Skip to content

Latest commit

 

History

History
50 lines (30 loc) · 2.74 KB

readme.md

File metadata and controls

50 lines (30 loc) · 2.74 KB
Raw

AutoRest: Literate file-format support

Purpose

To provide a means to tie together all related information for a given set of OpenAPI specifications including information regarding:

  • version of AutoRest to use.
  • global and code-generator specific settings
  • hints, overrides and information related to code generation that does not affect the wire-protocol
  • references to additional supplemental resources like test data, examples, and other documentation
  • metadata to track the publishing/release state of a given SDK

In a singular file, instead of passed on the command line and/or embedded in the OpenAPI input file, or batch files.

The design of the file formats chosen here promotes a documentation-centric model for both settings and OpenAPI spec authoring.

File Format

The 'Literate' file format is CommonMark document that has embedded code-blocks for machine-readable sections. The this encourages easy-to-author documentation while permitting the specificity desired when authoring and processing instructions.

Generally speaking, it is preferable that the documentation section that follows a given code-block applies to that code-block.

Scenarios to consider

Using AutoRest with Literate files is as simple as possible without the user having to understand convoluted mechanisms in order to do day-to-day work with the tool.

The usage is consistent regardless of how AutoRest is used (ie, from the command line, as a language service in VSCode, or in a CI/CD system.)

This includes authoring and execution support for using AutoRest

  • on a single OpenAPI file, with a specific set of settings
  • on a collection of related OpenAPI files, (like the now-defunct composite-modeler), but allow for individually generating SDKs, or merge them based into one.
  • on OpenAPI files, in parallel for a given set of language generators

Implementation

Since AutoRest has moved towards a pluggable extension model where generators, validators and transformer can be implemented out-of-proc in any language, the intention is that AutoRest implement the support to handle the Literate file formats, and the individual extensions can simply request the processed object model, and not have to do any coding to process or parse the CommonMark file formats.

Compatibility

AutoRest can transparently process a JSON file, a YAML file or a Literate document for any purpose that is necessary, and an author can substitute one for another if it contains the information required.

AutoRest can also be used to 'preprocess' a given file into a JSON or YAML for consumption by other tools that are not able to process a Literate file.

Specifications

The specification for the Literate Configuration