Build Status

Purpose

LutaML is a data model accessor that supports various data model formats, including:

  • EXPRESS (*.exp files)

  • OMG UML in XMI format (*.xmi files)

This plugin allows you to access LutaML models from within a Metanorma document.

Installation

$ gem install metanorma-plugin-lutaml

Usage with EXPRESS

General

LutaML supports accessing EXPRESS models via the Expressir parser.

Usage of the lutaml command

Given an example.exp EXPRESS file with content:

SCHEMA test_schema 'test';

(* Need select elements for measure_value *)
 REFERENCE FROM measure_schema
   (measure_value);

  TYPE my_type1 = EXTENSIBLE SELECT;
  END_TYPE;

  TYPE my_type2 = EXTENSIBLE ENUMERATION;
  END_TYPE;

  TYPE my_type3 = EXTENSIBLE ENUMERATION;
  END_TYPE;

  TYPE my_type4 = EXTENSIBLE ENUMERATION;
  END_TYPE;

  TYPE my_type5 = EXTENSIBLE ENUMERATION;
  END_TYPE;
END_SCHEMA;

And the lutaml block:

[lutaml,example.exp,my_context]
----

{% for schema in my_context.schemas %}
== {{schema.id}}

{% for entity in schema.entities %}
=== {{entity.id}}
{% endfor %}

{% endfor %}
----

Where:

  • content within the block is called the “template”;

  • {example.exp} is the location of the EXPRESS schema file (*.exp) that contains data to be loaded. Location of the file is computed relative to the source directory that [lutaml] is used (e.g., if [lutaml,example.exp,my_context] is invoked in an .adoc file located at /foo/bar/doc.adoc, the data file is expected to be found at /foo/bar/example.exp);

  • {my_context} is the name where the EXPRESS Repository read from the .exp file can be accessed with.

    • The context object is a serialized Expressir::Model::Repository object with all variable names available. See Expressir docs for reference. {my_context} has schemas method to access Expressir schemas

Will produce this output:

== test_schema

=== my_type1 === my_type2 === my_type3 === my_type4 === my_type5

This command also supports .lutaml files.

Instead of using the direct path to the file one can use lutaml-express-index document attribute to supply directory with express files or YAML index file to parse as well as the cache file location.

The syntax is as follows:

:lutaml-express-index: my_custom_name; dir_or_index_path[; cache=cache_path]

Where:

  • my_custom_name - is name of the parsed EXPRESS files context for the later use with lutaml macro

  • dir_or_index_path - location of directory with EXPRESS files or path to the YAML index file to parse

  • cache_path - optional, location of the Expressir cache file to use

Example of usage:

= Document title
Author
:lutaml-express-index: index_name; /path/to/express_files; cache=/path/to/cache_file.yaml

[lutaml,index_name,context]
----
{% for schema in context.schemas %}
== {{schema.id}}
{% endfor %}
----

Using config.yaml

This functionality allows [lutaml_express] blocks to load a full set of EXPRESS schemas in one index, and then provide a select ("filter") option per-block via a separate YAML file.

:lutaml-express-index: all_schemas; ../schemas_all.yaml;

[lutaml_express,all_schemas,context,leveloffset=+1,config_yaml=schemas.yaml]
---
{% assign selected = context.schemas | where: "selected" %}
{% render "templates/resources/schema" for selected as schema %}
---

Where schemas_all.yml provides all schemas:

---
schemas:
  action_schema:
    path: "../../schemas/resources/action_schema/action_schema.exp"
  application_context_schema:
    path: "../../schemas/resources/application_context_schema/application_context_schema.exp"
  approval_schema:
    path: "../../schemas/resources/approval_schema/approval_schema.exp"
...

And schemas.yaml only selects 2 schemas:

---
schemas:
  action_schema:
    anything: ...
  application_context_schema:
    anything: ...

The resulting block adds the select attribute to every schema of the the "context" object, which allows you to filter those out for complex operations via Liquid:

[lutaml_express,schemas_1,repo,leveloffset=+1,config_yaml=select.yaml]
---
{% assign selected = repo.schemas | where: "selected" %}
... do things with `selected` ...
Note
This functionality is used in the ISO 10303 SRL to load the full schema set at once but only render the selected schemas in individual documents.

Usage with UML

Rendering a LutaML view: lutaml_diagram

This command allows to quickly render a LutaML view as an image file.

Given a file example.lutaml file with content:

diagram MyView {
  title "my diagram"

  enum AddressClassProfile {
    imlicistAttributeProfile: CharacterString [0..1] {
      definition
        this is multiline with `asciidoc`
      end definition
    }
  }

  class AttributeProfile {
    +addressClassProfile: CharacterString [0..1]
    imlicistAttributeProfile: CharacterString [0..1] {
      definition this is attribute definition
    }
  }
}

The lutaml_diagram command will add the image to the document.

lutaml_diagram::example.lutaml[]

The lutaml_diagram command can also be used to denote a block with an embedded LutaML view.

For example:

[lutaml_diagram]
....
diagram MyView {
  title "my diagram"

  enum AddressClassProfile {
    imlicistAttributeProfile: CharacterString [0..1] {
      definition {
        This is multiline AsciiDoc content.
      }
    }
  }

  class AttributeProfile {
    +addressClassProfile: CharacterString [0..1]
    imlicistAttributeProfile: CharacterString [0..1] {
      definition this is attribute definition
    }
  }
}
....

Rendering a LutaML Enterprise Architect diagram from XMI: lutaml_ea_diagram

This command allows to quickly render a LutaML diagram as an image file by specifying the name of diagram.

lutaml_ea_diagram::[name="name_of_diagram",base_path="/path/to/xmi-images",format="png"]

The code will search the diagram with name name_of_diagram and then render it as:

[[figure-{{ diagram.xmi_id }}]]
.{{ diagram.name }}
image::{{ image_base_path }}/{{ diagram.xmi_id }}.{{ format | default: 'png' }}[]

Rendering a LutaML GML Dictionary: lutaml_gml_dictionary

This command allows to render a LutaML GML Dictionary by using Liquid Drop.

GmlDictionaryDrop has the following attributes:

  • name

  • file_name

  • dictionary_entry

Each dictionary_entry has the following attributes:

  • name

  • description

lutaml_gml_dictionary::/path/to/dictionary.xml[template="/path/to/template.liquid",context=dict]

The command accepts the options listed below:

  • /path/to/dictionary.xml specifies the path of xml file of the GML Dictionary.

  • template="/path/to/template.liquid" specifies the liquid template. For example, you can create a liquid template and link it by template.

  • context=dict define the context in the template.

[cols="3a,22a"]
|===
| Name | {{ dict.file_name }}

h| Code h| Description
{% for entry in dict.dictionary_entry %}
| {{ entry.name }} | {{ entry.description }}
{% endfor %}
|===

[.source]
<<source_link>>

In spite of specifying the path of the template, you can also define an inline template within a block by [lutaml_gml_dictionary,"/path/to/dictionary.xml",context=dict].

[lutaml_gml_dictionary,"/path/to/dictionary.xml",context=dict]
--
{% capture link %}https://www.test.com/{{ dict.file_name }}{% endcapture %}

[cols="3a,22a"]
|===
| File Name | {{ dict.file_name }}
h| URL | {{ link }}
h| Help | Description
{% for entry in dict.dictionary_entry %}
| {{ entry.name }} | {{ entry.description }}
{% endfor %}
|===

[.source]
<<source_link>>
--

Rendering a LutaML table of a class: lutaml_klass_table

This command allows to render a LutaML table of a class by using Liquid Drop.

The table will show:

  • Class Name

  • Class Definition

  • Inherited Properties

  • Self-defined Properties

  • Properties Inherited from Association

  • Properties Defined in Association

lutaml_klass_table::/path/to/example.xmi[name="NameOfClass",template="/path/to/templates/_my_klass_table.liquid"]

The command accepts the options listed below:

  • /path/to/example.xmi specifies the path of xmi file.

  • name="NameOfClass" specifies the name of the Class.

  • template="/path/to/templates/_my_klass_table.liquid" specifies the path of the liquid template. (Optional) By default, it will look for the template _klass_table.liquid defined in lib/metanorma/plugin/lutaml/templates. This template can be customized by changing the template path in the template option.

  • `guidance="/path/to/my_guidance.yml"`specifies the path of the yaml file of the guidance. (Optional)

The guidance file should be in the following format:

---
classes:
  - name: Name Of Class
    attributes:
      - name: Name Of Attribute (e.g. gml:boundedBy)
        used: false
        guidance: |
          Drop guidance message here.
...

If you want to define the guidance, you can define the name of the class under classes. Then define which attributes you want to add guidance by the name. Set used to show the attribute is used or not. Drop the message of guidance in guidance.

Generating UML class and attributes: lutaml_uml_class

This command allows rendering a definition clause for a UML class.

Given example.lutaml with this content:

class Register {
  definition {
    A register of information.
  }

  identifier: String[1] {
    definition {
      Unique identifier of the register.
    }
  }

  concepts: Concept[0..*] {
    definition {
      Concepts.
    }
  }
}

The command:

[lutaml_uml_class,views/Register_Register.lutaml,Register]

Will produce this output:

== Register

A register of information.

=== Attributes

==== identifier

Unique identifier of the register.

Table 1. Specification of Register.identifier

Value type and multiplicity

String [1]

==== concepts

Concepts.

Table 2. Specification of Register.concepts

Value type and multiplicity

Concepts [0..*]

The command accepts two options:

  • skip_headers=true (or just skip_headers). The initial heading (the UML class name) will not be generated. This is useful if additional content is to be supplied to the clause, such as diagrams that are defined outside the UML model.

  • depth={n}. (default: 2) This determines the depth of the generated headings. A depth of 2 means the initial heading will have 2 equal signs, and so forth. The heading depth of the attributes are in relation to the initial depth, so a depth of 2 will have the "Attributes" section at depth 3.

lutaml_uml_attributes_table

This command allows rendering definition tables for a UML model.

Given example.lutaml file with the content:

diagram MyView {
  title "my diagram"

  enum AddressClassProfile {
    imlicistAttributeProfile: CharacterString [0..1] {
      definition
        this is multiline with `ascidoc`
      end definition
    }
  }

  class AttributeProfile {
    +addressClassProfile: CharacterString [0..1]
    imlicistAttributeProfile: CharacterString [0..1] {
      definition this is attribute definition
    }
  }
}

And the lutaml_uml_attributes_table macro:

[lutaml_uml_attributes_table, example.lutaml, AttributeProfile]

Will produce this output:

=== AttributeProfile

Table 3. AttributeProfile attributes
Name Definition Mandatory/ Optional/ Conditional Max Occur Data Type

addressClassProfile

TODO: enum’s definition

M

1

CharacterString

imlicistAttributeProfile

this is attribute definition with multiply lines

M

1

CharacterString

In case of "enumeration" (AddressClassProfile) entity:

[lutaml_uml_attributes_table, example.lutaml, AddressClassProfile]

Will produce this output:

=== AddressClassProfile

Table 4. AddressClassProfile values
Name Definition

imlicistAttributeProfile

this is multiline with asciidoc

Usage of lutaml_uml_datamodel_description macro

This command allows to quickly render data model packages and its dependent objects for supplied XMI file.

Given an Enterprise Architect example.xmi file with 2 packages:

  • 'Another'

  • 'CityGML'

The lutaml_uml_datamodel_description macro can be used:

[lutaml_uml_datamodel_description, path/to/example.xmi]
--
[.before]
....
my text
....

[.diagram_include_block, base_path="requirements/", format="emf"]
....
Diagram text
....

[.include_block, package="Another", base_path="spec/fixtures"]
....
my text
....

[.include_block, base_path="spec/fixtures"]
....
my text
....

[.before, package="Another"]
....
text before Another package
....

[.after, package="Another"]
....
text after Another package
....

[.after, package="CityGML"]
....
text after CityGML package
....

[.after]
....
footer text
....
--
--

Where:

  • path/to/example.xmi - required, path to the XMI file to render

  • [.before] - block text that adds additional text before the rendered output, can be used only once, additional occurrences of macro will overwrite text, not that literal block style must be used in there(eg …​.)

  • [.after] - block text that adds additional text after the rendered output, can be used only once, additional occurrences of macro will overwrite text

  • [.after, package="Another"] - block text to be inserted before(after in case of .before name) the package

  • [.package_text, position="after", package="Another"] - include custom adoc code into package rendered body, position is a a required attribute which tells where to insert the code.

  • [.package_text, package="Another"] - same as above, but include block will be included only for supplied package name

  • [.diagram_include_block] - block text to automatically include diagram images. Attribute base_path is a required attribute to supply path prefix where to look for a diagram image. format is an optional attribute that tells what file extension to use when including diagram file.

    The logic is as follows:

{% for diagram in package.diagrams %}
[[figure-{{ diagram.xmi_id }}]]
.{{ diagram.name }}
image::{{ image_base_path }}/{{ diagram.xmi_id }}.{{ format | default: 'png' }}[]

{% if diagram.definition %}
{{ diagram.definition | html2adoc }}
{% endif %}
{% endfor %}

For instance, the script will take package diagrams supplied in the XMI file and will try to include image with the name equal to diagram' xmi_id attribute plus .png. Also one can add any text to the macro text, it will be added as paragraph before each image include.

  • [.diagram_include_block, package="Another"] - same as above, but diagram will be included only for supplied package name

  • [.include_block, base_path="spec/fixtures"] - macro to include files (.adoc or .liquid) for each package name. Attribute base_path is a required attribute to supply path prefix where to look for file to include. Macro will look for a file called base_path + / package_name(downcase, replace : → '', ' ' → '') + .adoc[.liquid], eg for package 'My Package name' and base_path eq to my/path, macro will look for the following file path: my/path/_my_package_name.adoc.

  • [.include_block, package="Another", base_path="spec/fixtures"] - same as above, but include block will be included only for supplied package name

Note
.after, .before, package_text and include_block macroses all can be used with additional option - liquid, if this option is supplied then the code inside block will be interpolated in liquid context

There are two other commands that are used to refer to LutaML generated document elements:

  • lutaml_figure. Provides a reference anchor to a figure defined in the XMI file, using its XMI ID for reference.

  • lutaml_table. Provides a reference anchor to the definition tables of a particular package, class, enumeration or data type object in the XMI.

The syntax is as follows:

// For lutaml_figure
This is lutaml_figure::[package="Wrapper root package", name="Fig B1 Full model"] figure

// For lutaml_table
This is lutaml_table::[package="Wrapper root package"] package
This is lutaml_table::[package="Wrapper root package", class="my name"] class
This is lutaml_table::[package="Wrapper root package", enum="my name"] enumeration
This is lutaml_table::[package="Wrapper root package", data_type="my name"] data type

This code will be transformed into [figure-{diagram.xmi_id}] and will point to diagram figure. One can only use this macro when document rendered lutaml_uml_datamodel_description macro as it needs diagram lookup table in order to reference package diagram.

Will produce this output:

my text
== CityGML package
=== CityGML overview

Diagram text

[[figure-EAID_ACBB5EE3_3428_40f5_9C7C_E41923419F29]]
.CityGML Package Diagram
image::requirements/EAID_ACBB5EE3_3428_40f5_9C7C_E41923419F29.png[]

BuildingFurnitureFunctionValue is a code list that enumerates the different purposes of a BuildingFurniture.

[[figure-EAID_938AE961_1C57_4052_B964_997D1894A58D]]
.Use of ISO and OASIS standards in CityGML
image::requirements/EAID_938AE961_1C57_4052_B964_997D1894A58D.png[]

The CityGML package is organized into
2 packages with 1 modules:

. Another package
. CityTML package

my text

Content for CityGML package

==== Defining tables

.<<section-EAPK_9C96A88B_E98B_490b_8A9C_24AEDAC64293>> -- Elements of &#8220;Another::AbstractAtomicTimeseries&#8221; (class)

[[section-EAPK_9C96A88B_E98B_490b_8A9C_24AEDAC64293]]
.Elements of &#8220;Another::AbstractAtomicTimeseries&#8221; (class)
[width="100%",cols="a,a,a,a,a,a,a,a"]
|===
h|Name: 7+| AbstractAtomicTimeseries
h|Definition: 7+|
h|Stereotype: 7+| interface
h|Abstract: 7+|
h|Associations: 7+| (none)
.4+h|Public attributes:
| _Name_
2+| _Definition_
| _Derived_
| _Obligation_
| _Maximum occurrence_
| _Data type_
| adeOfAbstractAtomicTimeseries
2+|
|
| C
| *
| ADEOfAbstractAtomicTimeseries
| observationProperty
2+|
|
| M
| 1
| CharacterString
| uom
2+|
|
| C
| 1
| CharacterString
h|Constraints: 7+| (none)
|===

=== Additional Information

text after CityGML package

In addition to the XMI file, this macro also supports a YAML configuration file that specifies:

  • What packages to include in the render;

  • What render style is desired;

  • Location of the root package (which package should the iterative process start at).

The format for using the YAML configuration file:

---
packages:
  # includes these packages
  - "Package *"
  - two*
  - three
  # skips these packages
  - skip: four
render_style: entity_list | data_dictionary | default
section_depth: 2

Where:

  • packages - required, root element with the list of strings or objects

  • Package * - pattern matching, specifies lookup condition for packages to render.

    Note
    In this example, it is equal to the following regular expression: /^Package.*$/
  • skip: four - object with package name to skip

  • render_style - what template to use to render packages, can be one of:

    • entity_list

    • data_dictionary; or

    • default

  • section_depth - what package to use as root package for render. e.g., a section_depth equal to 2 tells the processor to use the first nested package of the first root packages in XMI file.

    EXAMPLE: If the XMI file has this package structure, and we have section_depth equal to 2, root package will be one-1.

    [
      {
        name: 'One',
        packages: [{ name: 'one-1' }, { name: 'one-2' }]
      },
      {
        name: 'Two',
        packages: [{ name: 'two-1' }, { name: 'two-2' }]
      }
    ]

Usage with macro:

[lutaml_uml_datamodel_description, path/to/example.xmi, path/to/config.yml]
----
[.diagram_include_block, base_path="models/Images", format="png"]
...
...
----

The processor will read the supplied YAML config file (path/to/config.yml), and iterate through packages according to the order supplied in the file. All packages that matches skip in the YAML config file will be skipped during render.

Usage of lutaml_ea_xmi macro

This command is a replacement for lutaml_uml_datamodel_description to perform the same functionalities of lutaml_uml_datamodel_description, which is to render data model packages and its dependent objects for supplied XMI file, by using Liquid Drop. The performance of lutaml_ea_xmi can be improved by 10~20 times. (Tested with a 10.6MB XMI file with 120000+ lines)

To use this macro, you only need to replace lutaml_uml_datamodel_description by lutaml_ea_xmi.

Replace:

[lutaml_uml_datamodel_description, path/to/example.xmi]
...

By:

[lutaml_ea_xmi, path/to/example.xmi]
...

You can define guidance in the configuration file as well. The configuration file will looks like:

packages:
  - my_package
guidance: "path/to/guidance.yaml"

Please refer to lutaml_klass_table section for how to configure the guidance yaml file.

Documentation

Please refer to https://www.metanorma.org.