espr 0.3.0

A crate for implementing EXPRESS language compiler and related tools

Overview

┌────────────────┐
│ EXPRESS Schema │
└──┬─────────────┘
│ Tokenize
┌──▼─────────────────────────┐
│ Abstract Syntax Tree (AST) │
└──┬─────────────────────────┘
│ Legalize
┌──▼───────────────────────────────┐
│ Intermediate Representation (IR) │
└──┬───────────────────────────────┘ │
│ Code Generation                 │ espr's responsibility
┌──▼────────────────────────┐        ▼
│ Rust code with proc-macro │  ─────────────────────────────────
└──┬────────────────────────┘
│ proc-macro
┌──▼──────────────┐ call ┌──────────┐
│ Final Rust code ├──────► ruststep │
└─────────────────┘      └──────────┘

• Tokenize
• Read EXPRESS input and parse into abstract syntax tree (AST)
• [ast] module defines AST structs
• [parser] module defines parser combinator for parsing EXPRESS language
• Legalize
• Convert AST to IR (intermediate representation) to ready the following code generation
• [ir] module defines IR structs, and they implements [ir::Legalize] trait for legalizing from AST
• Code Generation
• [codegen::rust] module generates Rust code from IR

Introduction to STEP

As we describe in README, STEP consists of three components:

• Schema language called EXPRESS
• Data format called "exchange structure" or "STEP file"
• Schemas, e.g. AP203 for CAD components

This short section shows basic concept of EXPRESS schemas and exchange structures for developers new to STEP ecosystem.

EXPRESS language, standardized as ISO-10303-11, is a schema language as a part of STEP ecosystem.

SCHEMA my_first_schema;
ENTITY a;
x: REAL;
y: REAL;
END_ENTITY;

ENTITY b;
z: REAL;
a: a;
END_ENTITY;
END_SCHEMA;

This example schema written in EXPRESS language defines my_first_schema schema and two entities a and b. Entity is a collection of primitive types or other entities similar to struct in Rust.

Actual data will be stored in another data format called "exchange structure" or "STEP file", standardized as ISO-10303-21. Exchange structure consists of following sections:

• HEADER
• ANCHOR (optional)
• REFERENCE (optional)
• DATA
• SIGNATURE (optional)

A data section contains indexed instances:

DATA;
#1 = A(1.0, 2.0);
#2 = B(3.0, A((4.0, 5.0)));
#3 = B(6.0, #1);
ENDSEC;

This describes three instances, #1 of type A, and #2 and #3 of type B. Data section itself is independent from any schema. The grammar of exchange structure allows any number of fields, e.g. A(1.0, 2.0, 3.0). What data must be contained in the data section is specified in its header section:

HEADER;
FILE_SCHEMA(('my_first_schema'));
/* others */
ENDSEC;

The type A (B) corresponds to a (b) entity in schema. Programs reading this exchange structure should reject invalid data e.g. A(1.0, 2.0, 3.0) based on the schema specification, and this crate aims to help creating such programs easily.

Terms and definitions

These are excerpts from ISO 10303-11 3.3 "Other terms and definitions" to help understanding terms appears in this document.

Basic terms

Terms about entity

Advanced terms about entity

"data type" is ommitted here since "entity" and "entity data type" indicate same one in most situation.