fix deserialize timezone info and merge of _metadata #1000 [@jamie256]

fixing broken links in examples/README.md #999 [@FelipeAdachi]

Add actions to ConditionCountMetric #989 [@richard-rogers]

update ks experiments #995 [@FelipeAdachi]

add condition_meets constraints factory helper function for condition count metrics. #991 [@FelipeAdachi]

viz - make null or undefined statistics appear as "-" instead of 0 #986 [@FelipeAdachi]

replace UCI repository dataset link to local S3 one #993 [@FelipeAdachi]

Fix Open in Colab link on Constraints_Suite #992 [@jamie256]

add tooltip text to "Observations" in summary drift report and profile summary #988 [@FelipeAdachi]

Source code(tar.gz)
Source code(zip)
v1.1.14(Nov 22, 2022)
whylogs release v1.1.14

Hi everyone! We’ve now released whylogs 1.1.14 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Fix formatted output for warning message when referencing utc_now #984 [@jamie256]

default performance metrics log full data false #980 [@jamie256]

add feature weights example to repository #982 [@FelipeAdachi]

Add error/warning logging for old or future dataset_timestamps #977 [@jamie256]

Add summary information to Viz' Constraints Report #976 [@FelipeAdachi]

Add setter for dataset_timestamp and model_performance_metrics #978 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.13(Nov 15, 2022)
whylogs release v1.1.13

Hi everyone! We’ve now released whylogs 1.1.13 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

pyspark datasetschema #972 [@gramhagen]

local segmented profiles serialization fix #971 [@jamie256]

Delete datasets symbolic link #970 [@FelipeAdachi]

Add timezone info to default dataset_timestamp #969 [@jamie256]

Standardize using single quotes around pip install in examples #960 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.12(Nov 9, 2022)
whylogs release v1.1.12

Hi everyone! We’ve now released whylogs 1.1.12 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Fix whylabs upload bug #963 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.11(Nov 8, 2022)
whylogs release v1.1.11

Hi everyone! We’ve now released whylogs 1.1.11 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Allow passing schema to log method #957 [@jamie256]

add benchmark for KS test. #951 [@FelipeAdachi]

Add separate NaN and inf counts, added tests #955 [@jamie256]

Don't reopen tempfile when writing profiles #954 [@jamie256]

enable quantiles as a parameter for KS drift calculation #945 [@FelipeAdachi]

[fix] Docs: Replacing advanced with basic examples #944 [@murilommen]

Source code(tar.gz)
Source code(zip)
v1.1.10(Nov 1, 2022)
whylogs release v1.1.10

Hi everyone! We’ve now released whylogs 1.1.10 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Java Loggers #939 [@TheMellyBee]

✨ Features

implements hellinger distance as a drift calculation #940 [@FelipeAdachi]

Add zero() to result_set and DatasetProfileView and test merge #942 [@jamie256]

🐛 Bug Fixes

Avoid KeyError indexing into pandas DataFrame #928 [@jamie256]

[Fix] Fixing broken links on readme #941 [@murilommen]

📚 Documentation

Adjusting overall README examples and integrations #934 [@murilommen]

Source code(tar.gz)
Source code(zip)
v1.1.9(Oct 26, 2022)
whylogs release v1.1.9

Hi everyone! We’ve now released whylogs 1.1.9 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Add the Result Set to Java #908 [@TheMellyBee]

Add segmented profile serialization to rolling logger #935 [@jamie256]

remove TransientLogger::flush() and ::close() #931 [@richard-rogers]

Fix stubs to handle no pandas scenario #930 [@jamie256]

Update to latest stable release of whylabs-client and poetry update #924 [@jamie256]

Changing quantiles for KS drift detection from 9 to 100 buckets. #932 [@FelipeAdachi]

🐛 Bug Fixes

Renames add_metrics to add_metric as it only adds a singular metric #917 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)
v1.1.8(Oct 18, 2022)
whylogs release v1.1.8

Hi everyone! We’ve now released whylogs 1.1.8 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Adding GCS writer #913 [@murilommen]

Handle numpy.bool_ when resolving metrics for bool types #914 [@jamie256]

Allow test to run on automated version bump PR #906 [@jamie256]

Implements Profiles & Views in Java #867 [@TheMellyBee]

Bump version to release version #905 [@github-actions]

Source code(tar.gz)
Source code(zip)
v1.1.7(Oct 11, 2022)
whylogs release v1.1.7

Hi everyone! We’ve now released whylogs 1.1.7 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Count booleans in cardinality metric, and add tests #904 [@jamie256]

Make SummaryConfig optional #903 [@richard-rogers]

Wrong image for visualizer #899 [@dleybz]

Generalize CompoundMetric for better whylabs support #894 [@richard-rogers]

Adding Profile Store functionality #852 [@murilommen]

More links to WhyLabs #898 [@dleybz]

Bump version to release version #897 [@github-actions]

Source code(tar.gz)
Source code(zip)
v1.1.6(Oct 4, 2022)
whylogs release v1.1.6

Hi everyone! We’ve now released whylogs 1.1.6 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Enable Fugue Implementation to take arbitrary column names #896 [@goodwanghan]

Update image logging example notebook #886 [@richard-rogers]

Bump version to release version #895 [@github-actions]

Source code(tar.gz)
Source code(zip)
v1.1.5(Sep 30, 2022)
whylogs release v1.1.5

Hi everyone! We’ve now released whylogs 1.1.5 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Fixed standard aggregators/serializers/deserializers #893 [@andyndang]

Fix stubs to allow isinstance checks #891 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.4(Sep 27, 2022)
whylogs release v1.1.4

Hi everyone! We’ve now released whylogs 1.1.4 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Comment out .doc lfs pattern in .gitattributes #884 [@jamie256]

add dependencies to viz extra #883 [@FelipeAdachi]

Update version to 1.1.3 #882 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.1.3(Sep 27, 2022)
whylogs release v1.1.3

Hi everyone! We’ve now released whylogs 1.1.3 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

removing image compounded metric in profile visualization #880 [@FelipeAdachi]

Fix ImageMetric::merge() bug #878 [@richard-rogers]

Notebook visualizer support for ImageMetric #877 [@richard-rogers]

Update ecommerce.rst #871 [@FelipeAdachi]

Better whylabs support for logged images #870 [@richard-rogers]

Fix empty view to_pandas #868 [@jamie256]

MetricConfig was supposed to be an optional argument to Metric::zero() #863 [@richard-rogers]

Source code(tar.gz)
Source code(zip)
v1.1.2(Sep 22, 2022)
whylogs release v1.1.2

Hi everyone! We’ve now released whylogs 1.1.2 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Add 'requests' package as dependency of whylogs[whylabs] #855 [@TheMellyBee]

Update required protobuf version to 3.19.4 or higher #862 [@jamie256]

Add open in colab button and fix links #861 [@jamie256]

Speed up confusion matrix calculation and add test #859 [@jamie256]

Example updates - adding WhyLabs blurb to every examples #856 [@FelipeAdachi]

Move .dockerignore to root #853 [@jamie256]

✨ Features

Update whylabs-client version for feature importance #854 [@jamie256]

feature importance writers API #857 [@FelipeAdachi]

Source code(tar.gz)
Source code(zip)
v1.1.1(Sep 20, 2022)
whylogs release v1.1.1

Hi everyone! We’ve now released whylogs 1.1.1 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Update fugue integration #850 [@goodwanghan]

Fugue Integration #839 [@goodwanghan]

poetry update #847 [@jamie256]

Update example links to point to nbviewer #834 [@jghoman]

Uncomment pip install so notebook examples mostly just work #837 [@jamie256]

Update getting started with WhyLabs notebook to v1 #832 [@jamie256]

'Check out our docs' link in Inspecting Profiles notebook is 404ing #829 [@jghoman]

Add support for serder into bytes and pickle support for DatasetProfileView #827 [@andyndang]

StandardResolver and Supporting classes #779 [@TheMellyBee]

Added image logging example notebook #815 [@richard-rogers]

Initial Sagemaker example readme and link to v0 #826 [@jamie256]

Update README.md #824 [@FelipeAdachi]

Update weather dataset test dates to UTC time #777 [@bernease]

regression metrics example #818 [@FelipeAdachi]

classification metrics example #819 [@FelipeAdachi]

Writer Examples - fixing environment variables and deprecated parameters #810 [@FelipeAdachi]

🐛 Bug Fixes

Fix bug in compound metric to_summary_dict #841 [@richard-rogers]

Fix multi-column segment key translation #848 [@jamie256]

typing-extensions dependency entry update #845 [@jamie256]

Have Writer's Return a Response #823 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)
v1.1.0(Sep 14, 2022)
whylogs release v1.1.0

Hi everyone! We’ve now released whylogs 1.1.0 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Initialize RegressionMetricsMessage prediction_field to 0 #814 [@jamie256]

Handle missing PIL #812 [@richard-rogers]

Added top level performance metrics methods #811 [@jamie256]

segments example #806 [@FelipeAdachi]

Model Performance Metrics in v1.1 #796 [@jamie256]

Datasets Module - Add Ecommerce Dataset #797 [@FelipeAdachi]

Tweak the information collection sentence so it's clear immediately this can be turned off. #801 [@jghoman]

Fix whylabs writer uncompound on segmented profiles #803 [@jamie256]

Fixed handling int with frequent item bug #794 [@andyndang]

Image metric logging v1 #776 [@richard-rogers]

Initial segments support for v1.1 #792 [@jamie256]

Hardcode en/latest into Announcement URL. Otherwise is coming up 404. #795 [@jghoman]

Add .dockerignore to .gitignore #790 [@jamie256]

Source code(tar.gz)
Source code(zip)
v1.0.14(Sep 6, 2022)
whylogs release v1.0.14

Hi everyone! We’ve now released whylogs 1.0.14 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Update version of create-pull-request #789 [@jamie256]

Update whylogs-sketching version and update poetry.lock #788 [@jamie256]

Condition Validators #761 [@FelipeAdachi]

Add Serialization and Deserialization Registry #762 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)
v1.0.13(Aug 31, 2022)
whylogs release v1.0.13

Hi everyone! We’ve now released whylogs 1.0.13 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

move link check into separate action #783 [@jamie256]

including Dask example #780 [@murilommen]

Aggregator for Metric Components #757 [@TheMellyBee]

bumpversion and allow dirty fix #782 [@FelipeAdachi]

changing WhyLabs Writer example's section title #778 [@FelipeAdachi]

Enable overriding SSL cert store #773 [@andyndang]

Update Documentation - Getting Started #768 [@FelipeAdachi]

DatasetSchema no longer a @dataclass #750 [@richard-rogers]

Source code(tar.gz)
Source code(zip)
v1.0.12(Aug 23, 2022)
whylogs release v1.0.12

Hi everyone! We’ve now released whylogs 1.0.12 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Remove the java release's python integration now that we have pyspark #760 [@jamie256]

WhyLabsWriter's log_reference_profile #753 [@FelipeAdachi]

Adding Java to the CI #758 [@TheMellyBee]

Using our dedicated stats endpoint for telemetry collection #759 [@andyndang]

🐛 Bug Fixes

Add Twine to dev dependency #766 [@TheMellyBee]

Broken Link Detection Action #749 [@TheMellyBee]

📚 Documentation

Broken Link Detection Action #749 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)
v1.0.11(Aug 16, 2022)
whylogs release v1.0.11

Hi everyone! We’ve now released whylogs 1.0.11 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Java IntegralMetric and Components #719 [@TheMellyBee]

Update push-release to not block on autobump version #752 [@jamie256]

Update the telemetry account ID #751 [@andyndang]

Condition Count Metric Example #748 [@FelipeAdachi]

Datasets Module #724 [@FelipeAdachi]

Enabling whylogs Reader #745 [@murilommen]

Fix Schema Configuration example #744 [@FelipeAdachi]

Viz module UI/UX fixes #742 [@FelipeAdachi]

Fix Constraints Report's search feature #737 [@FelipeAdachi]

Update README.md #739 [@jamie256]

adding BigQuery example #734 [@murilommen]

📚 Documentation

Updates the ReadMe for Java #741 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)
v1.0.10(Aug 9, 2022)
whylogs release v1.0.10

Hi everyone! We’ve now released whylogs 1.0.10 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

just adding Custom Metrics in the documentation #735 [@FelipeAdachi]

Add some missing properties on metrics to cover things in summaries #731 [@jamie256]

make plots not sticky when scrolling horizontally #730 [@FelipeAdachi]

Condition count metric / value constraints #618 [@richard-rogers]

✨ Features

Add callback for rolling logger profile writes #725 [@jamie256]

Source code(tar.gz)
Source code(zip)
whylogs-1.0.10-py3-none-any.whl(1.66 MB)
whylogs-1.0.10.tar.gz(1.63 MB)
v1.0.9(Aug 2, 2022)
whylogs release v1.0.9

Hi everyone! We’ve now released whylogs 1.0.9 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

add bump2version to dev dependencies and sync poetry.lock file #721 [@jamie256]

Update .bumpversion.cfg #720 [@TheMellyBee]

Fix DatasetProfile not resolving with a custom schema #718 [@richard-rogers]

Making constraint helpers publicly available #715 [@murilommen]

Examples of custom Metrics #586 [@richard-rogers]

Source code(tar.gz)
Source code(zip)
v1.0.8(Jul 26, 2022)
whylogs release v1.0.8

Hi everyone! We’ve now released whylogs 1.0.8 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Add testing python for 3.10 and restrict compatibility to less than 3.11 #710 [@jamie256]

Copy minimal structure for whylogs java #680 [@andyndang]

Fix using None as type when pandas is not present #706 [@jamie256]

Allow python 3.10 in pyproject.toml #704 [@jamie256]

Stabilizing IPython init #707 [@murilommen]

Add python working directory for python commands in release workflow #701 [@jamie256]

✨ Features

single-profile visualization (Profile Summary) #705 [@FelipeAdachi]

Orders all views columns before sending to pandas #702 [@TheMellyBee]

📚 Documentation

Fix small typo. #711 [@jghoman]

Flask Integration of whylogs and WhyLabs Notebook #685 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)
whylogs-1.0.8-py3-none-any.whl(1.65 MB)
whylogs-1.0.8.tar.gz(1.62 MB)
v1.0.7(Jul 19, 2022)
whylogs release v1.0.7

Hi everyone! We’ve now released whylogs 1.0.7 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Decoupling Notebook Profile Vizualizer #661 [@murilommen]

add sort by drift p-values #699 [@FelipeAdachi]

Update dependency on whylogs-sketching that has kll merge fix #693 [@jamie256]

Change KLL accuracy according to numerical/submetrics type #690 [@FelipeAdachi]

Comment pip install in examples notebooks #687 [@FelipeAdachi]

Adding Push-Release #689 [@TheMellyBee]

Adds the BumpVersion for our CICD workflow #691 [@TheMellyBee]

📚 Documentation

Updates links for readme, merges it in, and removes an unavailable co… #698 [@TheMellyBee]

fix contribution.md hyperlink on readme #696 [@murilommen]

Source code(tar.gz)
Source code(zip)
whylogs-1.0.7-py3-none-any.whl(1.64 MB)
whylogs-1.0.7.tar.gz(1.61 MB)
v1.0.6(Jul 14, 2022)
whylogs release v1.0.6

Hi everyone! We’ve now released whylogs 1.0.6 🚀. whylogs is the open standard for data and ML logging created by WhyLabs. 👩🏽‍🔬 This version includes:

Fix README Docs hyperlink #683 [@murilommen]

Add WhyLabs Writer example notebook #672 [@FelipeAdachi]

Optionally skip lower-casing the strings #678 [@richard-rogers]

Make Profile/Profile Views retain the same order of columns as original DataFrame #675 [@FelipeAdachi]

Constraints - helper functions #652 [@FelipeAdachi]

make Writing_Profiles & Mlflow_Logging testable #670 [@FelipeAdachi]

[Fix] Making installs on examples "shell safe" #608 [@murilommen]

Adds in Release Drafter to v1 #688 [@TheMellyBee]

Source code(tar.gz)
Source code(zip)

Owner

WhyLabs

Observability for AI pipelines and applications. Instrument data pipelines, analyze data quality and drift, catch deviations before they cause model failures.

GitHub Repository http://docs.whylabs.ai/

jaxfg - Factor graph-based nonlinear optimization library for JAX.

Factor graphs + nonlinear optimization in JAX

134 Dec 21, 2022

An AutoML survey focusing on practical systems.

This project is a community effort in constructing and maintaining an up-to-date beginner-friendly introduction to AutoML, focusing on practical systems. AutoML is a big field, and continues to grow

16 Aug 14, 2022

🌲 Implementation of the Robust Random Cut Forest algorithm for anomaly detection on streams

416 Jan 06, 2023

Machine Learning University: Accelerated Natural Language Processing Class

Machine Learning University: Accelerated Natural Language Processing Class This repository contains slides, notebooks and datasets for the Machine Lea

2k Jan 01, 2023

Python ML pipeline that showcases mltrace functionality.

mltrace tutorial Date: October 2021 This tutorial builds a training and testing pipeline for a toy ML prediction problem: to predict whether a passeng

28 Nov 09, 2022

Dive into Machine Learning

Dive into Machine Learning Hi there! You might find this guide helpful if: You know Python or you're learning it 🐍 You're new to Machine Learning You

11.1k Jan 03, 2023

Automated Machine Learning with scikit-learn

auto-sklearn auto-sklearn is an automated machine learning toolkit and a drop-in replacement for a scikit-learn estimator. Find the documentation here

6.7k Jan 07, 2023

Petastorm library enables single machine or distributed training and evaluation of deep learning models from datasets in Apache Parquet format. It supports ML frameworks such as Tensorflow, Pytorch, and PySpark and can be used from pure Python code.

Petastorm Contents Petastorm Installation Generating a dataset Plain Python API Tensorflow API Pytorch API Spark Dataset Converter API Analyzing petas

1.6k Dec 31, 2022

Neighbourhood Retrieval (Nearest Neighbours) with Distance Correlation.

Neighbourhood Retrieval with Distance Correlation Assign Pseudo class labels to datapoints in the latent space. NNDC is a slim wrapper around FAISS. N

1 Jan 16, 2022

A statistical library designed to fill the void in Python's time series analysis capabilities, including the equivalent of R's auto.arima function.

pmdarima Pmdarima (originally pyramid-arima, for the anagram of 'py' + 'arima') is a statistical library designed to fill the void in Python's time se

1.3k Dec 22, 2022

Tribuo - A Java machine learning library

Tribuo - A Java prediction library (v4.1) Tribuo is a machine learning library in Java that provides multi-class classification, regression, clusterin

1.1k Dec 28, 2022

Tools for mathematical optimization region

15 Nov 30, 2022

MLR - Machine Learning Research

Machine Learning Research 1. Project Topic 1.1. Exsiting research Benmark: https://paperswithcode.com/sota ACL anthology for NLP papers: http://www.ac

69 Oct 20, 2022

CinnaMon is a Python library which offers a number of tools to detect, explain, and correct data drift in a machine learning system

67 Dec 28, 2022

whylogs: A Data and Machine Learning Logging Standard

Related tags

Overview

whylogs: A Data and Machine Learning Logging Standard

Getting started

Using pip

From source

Quickly Logging Data

Multiple Profile Plots

Profile Viewer

Documentation

Features

Data Types

Integrations

Examples

Roadmap

Community

Contribute

Comments

Problem

Fixes

Description

Description

Related

Description

General Checklist

Description

Description

General Checklist

General Checklist

Description

Description

Description

Description

Changes

Related

Description

Description

Releases(v1.1.20)

v1.1.20(Jan 3, 2023)

whylogs release v1.1.20

v1.1.19(Dec 27, 2022)

whylogs release v1.1.19

v1.1.18(Dec 20, 2022)

whylogs release v1.1.18

v1.1.17(Dec 13, 2022)

whylogs release v1.1.17

v1.1.16(Dec 6, 2022)

whylogs release v1.1.16

v1.1.15(Nov 29, 2022)

whylogs release v1.1.15

v1.1.14(Nov 22, 2022)

whylogs release v1.1.14

v1.1.13(Nov 15, 2022)

whylogs release v1.1.13

v1.1.12(Nov 9, 2022)

whylogs release v1.1.12

v1.1.11(Nov 8, 2022)

whylogs release v1.1.11

v1.1.10(Nov 1, 2022)

whylogs release v1.1.10

✨ Features

🐛 Bug Fixes

📚 Documentation

v1.1.9(Oct 26, 2022)

whylogs release v1.1.9

🐛 Bug Fixes

v1.1.8(Oct 18, 2022)

whylogs release v1.1.8

v1.1.7(Oct 11, 2022)

whylogs release v1.1.7

v1.1.6(Oct 4, 2022)

whylogs release v1.1.6

v1.1.5(Sep 30, 2022)

whylogs release v1.1.5

v1.1.4(Sep 27, 2022)

whylogs release v1.1.4

v1.1.3(Sep 27, 2022)

whylogs release v1.1.3

v1.1.2(Sep 22, 2022)