Skip to content

Conversation

@mfeurer
Copy link
Collaborator

@mfeurer mfeurer commented Apr 23, 2018

I think we're good to go for our first release on pypi as most bugs are resolved for now. I'd like to do this as a soft launch and not advertise it until with the next minor release the pypi site and documentation look perfect. I'll create a new milestone afterwards.

janvanrijn and others added 30 commits September 10, 2017 10:02
add task ids to docs of list_tasks
# Conflicts:
#	openml/datasets/functions.py
fix docstrings in list_datasets, some pep8
Cast data qualities to float
Make task the default concept to work with
janvanrijn and others added 26 commits February 20, 2018 17:57
* Fix ascii decoding problem with python 2

* Created separate pickle files for splits and datasets according to the python version

* Added production task to unit test

* Update test_task_functions.py
* Implementing dataset listing with more filters and adding unit tests

* Add to function documentation
* Created first basic template, removed redudant variable

* Improving list_datasets

* First implementation of the list_* with the limit tag active

* First implementation of the feature, fixed bugs and refactored the code

* Changing batch_size to be a keyword argument

* Fixed not considering initial offset, removing size and the double offset key from the filter dict

* Changing task_type_id argument name in accordance with the new implementation

* Reverting previous solution for task_type_id, implementing another fix

* Fix for python2 and changing the unit test which times out

* Added another test method and did a slight change in an existing test method

* Changing the assert value for the failing test method

* Added the implementation to filter by uploader for flows, filter by task_type for runs, filter by multipple operator for tasks and also refactored the code according to PEP8

* Refactored code as requested
* ADD unit test to ensure example listing

* Update test_study_examples.py
* MAINT bump required liac-arff version

* Update requirements.txt
* ADD contributing guidelines and github templates

* Update CONTRIBUTING.md

* ADD reference to scikit-learn to the license

* Added clarification for unit test on Windows.
* ADD hash checks to datasets

* split target on comma, allows errors on multiple targets

* FIX unit test
* MAINT move from cell executor to exec to test jupyter notebooks

* Mock calls to /run/upload/ in notebook unit tests

* FIX mock patch target

* Change backend
* Basic solution for private datasets

* Faulty and not necessary implementation of finally

* Fixed assertRaises call in test_get_data

* Updated get_dataset

* Fixed typo
* Have different directories for different servers

* simplify based on jans comments

* fix attribute access

* Change variable name

* Take into account Jans suggestions

* Harmonize import, fix rebase errors

* re-add accidentaly removed files

* First try at a solution

* Removing faulty fix

* Fix for bug in unit test, method _get_cached_task

* Removing FileNotFoundError as it does not exist in python2

* Fixing test_tagging

* Changing id according to new solution

* Change _remove_dataset_cache_dir to the new implementation
* Adding support for home directory symbol

* Fixing backward compatibility
* ADD introductory snippet to front page of the docs

* FIX issues with rst formatting
* FIX fixes and improves unit tests

* ADD additional asserts to hunt down bug

* ADD debug output

* remove print statement
* FIX circle ci yaml file (remove requirements.txt)

* circle ci: build docs on all branches
Copy link
Member

@janvanrijn janvanrijn left a comment

Choose a reason for hiding this comment

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

That's a lot of code for half a year

@mfeurer mfeurer merged commit 3449663 into master Apr 24, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

8 participants