* Initial stab at github actions * Fixed docker images * Added run * old new * update * Updated * Fixed (?) * Added certs * Proper safe head * Use proper API * Added project name * Updated docker image * Started reading the right comments * Added docker image to CI * Renamed CI thing * Increased swap * Fixed tests * Fixed docker hub * Added conditionals back in * minor refinements |
||
---|---|---|
.github/workflows | ||
app | ||
awsm-css@ad03dcf22f | ||
benchmarks | ||
github-action | ||
nix | ||
src | ||
test | ||
.gitignore | ||
.gitmodules | ||
.hlint.yaml | ||
action.yaml | ||
CHANGELOG.md | ||
default.nix | ||
fourmolu.yaml | ||
LICENSE | ||
openapi-diff.cabal | ||
README.md | ||
stack.yaml | ||
stack.yaml.lock |
openapi-diff
Compatibility checker for OpenAPI
Quick Start Guide
Your situation
You are developing a very important server with a REST API. You have clients who use your API that you do not control. Say, you are also developing a mobile app that uses your API and you can't force someone to update to the latest version. (Or you prefer not to for UX reasons.)
You have recently released version 1.0 and things are going great: user are downloading your app, servers are processing requests.
You describe your API in a file api-1.0.0.yaml
:
openapi: "3.0.0"
info:
version: 1.0.0
title: Swagger Petstore
license:
name: MIT
servers:
- url: https://example.com
paths:
/pets:
get:
parameters:
- name: limit
in: query
required: false
schema:
type: integer
maximum: 20
responses:
"200":
description: ""
headers:
x-next:
schema:
type: string
content:
application/json:
schema:
$ref: "#/components/schemas/Pets"
post:
requestBody:
content:
application/json:
schema:
$ref: "#/components/schemas/Pet"
responses:
"201":
description: ""
components:
schemas:
Pet:
type: object
required:
- id
- name
properties:
id:
type: integer
name:
type: string
minLength: 3
maxLength: 10
Pets:
type: array
items:
$ref: "#/components/schemas/Pet"
Evolving your product
Enthused over your initial success you hurry to release a new and improved version of your API and mobile app.
After a round of very intense programming you take a look at your new api-1.1.0.yaml
:
openapi: "3.0.0"
info:
version: 1.1.0
title: Swagger Petstore
license:
name: MIT
servers:
- url: https://example.com
paths:
/pets:
get:
parameters:
- name: limit
in: query
required: false
schema:
type: integer
maximum: 30
responses:
"200":
description: ""
headers:
x-next:
schema:
type: string
content:
application/json:
schema:
$ref: "#/components/schemas/Pets"
post:
requestBody:
content:
application/json:
schema:
$ref: "#/components/schemas/Pet"
responses:
"201":
description: ""
components:
schemas:
Pet:
type: object
required:
- id
- name
properties:
id:
type: integer
name:
type: string
minLength: 1
maxLength: 15
weight:
type: integer
Pets:
type: array
items:
$ref: "#/components/schemas/Pet"
Looking at the very large and complex API description, you grow more and more concerned that your old mobile app might stop working when you update the server. But the spec is too large and too complex to reasonably assess this manually.
Assessing compatibility automatically
Luckily, you have access to compaREST which can programmatically analyze your APIs and determine what, if anything, breaks compatibility and what doesn't.
You can call it, passing the API your client will be aware of, and the API your server will serve like so:
docker run --rm -v $(pwd):/data:rw typeable/comparest --client /data/api-1.0.0.yaml --server /data/api-1.1.0.yaml --output /data/report.md
Running this command will output a file report.md
, containing the compatibility report between the two APIs:
Summary
⚠️ Breaking changes 🙆 Non-breaking changes 🤷 Unsupported feature changes 5 6 0 ⚠️ Breaking changes
GET /pets
📱⬅️ JSON Response – 200
$[*].name(String)
Maximum length of the string changed from 10 to 15.
Minimum length of the string changed from 3 to 1.
POST /pets
📱➡️ JSON Request
$.weight
Values are now limited to the following types:
- Number
The property was previously implicitly described by the catch-all "additional properties" case. It is now explicitly defined.
$.weight(Number)
Value is now a multiple of 1.0.
🙆 Non-breaking changes
GET /pets
Parameter limit
JSON Schema
$(Number)
Upper bound changed from 20.0 inclusive to 30.0 inclusive.
📱⬅️ JSON Response – 200
$[*].weight
Values are now limited to the following types:
- Number
The property was previously implicitly described by the catch-all "additional properties" case. It is now explicitly defined.
$[*].weight(Number)
Value is now a multiple of 1.0.
POST /pets
📱➡️ JSON Request
$.name(String)
Maximum length of the string changed from 10 to 15.
Minimum length of the string changed from 3 to 1.
You now know exactly in what situations and in what way your 1.0 version of the app will break if you deploy your 1.1 version of the server.
Additional formats
You can also produce a self-contained HTML report that you can open in your browser by simply omitting the file extension of the output file:
docker run --rm -v $(pwd):/data:rw typeable/comparest --client /data/api-1.0.0.yaml --server /data/api-1.1.0.yaml --output /data/report
CLI docs
Usage: openapi-diff (-c|--client ARG) (-s|--server ARG)
[--silent | --only-breaking | --all] [-o|--output ARG]
[--folding-block-quotes-style | --header-style]
[--signal-exit-code]
A tool to check compatibility between two OpenApi specifications.
Usage examples
Compare files old.yaml with new.yaml and output the resulting report to
stdout:
openapi-diff -c old.yaml -s new.yaml
Only output breaking changes and write a styled HTML report to file
report.html:
openapi-diff -c old.yaml -s new.yaml --only-breaking -o report
Don't output anything, only fail if there are breaking changes:
openapi-diff -c old.json -s new.json --silent
Write full report suitable for embedding into a GitHub comment to
report.html:
openapi-diff -c old.json -s new.json --folding-block-quotes-style -o report.html
Available options:
-h,--help Show this help text
-c,--client ARG A path to the file containing the specification that
will be used for the client of the API. Can be either
a YAML or JSON file.
-s,--server ARG A path to the file containing the specification that
will be used for the server of the API. Can be either
a YAML or JSON file.
--silent Silence all output. Only makes sense in combination
with --signal-exit-code.
--only-breaking Only report breaking changes in the output.
--all Report both incompatible and compatible changes.
Compatible changes will not trigger a failure exit
code.
-o,--output ARG The file path where the output should be written. If
the option is omitted the result will be written to
stdout.
The file extension is used to determine the type of
the output file.
Supports many formats such as markdown, html, rtf,
doc, txt, rst, and many more.
Leave out the extension to produce a self-contained
HTML report with styling.
--folding-block-quotes-style
The report tree is structured using summary/detail
HTML elements and indented using block quotes. This
style renders well on GitHub.Intended for HTML output
format. Markdown has rendering bugs on GitHub.
--header-style The report tree is structured using increasing levels
of headers.
--signal-exit-code Signal API compatibility with the exit code.
Exit with 0 if there are no breaking changes.
Exit with 1 if there are breaking changes.
Exit with 2 if could not determine compatibility.