CHANGELOG-v2-and-older.rst

Change Log

This changelog file holds the changed smaller than version 3.0.0.

2.7.0 (31OKT19)

Warning

This is the last release that is compatible with Python 2.7, which is due for sunsetting in Januari 2020.

This is the last release that is compatible with the KE-chain 2 API (KE-chain API versions < 3.0).

Note

For releases of KE-chain >= v3.0, you need a pykechain >= 3.0.

• Added a function to retrieve the associated activities of a part: Part.associated_activities() and Property.associated_activities(). (#503 - Thanks to @raduiordache for the PR)
• Added a function to count parts Part.count_instances() using a lightweight call to the API. (#485 - Thanks to @raduiordache for the PR)
• Updated dependent versions for development: pytest (5.2.2),tox (3.14.0), twine (2.0.0), matplotlib (3.1.1), Sphinx (2.2.1), semver (2.9.0), flake8 (3.7.9), mypy (0.740), jsonschema (3.1.1), nbsphinx (0.4.3), pydocstyle (4.0.1)
• Added a source distribution to PyPI.

2.6.1 (17JUN19)

• Fixed a bug where in the move/copy functionality the options to ReferenceProperty and AttachmentProperty where not passed down. Thanks to @raduiordache. (#502)
• Updated dependent versions for development: requests (2.22.0), pytest (4.6.3),tox (3.12.1), twine (1.13.0), matplotlib (3.1.0), Sphinx (2.1.1).

2.6.0 (23APR19)

• Added the possibility to create a scope, clone a scope, and delete a scope. Check Client.create_scope(), Scope.clone and Scope.delete for documentation. (#359)

2.5.7 (18APR19)

• Added additional properties for the Service and ServiceExecution class. Now you can retrieve the Service.filename amoungst others. Please refer to the documentation of Service and ServiceExecution to see the properties that are now available (a feature request by @JelleBoersma). #480
• We added a utility function to parse_datetime strings into datetime objects. These strings are in a json response from the KE-chain backend and are now properly translated and timezoned. #482
• Updated dependent versions for development: pytest (4.4.1), mypy (0.701), tox (3.9.0).

2.5.6 (13APR19)

• Small patch release to ensure that the Activity2.assignees returns an empty list when nobody is assigned to the task. #477. Thanks to @raduiordache for finding it out.

2.5.5 (11APR19)

• Added properties to the Property to directly access properties such as unit, description and type. Property.type refers to a PropertyType enum. #469
• Added a property to the AttachmentProperty.filename to return the filename of an attachment. #472
• Added a property to retrieve the assignees list of an activity through Activity2.assignees. This will return a list of User's assigned to the activity. #473
• Added additional properties to Service such as name, description and version of a service. #469
• Added additional properties to Scope such as description, status and type. #469
• Updated dependent versions for development: matplotlib (3.0.3), jsonschema (3.0.1), pytest (4.4.0), sphinx (2.0.1), mypy (0.700), tox (3.8.6).

2.5.4 (28FEB19)

• Fixed a bug where the update of the single select list options could overwrite the existing validators. Thanks to @jelleboersma for finding this out and creating the PR. (#446)
• Updated dependent versions for development: sphinx (1.8.4), mypy (0.670), pytest (4.3.0), flake8 (3.7.7), jsonschema (3.0.0), pyOpenSSL (for python 2.7, 19.0.0).
• Updated security advisory to install requests package later than 2.20.0 (CVE-2018-18074).

2.5.3 (21JAN19)

• Fixed a bug where a numeric range validator from a property was not correctly instantiated for provided min/max values when the validator was retrieved from the KE-chain backend. Thanks to @bastiaanbeijer for finding this! (#435)
• Updated dependent versions for development: requests (2.21.0), sphinx (1.8.3), pytest (4.1.1), mypy (0.660), nbsphinx (0.4.2), tox (3.7.0).

2.5.2 (30NOV18)

• Fixed the customizations to be compatible with KE-chain 3: Custom Title replaced by Custom title; added the possibility to include the Clone button where applicable. The metaWidget now uses 'Set height' and 'Automatic height'. (#421) thanks to @raduiordache.
• Updated dependent versions for development: requests (2.20.1), sphinx (1.8.2), pytest (4.0.1), requests (2.20.0), matplotlib (3.0.2)

2.5.1 (05NOV18)

• patch release to include the dependency pytz in the normal list of dependencies, not only for development.

2.5.0 (1NOV18)

• Added the ability to set and retrieve the scope tags using the Scope.tags property. (#367)
• Added timezone, language and email to the user object. You can access this directly as a property on the User object. (#378)
• Ensured that you can now filter users on their name, username and email. (#373)
• Added the possibility to generate a PDF from an activity even with attachments included. The later is an async process on the KE-chain server and pykechain uses a 'hint' to retrieve the PDF once it becomes available on the server. It has an timeout of 100 seconds. (#406)
• included many updated tests for the copy_move functionality including cross reference properties. (#376)
• Updated dependent versions for development: semver (2.8.1), pydocstyle (3.0.0), mypy (0.641), requests (2.20.0), flake8 (3.6.0), matplotlib (3.0.1), pytest (3.9.3), tox (3.5.3)

2.4.1 (26SEP18)

• Added support for the Scope.team property. Will return a Team object if the project has a team associated to it, otherwise None. (#392)
• Included Team object in the API documentation.

2.4.0 (26SEP18)

• Added the Team concept. You can now query the API to retrieve Teams using client.team(name='My own team'). You can also now Team.add_members and Team.remove_members with their TeamRoles. (#391)
• Updated dependent versions for development: twine (1.12.1)

2.3.3 (24SEP18)

• Fixed an issue with the scope.edit() method. It will handle now the assignment of the team with a team_id correctly. Thanks @stefan.vanderelst (#388)
• Updated dependent versions for development: tox (3.4.0), pytest (3.8.1), sphinx (1.8.1)

2.3.2 (19SEP18)

• The setting of the min and max value of the numeric range validator could not correctly deal with a value of None. That is fixed. Thanks to @JelleBoersma for the fix! (#382)
• Additional widgets are introduced in KE-chain or in the process of being introduced, so we updated the enumerations. In this process we also updated the jsonschema of the widget to check against before uploading a customization to KE-chain (#369)
• Updated dependent versions for development: tox (3.2.1), pytest (3.8.0), nbsphinx (0.3.5), sphinx (1.8.0), mypy (0.630) and matplotlib (3.0.0)

2.3.1 (2AUG18)

• The details of a scope can now be edited using Scope.edit() method. This contains action already prepared for the KE-chain 2.16.0-143 release (Mid August). (#357)

  For example:

  >>> from datetime import datetime
  >>> project.edit(name='New project name',description='Changing the description just because I can',start_date=datetime.utcnow(),status=ScopeStatus.CLOSED)

• Updated dependent versions for development: pytest (3.7.0)

  For example:

  >>> from datetime import datetime
  >>> project.edit(name='New project name',description='Changing the description just because I can',start_date=datetime.utcnow(),status=ScopeStatus.CLOSED)

• Updated dependent versions for development: pytest (3.7.0)

  For example:

  >>> from datetime import datetime
  >>> project.edit(name='New project name',description='Changing the description just because I can',start_date=datetime.utcnow(),status=ScopeStatus.CLOSED)

• Updated dependent versions for development: pytest (3.7.0)

  For example:

  >>> from datetime import datetime
  >>> project.edit(name='New project name',description='Changing the description just because I can',start_date=datetime.utcnow(),status=ScopeStatus.CLOSED)

• Updated dependent versions for development: pytest (3.7.0)

For example:

>>> from datetime import datetime
>>> project.edit(name='New project name',
...              description='Changing the description just because I can',
...              start_date=datetime.utcnow(),  # naive time is interpreted as UTC time
...              status=ScopeStatus.CLOSED)

• Updated dependent versions for development: pytest (3.7.0)

2.3.0 (26JUl18)

• We added additional utilities to help pykechain script developers to Part.copy(), Part.move() and Part.clone() part models and part instances. (#343)

For example; To move part models, their children (subtree) and their instances:

>>> model_to_move = project.model(name='Model to be moved')
>>> bike = project.model('Bike')
>>> model_moved = model_to_move.move(target_parent=bike, name='Moved model',
>>>                                  include_children=True,
>>>                                  include_instances=True)

• We added show headers and show columns in the arguments of the property grid to align to KE-chain functionality of the widget. (#350)
• We added the posibility to use a JSON widget to the list of allowed widgets. (#351)
• We added the posibility to update the options of a reference property. (#352)
• Updated dependent versions for development: pytest (3.6.3), tox (3.1.2), sphinx (1.7.6), mypy (0.620)

2.2.4 (22JUN18)

• An issue was fixed where the suppress_kevents flag was not correctly injected in the API request for all functions that created parts. (#340)
• Changed the way the cached children are stored when the Part.children() method is used. It is now cached as a List instead of a Partset and you can iterate over the List many times in your application. Thanks to Jelle Boersma for finding this. (#341)
• Updated dependent versions for development: pytest (3.6.2), mypy (0.610), requests (2.19.1)

2.2.3 (5JUN18)

• An issue was fixed in the Activity2.siblings() function. It now returns the actual siblings (other children of the common parent (subprocess)) for WIM2 based Activities. Thanks to @bastiaanbeijer for finding it, thanks to @raduiordache for fixing it. (#332)
• Updated dependent versions for development: pytest (3.6.1), semver (2.8.0), pyopenssl (18.0.0), sphinx (1.7.5)

2.2.2 (27MAY18)

• We fixed an issue with the pdf download option to ensure that the attachements property is passed as well in preparation for the async pdf downloader release in KE-chain 2.13.0-140 (#329). Found and fixed by @raduiordache; thanks!

2.2.1 (23MAY18)

• We ensured that old pykechain code to create a property model when not using the PropertyType enums is still compatible with changes introduced in version 1.16.0 (MAR18). We improved the documentation for Client.create_property() and enums.PropertyTypes. Using 'CHAR' (pre 1.16 style) instead of 'CHAR_VALUE' (enum style) will result in a warning (with suggestion to change this) for old code and will be corrected. Using an invalid property_type will result in an IllegalArgumentError. (#326)

2.2.0 (14MAY18)

Major feature: Property validators

• We added support for validators to KE-chain v2.12.0-139 and pykechain. Validators objects are stored on a property and can be used to validate the value of a property. The validator objects are also visualised in the KE-chain frontend. (#317)

Validators have a representation in the frontend of KE-chain 2 (see also documentation on: https://support.ke-chain.com/). The validators are stored on the Property object and currently the following validators are implemented:

• :class:`NumericRangeValidator`: When you provide a range, the validate can check if the value of the property is within range. It can even check a stepsize. See the documentation for :class:`NumericRangeValidators`. A representation in KE-chain is available when the value does not conform to this range.
• :class:`RequiredFieldValidator`: When you add this to a property (model), the property validates when a value is provided. There is a representation in KE-chain frontend available.
• :class:`RegexStringValidator`: A special validation to check a string (eg textfield) against a regex pattern. There no representation in KE-chain 2 in version v2.12.0-138.
• :class:`OddNumberValidator` and :class:`EvenNumberValidator`: a validator that checks a numeric field (decimal or integer field) if it is an even or odd number. There no representation in KE-chain 2 in version v2.12.0-138.
• :class:`SingleReferenceValidator`: a special validator that ensures that there can only be a single referenced part selected in a (multi) reference property.

To validate the property object there are several new functions available. :meth:`Property.validate()` to validate all validators attached to the property using the :attr:`Property.value` as basis for the validation. You will be provided back a resulting list with all validations including their validation reason.

To only check if the Property and its value conforms to the list of Validators, use the :attr:`Property.is_valid` and :attr:`Property.is_invalid` properties.

To retrieve the :class:`PropertyValidator` objects that are stored on the Property use the property :meth:`Property.validators`. You can set a list of :class:`PropertyValidator` objects to this property as well, which will be stored on the Property in KE-chain using an API call.

To add validators to a property (model):

>>> bike_model = project.model(name='Bike')  # type: Part
>>> electric_range = bike_model.property('electric_range')  # type: Property
>>> range = NumericRangeValidator(minvalue=0, maxvalue=100)  # instantiate a range validation between 0 and 100
>>> reqd = RequiredFieldValidator()  # instantiate a requiredFieldValidator
>>> electric_range.validators = [range, reqd]  # save the validators on the property to KE-chain

To validate a value against a validator:

>>> bike = project.part(name='Bike')  # type: Part
>>> electric_range = bike.property('electric_range')  # type: Property
>>> electric_range.value
None
>>> electric_range.is_valid  # No value set, invalid according to the requiredFieldValidator
False
>>> electric_range.value = 50
>>> electric_range.is_valid  # Value is provided AND value is within the range (0, 100)
True
>>> electric_range.value = -1
>>> electric_range.is_valid  # However, the value itself is invalid according to the range validation
False
>>> electric_range.validate(reason=True)  # use the explicit validation
[(False, "Value '-1' should be between 0 and 100"), (True, "Value is provided")]

For more documentation of Validators, please refer to the API documentation at: http://pykechain.readthedocs.io/en/latest/developer_api.html

Fixes and improvements

• A fix was made for the the Part.populate_descendants() to be working for part of category MODEL too. Thanks to a fix of @raduiordache. (#320)

2.1.1 (10APR18)

• We fixed an issue with the caching of the children of a Part when you retrieve children with additional filters on it. (#312)

2.1.0 (6APR18)

• We added an optimisation to the Part. When you use the Part.children() method, the children are cached for later re-retrieval. In order to boost performance even more, you can use the Part.populate_descendants() function to pre-populate all children for the whole subparttree inside the Part. You can easily then access its children without further expensive API calls. (#306)
• We brought the capabilities of the Text Widget up to specification with the KE-chain 2.10 release. We can now also provide and set the collapsed initial state of the text widget (#310). Thanks to @raduiordache.
• We added a function to download an activity as PDF (#286). Thanks to @raduiordache.
• Updated dependent versions for development: pytest (3.5.0), mypy (0.580), nbsphinx (0.3.2), tox (3.0.0), matplotlib (2.2.2), twine (1.11.0), shpinx (1.7.2)

2.0.0 (14MAR18)

This is a major release of pykechain, adding support for the legacy version of the Workflow Information Module (WIM) in KE-chain as well as the new version WIM2. Based on the version number of the WIM, either an Activity or an Activity2 class is provided.

Major differences

The main diferences in the concepts between WIM1 Activity and WIM2 Activity2 are:

• In WIM1: The root object is not an Activity, while in WIM2 the root object is an Activity2. Use predicates such as is_root to check this.
• In WIM1, an Activity that exist of the rootlevel, returns a NotFoundError when you search for its parent (using the subprocess() or parent method). In WIM2 you will get the root object back. Use the predicate is_rootlevel to help you assess if the Activity is indeed on the root level of the project.
• In WIM1, the types of activity are actually called a activity_class, while in WIM2 this is called an activity_type. A UserTask in WIM1 is a TASK in WIM2, and a Subprocess in WIM1 is a PROCESS in WIM2. The enums.ActivityTypes are updated accordingly.
• In WIM1, the assignees where to be assigned using usernames, in WIM2 user_ids need to be provided. For the sake of compatibility pykechain helps you in this. You can provide usernames to a Activity2.edit() and it is automatically translated in user_ids with additional calls to KE-chain.

Other changes

• Revamped the activity API endpoints and functionality to work with the new WIM2 implementation of KE-chain 2.10 (MAR18)
• Added a number of predicated on the Activity object to simplify the introspection of the Activity, eg. is_rootlevel, is_root, is_workflow...
• We added a translation layer that automatically detects if you connect to WIM1 or WIM2 and automagically translates activity_class and activity_type and the assigneesids (in lieu of usernames).
• Added the Actvity2.parent() function to retrieve the parent (in lieu for subprocess())
• Added User object in pykechain to check the users in a KE-chain instance.
• KE-chain for WIM2 added also a version endpoint to check the version of the individual KE-chain 'apps' such as WIM. It is used to automatically give you back the Activity class based on the version you are using. You can check out the client.app_versions property.
• Fixed the way you limit the scope search in Scope.activities() and Scope.activity()

Backward incompatible changes

• Deprecated the 'inspector components' including its base classes.
• Deprecated the 'single reference property', which is replaced by the MultiReference property.

Pending Deprecation Warnings

• In May 2018 we will deprecate the support of WIM1 in pykechain. PendingDeprecationWarnings are in place when you use WIM1 Activity.

1.16.0 (14MAR18)

This is the last release in preparation for the WIM2 release of KE-chain and consequently pykechain. In the next version of pykechain, some backward incompatible changes will happen. A migration path is provided as well.

• Implemented new functions for adding the following widgets: Basic table, Paginated table, JSON, Script, Notebook, Text, Attachment viewer and Navigation Bar. (#280)
• Added two new enums (SortTable and NavigationBarAlignment) which can be used when adding new widgets. (#280)
• Added additional enums WidgetNames for the proper names of the widgets in the customisation dialog in KE-chain. (#280)
• Updated the documentation regarding Property Types. (#280)
• Fixed the enums so they now work correctly for each PropertyType. (#280)
• Wrote a test that tests each property type (we didn't have this before). This will increase the test coverage to ~95% for KE-chain. (#280)
• Added a new function called Client.property(), which allows the user to retrieve one property. (#296)
• Improved the overall test coverage for ExtCustomization class to 100%.
• increased the coverage of MultiReference property tests to 100%. (#296)
• increased the coverage of SelectListProperty tests to 100%. (#296)
• setting the value of a SelectListProperty instance to None empties it.
• increased the coverage of Scope tests to 100%. (#296)
• increased the coverage of Client tests to 99%. (#296)
• increased the coverage of Service tests to 90%. (#296)
• Updated dependent versions for development: betamax (0.8.1), twine (1.10.0), matplotlib (2.2.0), pytest (3.4.2), mypy (0.570), sphinx (1.7.1)

1.15.4 (15FEB18)

• Fixed an issues where the MultiReference property did not provide the correct choices. Also fixed a bug where the setting of the value is now performed more robust. (#282)

1.15.3 (8FEB18)

• Fixed a bug where the MultiReference property only provided back the 'last' Part in the internal value due to the way a library parses a list. Fixed that and added tests (#276). Thanks again to @raduiordache.

1.15.2 (5FEB18)

• Fixed a bug where the MultiReference property could not retrieve parts through the API based on the value of the MultiReference property as it incorrectly retrieved the 'id' from the value list (#274). Thanks to @raduiordache!

1.15.1 (2FEB18)

• The Part.property() method was slightly changed in 1.15 (the argument name was name and became name_or_id). This is reverted to name to be compatible with older pykechain releases. (#271)
• Updated dependent versions for development: pytest (3.4.0)

1.15.0 (25JAN18)

• added ability to provide additional keyword=value arguments to many of the part and property methods that either create parts or update properties (#260). This facilitates the use of suppress_kevents=True that you might want to use for a backend performance boost. This is a trade-off that the frontend will not be informed of any property updates or new parts until after a reload of the page in the KE-chain frontend application. When you can use suppress_kevents=True in the method, it is documented in the function. This can be found in de Developer API docs. Examples of functions that can handle the supress_kevents=True as additional keyword=value argument are: Client.create_part(), Client.create_model(), Part.add(), Part.update(), Part.edit() and more like these.
• added validation of a single select list. The value is not set when it is not in the list of options (#259).
• enabled to use of property model UUID in the Part.add_with_properties() next to using property names. (#258)
• enabled to search for properties using UUID next to using property names. This is provided for you in Part.property().
• The Part.update() is now considerate if you provide the property UUIDs inside the update_dict as well as property names. You can even mix UUIDs and property names together. (#263) Thanks to @raduiordache.
• functions and methods that check if the correct type was provided to the method as arguments that raised TypeError before, are now raising IllegalArgumentError.
• The Activity.customize() method and the InspectorComponents are now deprecated (since Nov 17) and will raise deprecation errors when called. Use Activity.customization() to retrieve the new activity Customization objects.
• updated dependent versions for development: pytest (3.3.2), sphinx (1.6.6), nbsphinx (0.3.1), matplotlib (2.1.2), mypy (0.560)
• updated documentation with additional crosslinks and better references.
• added source code to all API documentation

1.14.0 (11DEC17)

• In preparation for the release of KE-chain 2.7.0-132, we added support for multireference properties in pykechain. Pykechain 1.14 is compatible with both older versions of KE-chain as well as the ones supporting multireference properties. The main difference is that you need to provide a list of Part`s or `part_id`s instead of a single `Part or part_id. It will override the value in KE-chain fully, no adding or substraction methods are provided, you need to do that in your own code.
• Updated documentation for the MultiReferenceProperty.
• updated dependent versions for development: pytest (3.3.1)

1.13.3 (5DEC17)

• added the 'type' attribute to a property (#248)
• updated the enums documentation to include all the possible enums available (#247)

1.13.2 (4DEC17)

• A wrong statuscode check prevented the upload of a script to complete fully in pykechain. The script is properly uploaded, but pykechain checked against wrong code (#246).
• updated dependent versions for development: pyopenssl (17.5.0), pytest (3.3.0), nbsphinx (0.2.18)

1.13.1 (16NOV17)

• Added additional xtypes to the list of allowed xtypes in the customizations in order to support our new widgets. (#240)
• updated dependent versions for development: pytest (3.2.5)

1.13 (9NOV17)

• Added Service and ServiceExecution models to pykechain (#231). This includes the ability to create, retrieve, edit, destroy and upload kecpkg files to KE-chain. Also the retrieve, terminate, and download log results for ServiceExecution`s (which are associated to `Service`s) are available. This brings `pykechain in line with the full abilities in KE-chain 2 SIM release (31OKT17) (SIM module license needed). Also see the PyPI package kecpkg-tools from KE-works to help you smoothen the workflow of creating custom KE-chain supported python packages (kecpkg) that can be executed by the KE-chain SIM module.
• Added additional keywords arguments in the scope and activity searchers. You can now craft complex search queries to the KE-chain API (#231)
• Prevented the creation of Activities with incorrect activity_class. This is now prevented in pykechain (#225)
• Added an option for all models to reload (will return a new object) and refresh (will update in place). (#232)
• Added additional tests and improved documentation for Service and ServiceExecution models.
• updated dependent versions for development: matplotlib (2.1.0), nbsphinx (0.2.16), flake8 (3.5.0), sphinx (1.6.5), mypy (0.540), pydocstyle (2.1.1)

1.12.9 (5OCT17)

• Improved scope control for activity queries. Will ensure that the scope_id of an acitivity is properly retrieved and checked for in case of subqueries such as Activity.children(), siblings, subprocess.
• Updated dependent versions for development: pytest updated to 3.2.3 (#215)

1.12.8 (2OCT17)

• Fixed a bug where the scope object was ambigously retrieved during the edit assignees of an activity action. It failed when the scope was closed. Thanks to @raduiordache for its find! (#211)
• Updated dependent versions for development for tox to 2.9.1 and Sphinx to 1.6.4 (#198, #209)

1.12.7 (2OCT17)

• Fixed a bug where a model without an instance raises an incorrect Error. Now it will raise a NotFoundError (#207).

1.12.6 (28SEP17)

• Fixed a bug in the models.customisation. After a succesfull save of a customisation to an activity, the activity could not be retrieved from KE-chain if the activity was part of a closed scope (#205).

1.12.5 (28SEP17)

• The get_project() helper method will now retrieve a scope a status other than 'ACTIVE' only (#203).
• Updated the documentation to fix wrongly formatted examples.

1.12.4 (26SEP17)

• Fixed a bug in the customization code by which the activity was incorrectly updated after a correctly saved customization to the KE-chain server. In some cases the incorrect customisation was retrieved on name basis, which may resulted in an error raised. Thanks to @raduiordache for finding it (#200).
• Added **kwargs to the Part.children(), Part.siblings(), Part.instances(), Activity.children(), and Activity.siblings() methods. This will enable more comprehensive searches, eg. by the name of children using Activity.children(name='Some childs name') (#199).

1.12.3 (21SEP17)

• Fixing the warning: 'could not any envfile' from envparse. Which is suppressed for cosmetics. It is advised to provide a pathname for the envfile when you want to load the environment variables from an envfile (#195).
• Fixed tests for the envparse warning and refactored the tests to better deal with in-test settings of the environment.

1.12.2 (15SEP17)

• Removed a logical error in the checking of the existing of the environment variables.

1.12.1 (15SEP17)

• Added the ability to enforce the use of environment variables when the KECHAIN_FORCE_ENV_USE is set to a true value in the environment. Altered documentation and altered tests for that (#193).

1.12 (14SEP17)

• Added a new helper get_project() to bootstrap a pykechain client and return a project (aka Scope) immediately. You can retrieve a project using direct arguments url, token (or username and password), and scope_id (or scope name). Alternatively, you can provide an .env file or provide the arguments from the environment as the environment variables KECHAIN_URL, KECHAIN_TOKEN (or KECHAIN_USERNAME and KECHAIN_PASSWORD), and KECHAIN_SCOPE_ID (or KECHAIN_SCOPE) (#185). This is ideal for pykechain scripts in the KE-chain SIM, as we provide support for this to make your scripting experience in KE-chain buttersmooth. An example:

  >>> from pykechain import get_project
  >>> project = get_project(url='http://localhost:8000', username='foo', password='bar', scope='Bike Project')
  >>> print(project.name)

• Added additional checks for the Client to check if the url provided is correct (#185).

• Improved the state of the project on codacy, a nice code quality monitor, from B to A grade. Removed over 100 insecure code elements, according to codacy. See: https://www.codacy.com/app/KE-works/pykechain/dashboard (#187).

• Updated dependent versions of pyopenssl to 1.1.2 (#188), pytest to 3.2.2 (#183) and tox to 2.8.2 (#184).

• Updated coverage of the files to internal standards. The critical models are now 100% tested such as the Client, Activity and Part. (#190) see: https://coveralls.io/github/KE-works/pykechain.

1.11.1 (4SEP17)

• Added the ability to clear and attachment field (unlink the attachment). Please refer to the AttachmentProperty.clear() method.
• Ensured a more robust updating of property value all over by altering Property._value and Property._json_data['value'] after you set a value on a property.

1.11 (4SEP17)

• In KE-chain 2.5 the way we use task customization has changed drastically. Pykechain (from 1.11 onwards) supports this by implementing a new concept in the activity called Activity.customization() (#161). This provides you an ExtCustomization object, which you can inspect and add new widgets. Please see the documentation on ExtCustomization and Activity.customization() for more details. An example to use is:

  >>> activity = project.activity(name='Customizable activity')
  >>> customization = activity.customization()
  >>> part_to_show = project.part(name='Bike')
  >>> customization.add_property_grid_widget(part_to_show, custom_title="My super bike"))

• Removed previously announced deprecated method for activity.create_activity() (use Activity.create()).

• Added deprecation warnings when using InspectorComponent objects and old style Customization components. They will be removed in November 2017 (introduced in pykechain 1.9)

• Added the ability to retrieve a list of project members and managers with the Scope.members() method (#169)

• Added the ability to manage member and managers of a scope. See the Scope.add_member, add_manager, remove_member, remove_manager (#175)

• Added the ability to add additional keyword arguments for the methods Part.update(), Part.add_with_properties() and Part.edit(). This will allow to provide additional (including undocumented) arguments to the KE-chain API. (eg. 'suppress_kevents=True') (#177)

• Added the ability to edit the name of the property, its description and the unit (#146, PR #179)

• Added classification enumeration (#175)

• Updated the documentation structure to better access all the pykechain models related documentation. See http://pykechain.readthedocs.io/en/latest/developer_api.html

• Updated dependent version of tox to 2.8.0 (#178) and further to 2.8.1 (#180)

• Updated all tests such that our coverage aim of 96%+ is maintained.

1.10.3 (28AUG17)

• Corrected the creation of partmodels (Part with category MODEL) with multiplicities other than ZERO_MANY as the provided multiplicity option was not respected in the create_model() method of Client and Scope. Thanks @raduiordache for the find. (#170)
• Updated tests.

1.10.2 (22AUG17)

• Corrected the ability to assign multiple assignees, using a list of assignees to an activity using the Activity.edit() method. (#167)
• Updated tests.

1.10.1 (18AUG17)

• updated incorrect tests related to Activity.associated_parts(). (#96, #149)

1.10.0 (18AUG17)

• Ability to edit the status of an Activity. Please refer to the ActivityStatus enumerations. (#163)
• Ability to sort properties of a Part model. (#141)
• Upgraded the requirements of dependent packages for development. (#152, #160, #159, #153, #157, #154)
• Added tests for all new features to get the > 95% coverage
• Updated the documentation.

1.9.1 (27JUN17)

• Improved testing. Notably on the new inspector objects. No functional change only that we want to reach our goal of 95% test coverage! Thanks to @raduiordache (#137)

1.9.0 (23JUN17)

• Added a major new feature to create Customization`s of activities in KE-chain 2 all programmatically and pythonic. We provide building block classes such as `SuperGrid, PaginatedGrid`s and `PropertyGrid`s to make your own task customization. All is documented with examples. A `validation() method is available. (#110)

  >>> my_task = project.activity('my task')
  >>> bike = project.part(name='Bike')
  >>> customization = Customization()  # init customization object for the task
  >>> my_prop_grid = PropertyGrid(part=bike, title=bike.name)  # create a PropertyGrid
  >>> customization.add_component(my_prop_grid)  # add PropertyGrid to the Customization component list
  >>> customization.validate()  # you can validate the customization
  >>> my_task.customize(customization)  # upload/set the Customization. Ensure you have data access set correctly.

• Updated the way the Activity.customize() method works. This method now accepts a Customization object or a josn (as a python dict). It uses the Customization.validate() method to validate if it conforms to the required json structure before uploading it to KE-chain.

• Improved test coverage and refactored the HTTP codes to human readable form. (#128)

• Added the ability to edit the description of property models. This was included in Part.create_property(... description=...) (#135)

• Add Part.as_dict() method to retrieve the properties of a part in pykechain as a python dictionary as {<property_name> : <property_value>} (#131)

• Added the ability to optionally update the name of a part together with the value of its properties. See the Part.update() method. (#126)

• Deprecated the Activity.create_activity() method in favor of Activity.create(). Use the latter. Will warn with a DeprecationWarning until removed.

1.8.0 (05JUN17)

• Added Part.instances() method for models to find their associated instances. (#113) Also added a Part.instance() method if you for sure that you will get only a single instance back.
• Added Activity.subprocess(), Activity.siblings() and Activity.children() methods to the Activity. It eases relative retrieval of other tasks in the task tree. Documentation is included. (#100)
• added Activity.activity_type property to the Activity.
• added ActivityType enumeration. This can be used to check if the activity_type of an Activity is either a Usertask or a Subprocess.
• Added ability to retrieve an Activity based on an id. As this included in the low level Client object, it can be used almost everywhere to retrieve an activity by its id (or primary key, pk) eg. in the Scope.activity.
• Added ability to add additional keywords to the activities searcher to be able to search by name, pk, container etc.
• Added a FutureDeprecationWarning to the Activity.create_activity() method. This will is replace with the Activity.create() method. Update your code please!
• Added a convenience method to retrieve models and instances related to a task at once: Activity.associated_parts(). Making use of the already provided method in Activity.parts(). (#118)
• Added missing tests for Activity.parts() and Activity.associated_parts()
• added tests for all new features.
• Updated the documentation.

1.7.3 (01JUN17)

• Updated documentation for activity startdate and duedate editting using timezone supported datetime objects. If a user want to make use of timezone aware datetime the best way to do it is:

  >>> my_tz = pytz.timezone('Europe/Amsterdam')
  >>> start_date = my_tz.localize(datetime(2017,6,1,23,59,0))
  >>> due_date = my_tz.localize(datetime(2017,12,31))
  >>> my_task.edit(start_date = start_date, due_date = due_date)
  

• Fixed a bug where a naive due_date and no provided start_date resulted in an error. Keep them bugs comin'!

1.7.2 (01JUN17)

• updated property.part property that gets the part for its property. For model this did not work as underlying only category=INSTANCES were retrieved. Thanks to @joost.schut for finding it and reporting.
• updated requirements for development.

1.7.1 (29MAY17)

• Added Part.multiplicity property method. Use the pykechain.enums.Multiplicity to check the multiplicity of a part against the set multiplicities.
• Updated documentation (a.o. the Basic Usage notebook).

1.7.0 (29MAY17)

• Added ReferencyProperty.choices() convenience method to provide you the list of parts (instances) that are acceptable as a choice for the value of the reference property.
• Added Part.proxy_model() method that will return the model that is used as the basis for the proxied model. A proxied model is a linkedcopy of the proxy_model that can have a different name, but follow that exact model definition of the proxy_model including its properties and submodel parts. A handy way to create model structures in the catalog world and use a multiple of those in the product world. An example is provided in the documentation.
• Added the Activity.edit() method to be able to update the activity name, description, start_date, due_date and assignee.
• Added the ability to customize an activity using the Activity.customize() method using a proper inspector NG json.
• Upgraded package requests to the latest version.
• Updated documentation according to PEP257.
• Updated requirements for development.

1.6.0 (3MAY17)

• Added a Part.model() method to retrieve the model from an instance.
• (Backwards Incompatibile) The task configuration (association) API is updated to the latest KE-chain release (release 2.1.0b-sprint119 30MAR17). This affects the activity.configure() method. This change is not compatible with older KE-chain 2 releases. For older KE-chain 2 releases use a pykechain version < 1.6
• Added Getting Started documentation page for pykechain using jupyter notebooks
• Documentation update for the reference property
• Updated documentation according to PEP257

1.5.1 (6APR17)

• Patch release to include the python package typing.

1.5.0 (6APR17)

• Added ability to edit the part name and description functionality. See the Part.edit() method.
• Added the ability to use the bulk_update_properties API endpoint for KE-chain releases later then 2.1.0b. No need to alter your pykechain code. The implementation of Part.update() method is augmented to use this faster method of uploading changes to property values. For connections to legacy KE-chain 2 instances, use the switch bulk=False.
• Added the ability to create a new part and provide its properties values for KE-chain releases later then 2.1.0b. You can use the new Part.add_with_properties() method and it will connect to the new KE-chain API endpoint of 'new_instance_with_properties'. Properties are provided by name and value in a dict. For examples see the docs.
• Reference properties can now be set with a Part directly. Setting a reference property to None will clear the value.
• Added the ability to create a proxy model with Part.add_proxy_to() and Client.create_proxy_model(). For exmaples see the documentation.
• Added enumerations for Category and Multiplicity in pykechain.enums. You can use these constants to ensure that these values are correct, aligned and thusfor accepted by KE-chain. Examples are included in the documentation.
• Attachment properties have now a value set if there is a file attached in KE-chain. Otherwise the value is None. Now you are able to check if there is a file attachment set before you download or upload. See the docs for examples.
• Added type annotations throughout the code and added mypy to the continuous integration pipeline to ensure high quality of the code provided.
• Improved the documentation

1.4.0 (17FEB17)

• Added functionality to create part models, just as you create part instances.
• Added functionality to upload files (using filename), python objects (as json) and matplotlib figures as attachments
• Added functionality to download attachments directly as file or python objects (from json).
• Improved logic if you request children or siblings that the corresponding category (Model or Instance) is retrieved.
• Improved continuous integration.
• Improved documentation

1.3.0 (16FEB17)

• Added functionality to support a select list property type from KE-chain in pykechain. Now you can inquire for the options and set the list of options (on the model) to choose from. See documentation of SelectListProperty.
• Added additional keyword arguments to the Client.parts() method. This allows access to additional filters on the KE-chain REST API.
• Fixed a bug that shows a warning when importing pykechain without a .env file being present. Improved documentation of the Client.from_env() method. Including an example of this .env file.
• Improved documentation
• Improved testing (and coverage)
• Improved introspection of PartSet.

1.2.0 (14FEB17)

• batch updates of properties in a part using a dictionary is now possible using the part update({'prop_name': val}) Part method.
• added relational methods on the part like: Part.parent(), Part.children() and Part.siblings(). See documentation of Part for that.
• improved documentation
• improved testing
• improved introspection of objects due to correct representation for debugging
• version number now available through pykechain.version

1.1.2 (7FEB17)

• improved (iterative) part retriever capability with a batch processed request. Will enable to retrieve large datasets than normal, that take longer than a standard timeout. Will concatenate the results of the various requests. Check out the documentation for the new limit and batch parameters on the Client.parts method.
• improved upload of files and attachments
• added bucket and limit filters to limit the number of parts retrieved
• improved testing
• improved coverage
• updated dependencies
• improved documentation

1.0.0 (3JAN17)

• First public release of pykechain
• ability to create a client to connect to a KE-chain instance
• ability to retrieve parts and properties within a KE-chain scope (project)
• ability to retrieve activities with a KE-chain scope
• ability to upload and download a property value

0.1.0.dev0 - 0.1.1.dev0 (23DEC16)

• Initial pre-release of pykechain
• ability to create a client to connect to a KE-chain instance
• ability to retrieve parts and properties within a KE-chain scope (project)
• limited ability to upload and download a property value