Tepi is a test runner for .http files.
Write your tests in .http files and run them with tepi
.
$ cat test.http
GET http://localhost:3000 # fetch a GET Request
HTTP/1.1 200 OK # assert a 200 OK response
Content-Type: text/plain # assert a text/plain content type header
Hola mundo! # assert a body with the text "Hola mundo!"
$ tepi test.http
Test your HTTP APIs with standard http syntax
- π Write end to end API REST tests in
.http
files - π Validate Response status, headers and/or body.
- π₯ Interpolate javascript with eta template
<%= %>
eta url: - π Write metadata as frontmatter yaml
- π¦ Reference by id another test to run them in advance
- β± Set a timeout for each test or globally in milliseconds. After the timeout, the test will fail.
- π¨ Stop running tests after the first failure.
- π Use env files to load environment variables
- π Fully featured and colorful display modes. (none, minimal, default and full)
- π Watch files for changes and rerun tests.
- π― Standard Response and Request with a automatic getBody()
deno install --global --reload --allow-read --allow-env --allow-net --allow-run -f -n tepi https://tepi.deno.dev/src/cli.ts
Or run remotely with:
deno run --allow-read --allow-env --allow-net --allow-run https://tepi.deno.dev/src/cli.ts
--allow-read
Needed to read files from the file system.--allow-net
Needed to make HTTP requests.--allow-env
(optional) Needed to load and read environment variables. Required if you use the --env-file option.--allow-run
(optional) Required if you use the --upgrade option.
tepi [OPTIONS] [FILES|GLOBS...]
-w
--watch
Watch files for changes and rerun tests.--watch-no-clear
same but without clearing the screen.-t
--timeout
Set the timeout for each test in milliseconds. After the timeout, the test will fail.-f
--fail-fast
Stop running tests after the first failure.-d
--display
Set the display mode. (none, minimal, default, truncate, full and verbose) - none: display nothing - minimal: display only a minimal summary - default: list results and full error summary - truncate: list results and full error summary but truncate data - full: display also all HTTP requests and responses and not truncate data - verbose: display also all metadata-e
--env-file
load environment variables from a .env file--no-color
output without color--no-animation
output without terminal animations--upgrade
upgrade to the latest version--version
output the version number-h
--help
output help minimal information-r
--readme
output usage full information
tepi
Run all .http in the current directory and folders. (same as tepi ./**/*.http)
tepi test.http ./test2.http
Run test.http and test2.http
tepi **/*.http
Run all .http in the current directory and folders.
tepi rest.http --watch
Run rest.http and rerun when it changes
tepi rest.http --watch "src/**/*.ts"
Run rest.http and rerun when any .ts file in the src folder changes.
tepi rest.http --watch "src/**/*.json" --watch "src/**/*.ts"
You can use multiple --watch flags. Note: You can use globs here too, but use quotes to avoid the shell expanding them.
tepi --timeout 10000
Set the timeout for each test in milliseconds. After the timeout, the test will fail.
tepi --fail-fast
Stop running tests after the first failure.
tepi --display minimal
Set the display mode. (none, minimal, default and full)
tepi --env-file .env --env-file .env.test
Load environment variables from a .env and .env.test
- You can use the standard HTTP syntax in your .http files to run a request and response validation.
- Use the
###
to separate the requests. - Use
#
to comment. - Use front matter yaml to set metadata.
For example, validate the headers, status code, status text and body:
GET https://faker.deno.dev/?body=hola&status=400
HTTP/1.1 400 Bad Request
content-type: text/plain; charset=utf-8
hola
It's deno π₯
Uses eta as template engine, see docs: https://deno.land/x/eta
Use <%= %>
to interpolate values.
All the std assertion module is available: https://deno.land/std/testing/asserts.ts
Use <% %>
to run custom assertions or custom JS.
For example:
GET http://localhost:3000/users
<% assert(response.status === 200) %>
Or:
<% if (Math.random() > 0.5) { %>
GET http://localhost:3000/users/1
<% } else { %>
GET http://localhost:3000/users/2
<% } %>
In the Interpolation <%= %>
or <% %>
you have access to any Deno API and the following variables:
- request: The Request from the actual block.
- meta: The metadata from the actual block.
- response: The standard Response object from the fetch API from the actual request. (only available in the expected response, after the request)
- body: The extracted body an alias of
await response.getBody()
(only available in the expected response, after the request) - [id]: the id of a block already run for example:
<%= login.body.jwt %>
or<%= login.response.status %>
The Block signature is:
type Block = {
meta: {
[key: string]: any,
},
request?: Request,
response?: Response,
expectedResponse?: Response,
error?: Error,
body?: any,
}
The request, response and expectedResponse has a custom method async getBody()
to extract the body as json, text or blob depending on the content-type.
The body
is an alias for await response.getBody()
.
For example:
---
id: hello
---
GET https://faker.deno.dev/?body=hola
HTTP/1.1 200
hola
###
POST https://faker.deno.dev/
<%= hello.body %>
HTTP/1.1 200 OK
hola
You can add metadata to your tests using front matter yaml.
For example:
---
id: hello
description: This is a test
needs: login
---
GET https://faker.deno.dev/?body=hola
This metadata is available in the interpolation scope as meta
and in the meta
property of the block.
You can set global metadata in the first block of the file. This metadata will be available in all the blocks. This apllies to all the properties except meta.id
and meta.description
, meta.ignore
and meta.only
.
For example:
---
display: verbose
timeout: 1000
host: https://example.com
---
###
GET / # this request will have the host https://example.com and will display the verbose output and will timeout in 1000ms.
###
POST / # This too
There are some especial metadata keys used by tepi, as: meta.needs, meta.id, meta.description, meta.display, meta.timeout and meta.import
The meta.delay allows you to delay the execution of the request fetch for a specific time in milliseconds.
The meta.timeout allows you to override the global timeout for a specific test. If the request takes longer than the timeout, the test will fail. The delay is not included in the timeout.
The meta.needs is a special metadata value that allows you to run a test in advance and use the result in the current test if needed.
For example:
---
needs: login
# will run the login test before this one
---
GET https://example.com/onlyAdmin
Authorization: Bearer <%= login.body.jwt %>
Content-Type: application/json
###
---
id: login
---
POST https://example.com/login
Content-Type: application/json
{"user": "Garn", "password": "1234"}
HTTP/1.1 200 OK
The meta.id allows you to identify a test for reference. The meta.description it's used to display the test name in the console if not set, it will use the meta.id.
The meta.import allows you to import a file before running the test. The imported file will run before the file that imports it.
The meta.display allows you to override the global display mode for a specific test.
For example:
---
display: verbose
---
GET https://example.com/get
You can use the vscode extension to run the tests from the editor.
Just install the extension and open a .http file and click the hover button to run the test.
https://marketplace.visualstudio.com/items?itemName=jupegarnica.tepi