Formalize and develop parser for EXPRESS MIM mapping

[ronaldtse]

MIM documents contain "reference path" mappings that relates to the EXPRESS schemas and their items.

Specifically:

Reference path: This section contains:

• the reference path to its supertypes in the common resources, for each MIM element created within this part of ISO 10303;
• the specification of the relationships between MIM elements, when the mapping of an application element requires to relate instances of several MIM entity data types. In such a case, each line in the reference path documents the role of a MIM element relative to the referring MIM element or to the next referred MIM element.

For example in ISO 10303-1001 (in the SMRL), it contains:

presentation_style_assignment
presentation_style_assignment.styles[i] ->
presentation_style_select

Is to be understood by these rules:

symbol	definition
->	the attribute, whose name precedes the -> symbol, references the entity or select type whose name follows the -> symbol;
[i]	the attribute, whose name precedes the [i] symbol, is an aggregate; any element of that aggregate is referred to;

So the element presentation_style_assignment.styles is an aggregate, and every item in presentation_style_assignment.styles[i] refers to the select type presentation_style_select.

The challenges are:

1. This mapping specification is not machine-readable. It is only read by humans.
2. We don't really know if the EXPRESS objects (e.g. presentation_style_select) referred exist or has changed based on the actual EXPRESS schemas.

What we need to do:

1. Formalize the language, based on the specification described in "5.1" below.
2. Create a parser to parse this language, link its nodes to Expressir, in order to perform validation with the actual EXPRESS schemas referred.

---

The full specification here.

5.1 Mapping specification

In the following, "Application element" designates any entity data type defined in Clause 4, any of its explicit attributes and any subtype constraint. "MIM element" designates any entity data type defined in Clause 5.2 or imported with a USE FROM statement, from another EXPRESS schema, any of its attributes and any subtype constraint defined in Clause 5.2 or imported with a USE FROM statement.

This clause contains the mapping specification that defines how each application element of this part of ISO 10303 (see Clause 4) maps to one or more MIM elements (see Clause 5.2).

The mapping for each application element is specified in a separate subclause below. The mapping specification of an attribute of an ARM entity is a subclause of the clause that contains the mapping specification of this entity. Each mapping specification subclause contains up to five elements.

Title: The clause title contains:

• the name of the considered ARM entity or subtype constraint, or
• the name of the considered ARM entity attribute when this attribute refers to a type that is not an entity data type or a SELECT type that contains or may contain entity data types, or
• a composite expression, <attribute name> to <referred type>, when this attribute refers to a type that is not an entity data type or a SELECT type that contains or may contain entity data types.

MIM element: This section contains, depending on the considered application element:

• the name of one or more MIM entity data types;
• the name of a MIM entity attribute, presented with the syntax ., when the considered ARM attribute refers to a type that is not an entity data type or a SELECT type that contains or may contain entity data types;
• the term PATH, when the considered ARM entity attribute refers to an entity data type or to a SELECT type that contains or may contain entity data types;
• the term IDENTICAL MAPPING, when both application objects involved in an application assertion map to the same instance of a MIM entity data type;
• the term NO MAPPING EXTENSION PROVIDED, when the extension to an extensible SELECT TYPE has no impact;
• the syntax /SUPERTYPE()/, when the considered ARM entity is mapped as its supertype;
• one or more constructs /SUBTYPE()/, when the mapping of the considered ARM entity is the union of the mapping of its subtypes.

When the mapping of an application element involves more than one MIM element, each of these MIM elements is presented on a separate line in the mapping specification, enclosed between parentheses or brackets.

Source: This section contains:

• the ISO standard number and part number in which the MIM element is defined, for those MIM elements that are defined in a common resource document;
• the ISO standard number and part number of this part of ISO 10303, for those MIM elements that are defined in the MIM schema of this part.

This section is omitted when the keywords PATH or IDENTICAL MAPPING or NO MAPPING EXTENSION PROVIDED are used in the MIM element section.

Rules: This section contains the name of one or more global rules that apply to the population of the MIM entity data types listed in the MIM element section or in the reference path. When no rule applies, this section is omitted.

A reference to a global rule may be followed by a reference to the subclause in which the rule is defined.

Constraint: This section contains the name of one or more subtype constraints that apply to the population of the MIM entity data types listed in the MIM element section or in the reference path. When no subtype constraint applies, this section is omitted.

A reference to a subtype constraint may be followed by a reference to the subclause in which the subtype constraint is defined.

Reference path: This section contains:

• the reference path to its supertypes in the common resources, for each MIM element created within this part of ISO 10303;
• the specification of the relationships between MIM elements, when the mapping of an application element requires to relate instances of several MIM entity data types. In such a case, each line in the reference path documents the role of a MIM element relative to the referring MIM element or to the next referred MIM element.

For the expression of reference paths and of the constraints between MIM elements, the following notational conventions apply:

symbol	definition
[]	enclosed section constrains multiple MIM elements or sections of the reference path are required to satisfy an information requirement;
()	enclosed section constrains multiple MIM elements or sections of the reference path are identified as alternatives within the mapping to satisfy an information requirement;
{}	enclosed section constrains the reference path to satisfy an information requirement;
<>	enclosed section constrains at one or more required reference path;
\|\|	enclosed section constrains the supertype entity;
->	the attribute, whose name precedes the -> symbol, references the entity or select type whose name follows the -> symbol;
<-	the entity or select type, whose name precedes the <- symbol, is referenced by the entity attribute whose name follows the <- symbol;
[i]	the attribute, whose name precedes the [i] symbol, is an aggregate; any element of that aggregate is referred to;
[n]	the attribute, whose name precedes the [n] symbol, is an ordered aggregate; member n of that aggregate is referred to;
=>	the entity, whose name precedes the => symbol, is a supertype of the entity whose name follows the => symbol;
<=	the entity, whose name precedes the <= symbol, is a subtype of the entity whose name follows the <= symbol;
=	the string, select, or enumeration type is constrained to a choice or value;
\	the reference path expression continues on the next line;
*	one or more instances of the relationship entity data type may be assembled in a relationship tree structure. The path between the relationship entity and the related entities, is enclosed with braces;
--	the text following is a comment or introduces a clause reference;
*>	the select or enumeration type, whose name precedes the *> symbol, is extended into the select or enumeration type whose name follows the *> symbol;
<*	the select or enumeration type, whose name precedes the <* symbol, is an extension of the select or enumeration type whose name follows the <* symbol;
!{}	section enclosed by {} indicates a negative constraint placed on the mapping.

[ronaldtse]

Ping @TRThurman the above is based on our last discussion at WG 12, please let me know if this is accurate. Thanks!

[TRThurman]

There is a missing optional element from the list above. /MAPPING_OF/ which is a predefined template.

See following: The definition and use of mapping templates are not supported in the present version of the application modules. However, use of predefined templates /MAPPING_OF/, /SUBTYPE/, and /SUPERTYPE/ is supported.

[TRThurman]

The newline character is part of the syntax. However, the use of newline to separate this declaration is not enforced by the XSL scripts.

This: presentation_style_assignment presentation_style_assignment.styles[i] -> presentation_style_select

might also be encoded as: presentation_style_assignment presentation_style_assignment.styles[i] -> presentation_style_select

[TRThurman]

mim elements are usually repeated in the reference path when an attribute is specified This usually appears ... presentation_style_assignment presentation_style_assignment.styles[i] -> presentation_style_select

---

This might appear presentation_style_assignment.styles[i] -> presentation_style_select

[TRThurman]

When PATH appears, any source entry shall be ignored. (A warning could be issued...)

[TRThurman]

The formal specification for mapping syntax and semantics is SC4N2661. If there are inconsistencies between that document and the above extract of clause 5.1, they need to be brought to the attention of WG12. SC4N2661 Guidelines for the development of mapping specifications.doc.zip

[TRThurman]

I could not find this document: STONIS, Alfonsas; Mapping syntax extensions, ISO TC 184/SC4/QC N203, 2001-06-11.

[TRThurman]

I found this document. qcn163.htm.pdf

[ronaldtse]

Thank you @TRThurman !

[ronaldtse]

@TRThurman specifications from SC4N2661 and qcn163.htm.pdf should probably be published as a standard in order to formalise the mapping language.