Gemd

Improvements

This update modifies the cake sample data to use the InChI of vanillin rather than the InChI for sodium chloride.

What’s New
* We've added support for [molecular structures](https://citrineinformatics.github.io/gemd-docs/specification/value-types/#molecular-structure). Now information can be stored and communicated about the structure of a molecule in either SMILES or InChI format.

Improvements
* We've updated the documentation on [how to contribute](https://github.com/CitrineInformatics/gemd-python/blob/master/CONTRIBUTING.md). It provides detailed information on how to deploy a new version to pypi as well as clear documentation of our coding standards.
* We've augmented our `cake` demo to include the new molecular structure bounds and value types. Take a look at the Abstract Salt and Abstract Sugar nodes.

Fixes
* We fixed an error with the fuzzing behavior on `cake` examples so that you can no longer get volume fractions greater than 1.


Obsoleted
* With the transition past `v0.7.0`, we've removed the now-obsolete support for the taurus namespace.

This release improves the Strehlow and Cook demo data.

What's New
- The Strehlow and Cook demo template scope is now configurable. The desired template scope can be passed as an argument to the command line invocation or to any method that uses it. The scope defaults to the previous value to ensure backwards compatibility

Fixes
- The Strehlow and Cook demo now uses unified material specs for replicate chemical formulas, as opposed to the previous behavior of creating a new one each time.

This release migrates to the official `gemd` name. The contents of the `taurus` module have been moved to `gemd`, but the `taurus` module was deprecated rather than removed. It will be removed in `0.7.0`.

What's new
* The `gemd` module! GEMD stands for "Generalized Expression of Materials Data" and is the official name of Citrine's data model! The "taurus" code name served us well and will be missed.

Improvements
* The `gemd.entity.bounds`, `gemd.entity.attribute`, `gemd.entity.value`, and `gemd.entity.template` modules join `gemd.entity.object` in importing their members, allowing imports like `from gemd.entity.template import ProcessTemplate, ParameterTemplate`.

Deprecated
* The `taurus` module is deprecated and will be removed in the next feature release. Migration involves replacing `taurus` with `gemd` in imports and `TaurusEncoder`/`TaurusJson` with `GEMDEncoder` and `GEMDJson`.

0.6.0 is released! This release focuses on significant changes to JSON support, which should provide a more intuitive experience to users and a more expressive and extensible experience to developers.

What’s New
* `taurus.json` has been added as a drop-in replacement for python's builtin `json` module. (https://github.com/CitrineInformatics/taurus/pull/86)
* `taurus.json.TaurusJson` has been added to encapsulate the configuration of a taurus-compatible json serializer/deserializer. (https://github.com/CitrineInformatics/taurus/pull/86)
* `TaurusJson` objects have a new `register_classes` method to support the extension of json support to other classes in packages that develop on top of `taurus`. (https://github.com/CitrineInformatics/taurus/pull/86)
* `taurus.util` includes a public `writeable_sort_order` method that accepts either data objects or type strings. (https://github.com/CitrineInformatics/taurus/pull/85)

Improvements
* Output materials are now included when serializing processes. (https://github.com/CitrineInformatics/taurus/pull/87)

Bugfixes
* Measurements are now correctly included when serializing materials. (https://github.com/CitrineInformatics/taurus/pull/84)

Changes
* `taurus.client.json_encoder` has been moved to `taurus.json`. The `taurus.client.json_encoder` is now deprecated and will be removed in `v0.7.0`. (https://github.com/CitrineInformatics/taurus/pull/86)

What’s New
* The cake examples has been extended to demonstrate more of the data model
* The Strehlow and Cook example now imports directly from the Citrination PIF and includes a substantially increased number of attributes

Improvements
* The `flatten` method has been altered so that entire material histories can be flattened. The order of the returned objects is such that they can be created and re-linked in that order.
* The `raw_dumps` and `raw_loads` methods were added to the `json_encoder` to provide access to the custom class encoder/decoder without the normalization/denormalization logic present in the full `dumps` and `loads` methods.