CHANGES
=======

1.1.0
-----

* Remove vendorized aiohttp-apispec
* fix unit tests
* fix pep8
* add training\_args, training\_output and training\_duration to the train\_response
* remove misleading headers from the UI

1.0.1
-----

* Fix failining unit test on cover env
* Set basePath on swagger.json when serving under OpenWhisk
* Return doc and swagger links on /
* Temporary vendorize aiohttp-apispec
* Changed typo in CONTRIBUTING.md
* Update documentation
* Fix cancellation of trainings

1.0.0
-----


1.0.0.rc4
---------

* Limit installation to Py3 versions
* Use release/1.4.0 of jenkins library to handle Python3 PyPI delivery
* Enable OWASP check
* Require pbr>=5.3.0
* Remove accented vowels from README

1.0.0.rc1
---------

* Use release as version in docs (#69)
* Move Swagger web UI to /ui rather than /docs
* fix default content type management
* Homogenize route declaration indentation
* Remove HEAD methods from GET routes
* debug: do not print debug marker is debug endpoint is not enabled
* Support for different response content types
* wsk: return 200 whenever we have a response from the app
* wsk: handle empty bodies in responses
* Add swagger.json on OpenWhisk functions
* Fix pep8 errors on Py3.6
* Jenkins: bump Python version to 3.6
* Disable unit test due to spurious failure
* testing: use shared fake responses
* wsk: add proxy testing
* Fix V2 models unittest
* Fix pep8 errors
* wsk: add wsk\_args to all handlers
* fix catch\_error docs
* wsk: disable trainings on OpenWhisk
* Move options to correct place
* wsk: fix OpenWhisk and aiohttp integration
* Use trailing slashes when defining endpoints
* wsk: return an aiohttp.web.Response on OpenWhisk proxy
* Remove empty return
* wsk: await the get\_app async function
* update error docs
* order user args
* add warm option to cli
* add client-max-size to cli
* fix client filesize limit
* typo fix
* update to new UploadedFile method
* Fix error on PyPI metadata generation
* Fix pep8 errors
* Fix V1 file passing
* Include citation information
* Update documentation logo and style
* Add correct support email address
* add python version requirement
* Include rich API info metadata
* Fix typo ¡n class name and unittests
* Fix pep8 errors and add correct doc links
* Add 204 response if debug endpoint is not enabled
* Return /debug responses as text/plain
* Fix documentation
* Allow loading arguments using the old way
* Fix "AssertionError: group argument must be None for now" errors
* Add banner to startup
* Improve documentation
* Document passing file arguments
* Remove trainings at all fro API
* Simply setup\_routes functions
* Run training and prediction as tasks
* Use asyncio in ModelWrapper and adapt API
* Move model wrapper into its own module
* Remove \_call\_method helper method and use a decorator
* test model: add example of argument with valid choices
* Drop Flask in favour of aiohttp
* Drop Python 2 support
* doc: update CLI man pages
* Add debug endpoint
* Disable reloading when running as debug
* V2: Initial version of the new API
* Include upgrade notes in documentation
* Do the plumbing for loading different V2 and V1 models
* Disable loading of V1 version and add CLI option
* Add V2 skeleton, following V1 API but on separate modules
* Add a version endpoint at the root URL
* Refactor API namespaces loading, using Flask blueprints
* Use release\_string as version, as it includes VCS tags
* tox: Include testing for Py3.7
* Change documentation theme
* Update paper.md
* Fix DOI in JOSS paper
* Fix bibiography entry in JOSS paper
* fix parsers overwrite (#62)
* Correct bibliography with correct entries and add missing DOI items
* Update README
* Adjust tox.ini settings and environments
* Install netifaces from APT repo

0.5.2
-----

* Fix documentation build error

0.5.1
-----

* Fix test args parser
* Fix test arguments
* docs fix
* fix style
* update docs and docstrs

0.5.0
-----

* Prepare 0.5.0 release
* Bump gevent dependency version
* add get\_train/test\_args docstrings
* add test\_args functionality

0.4.1
-----

* Update requirements.txt for Werkzeug>=0.15.0
* Update version on Dockerfile

0.4.0
-----

* Update Dockerfile for Docker container creation
* Add unit tests for cmd module
* Rename "openwisk" to "openwhisk" module
* tox: remove Python 3.5 testing
* Handle SIGTERM signal explicitly
* Remove leftover debug print statement
* Migrate to jenkins library 1.3.5 (fix in DockerClean)
* Set new Jenkins badge URL
* Set PyPI delivery only on new release
* Move to version 1.3.4
* Move to 1.3.3
* fixing typos
* Add initial version of paper for Journal of Open Source Software
* Comment out OWASP stage
* Move to pipeline==1.3.1
* Align with 1.0.0 version of pipeline library

0.3.0
-----

* Set absolute path to HTMLreport
* Disable Swagger documentation for OpenWhisk
* Add support to run DEEPaaS as an OpenWhisk Docker action
* Point to correct documentation
* Fix typo on documentation

0.2.0
-----

* Add missing release notes
* get\_train\_args() docs
* Initial support for passing down training arguments
* Add WP2 main contacts as watchers of JIRA issues
* Use indigo-dc.deepaas Ansible role
* Add badge for Jenkins build status

0.1.2
-----

* setup.cfg: fix classifiers
* docs: fix document generation on Read the Docs

0.1.1
-----

* tox: create environment for checking dependencies
* Include release notes in documentation
* Add initial documentation using Sphinx
* Generate configuration file through tox
* Improve versioning information within package
* Update README with documentation on DEEPaaS usage
* Update setup.cfg file
* Add PYPI delivery
* Add link to dockerhub image
* Fix: dockerhub\_image\_id variable interpolation

0.1.0
-----

* Fix: image\_id to dockerhub\_image\_id
* Add correct Dockerfile version
* Add OWASP dependency check
* Move Jenkinsfile's tox environments to tox.ini
* Create JIRA issue on new builds
* Provider requirements as a list
* Add a Makefile to quickly run DEEPaaS
* Add unit testing for new model wrapper and base classes
* Define a base class for all models
* Migrate from testr to stest and adjust Python supported versions
* Start using logging
* tests: use oslo\_config.fixture in the unit tests
* Computes SLOC data (#9)
* Switch to shared libraries (#8)
* Indent fix
* Move Dockerfile to repository's root path
* Fix double underscore typo
* Add credentials for pushing the image to DockerHub
* Add docker image build step
* Add Dockerfile to codebase
* Fix Jenkinsfile indentation
* Generate Cobertura report
* Get and publish code coverage
* Parse flake8 logs
* Test flake8 from different stage
* Separate stage for checkout code
* jenkins: added code style stage
* Allow uploading files as well as pointing to URLs
* Rework model loading
* Allow loading different models
* Allow configuring IP and port where we should listen
* tests: add cmd testing
* testing: fix package name
* Add some boilerplate README text
* Add initial unit testing
* Add initial unit testing plumbing
* Ignore editor files
* Separate model loading into its own module
* Add werkzeug as an explicit dependency
* Fix pep8 errors
* Naive model loading
* Fix data upload and argument parsing
* Update response model
* Add initial loading plumbing
* Return model metatada
* Allow uploading files and URLs in the POST request
* Split API and namespaces into separate files
* add python shebang
* Add initial code skeleton
* add tox with initial pep8
* Add community and informative files
* Initial commit
