Develop Reference

Generate Joi Schema from Typescript types/interfaces

I want to generate some joi schema object from Typescript types or interfaces. In my initial searching I found some things that do the opposite (generate Typescript types/interfaces from joi schemas), and ts-interface-builder + ts-interface-checker that offer some ability to create runtime checkers based on Typescript types/interfaces but were still lacking in feature support, and a whole bunch of gnarly things using classes and decorators around props/methods to accomplish this.
Is there something out there to generate such joi schemas? Or a more mature alternative for runtime checking of interfaces/types? (Useful when pulling from a database and ensuring that the response from a DB is in the correct structure)
EDIT: I guess there's a pretty good thread about this kind of problem on this io-ts github issue.

I found something that works to my satisfaction for this: typescript-is.
Swap out what typescript compiler you're using to ttypescript. Instead of calling tsc, call ttsc.
Add the following to tsconfig.json:
{
"compilerOptions": {
// Rest of ts config options...
"plugins": [
{
"transform": "typescript-is/lib/transform-inline/transformer"
}
]
}
}
$ npm install ttypescript typescript-is
Now you can add the following to your code:
import { is, assertType } from 'typescript-is'
interface SomeInterfaceOrType {
aString: string
}
// To do a simple boolean check, do the following:
const nonConformingObj = { somethingOutOfLeftField: 42 }
const conforms = is<SomeInterfaceOrType>(nonConformingObj) // false
// Or to get better details on what's wrong, do the following
const anotherNonConformer: unknown = { aString: 1337 }
try {
assertType<SomeInterfaceOrType>(anotherNonConformer)
} catch(error) {
console.log(error.message) // logs: "validation failed at anotherNonConformer.aString: expected a string"
}
At compile time, the is<T>() and assertType<T>() calls will be transformed into working runtime checks. Not necessarily a JOI schema, but still pretty neat, and definitely good enough for my use case.

How do I correctly configure mocha tests with ts_transformer_keys?

I can't seem to set the custom transformer for ts-transform-keys with my mocha tests.
I’m using mocha 6.1.4
ts-node 8.3.0 https://www.npmjs.com/package/ts-node
ts-trasnformer-keys 0.3.5 https://github.com/kimamula/ts-transformer-keys
ttypescript 1.5.7 https://github.com/cevek/ttypescript
The ts-node documentation says that you cannot set a custom transformer on the CLI, only programatically. So I'm trying to use ttypescript to get around that restriction.
I've tried the following...
Note: test.ts contains the following
import { keys } from 'ts-transformer-keys';
describe("xyz"), () => {
it("123", (done) => {
keys<CustomInterface>();
});
});
Attempt 1) - Set the ts-node with an environment variable
TS_NODE_COMPILER="ttypescript" mocha test/test.ts --require ts-node/register
Then I have the following in test/tsconfig.json
{
"compilerOptions": {
"plugins": [
{ "transform": "ts-transformer-keys/transformer" }
]
}
}
This results in Uncaught TypeError: ts_transformer_keys_1.keys is not a function which indicates that the custom transformer wasn't used at compile time.
Attempt 2) Following the typescript API example from ts-transformer-keys
I added a mocha.opts file with the following
--file test/transformer-config.js
and a transformer-config.js file with the following
const ts = require('typescript');
const keysTransformer = require('ts-transformer-keys/transformer').default;
const program = ts.createProgram(['test/test.ts'], {
strict: true,
noEmitOnError: true,
target: ts.ScriptTarget.ES5
});
const transformers = {
before: [keysTransformer(program)],
after: []
};
const { emitSkipped, diagnostics } = program.emit(undefined, undefined, undefined, false, transformers);
if (emitSkipped) {
throw new Error(diagnostics.map(diagnostic => diagnostic.messageText).join('\n'));
}
Then I invoke it like this mocha test/test.ts --require ts-node/register
This results in the following error
/Users/jjohnson/Documents/OCS/hmp/git/hmp-server/server/test/ttypescript-register.js:17
throw new Error(diagnostics.map(diagnostic => diagnostic.messageText).join('\n'));
^
Error: [object Object]
[object Object]
[object Object]
at Object.<anonymous> (/Users/jjohnson/Documents/OCS/hmp/git/hmp-server/server/test/ttypescript-register.js:17:9)
at Module._compile (internal/modules/cjs/loader.js:777:30)
...
It feels like in Attempt 1 it wasn't ever calling the code that sets the custom transformer in tsconfig.json or if it was getting called the code was failing silently.
It feels like in Attempt 2 I'm creating a new instance of the typescript program and then that fails for some reason. And even if it succeeded I'm not sure that this is the right way to go about configuring things since the ts.createProgram wants a list of RootNames for the files it will transpile.
Maybe my entire approach is wrong.
All I really want is a way that in my mocha tests I can verify that the expected result type is what the method returned. And I'd like to be able to do this w/out touching too much of the source code.

you should be able to define your required module (see below) and run ts-node programmatically. In this way, you can safely use any customer transformer.
// tsnode.js
const transformer = require('ts-transformer-keys/transformer').default;
require("ts-node").register({
transformers: program => ({
before: [
transformer(program)
]
})
});
then you can run mocha with require
mocha --require './tsnode.js' --watch-extensions ts,tsx "test/**/*.{ts,tsx}

You can tell ts-node which compiler to use in tsconfig.json. This is covered in the ts-node docs. If your using transformers presumably your also using ttypescript compiler. You just need to add this:
"ts-node": {
"compiler": "ttypescript"
}

how to fix error for "Expected an assignment or function call and instead saw an expression" [duplicate]

In my Chai tests I often find myself wanting to use their assertions that are something like .to.be.empty, .to.be.true e.t.c., because I find them to be cleaner to read than .to.be.length(1) or .to.be.equal(true). However, this breaks my linter (I'm using default Airbnb linting).
I could use the // disable-eslint-line syntax, but then I'd have to add it to every single line that reads like that and that seems tedious.
I've also read about the DirtyChai library, but that would require me to go back through my entire testing library adding brackets to them all which seems like something I shouldn't have to do simply to get my linter to pass something it should probably be OK with in the first place.
Does anyone know a nicer way to handle this than the ways I've outlined above?

You can disable the rule for the entire file using eslint-disable at the top of the file in question:
/* eslint-disable no-unused-expressions */
expect(someTrueValue).to.be.true;
However, adding this at the top of every test file can be tedious. To disable this rule for all relevant files, you can:
Put a new .eslintc configuration file in the same directory as your test files, configured to disable that rule. This allows you to use the default configuration for all other rules while ignoring that rule specifically only on files in that folder. ESLint calls this Configuration Cascading.
{
"rules": {
"no-unused-expressions": "off"
}
}
Use the overrides key in your main .eslintrc file to disable rules for groups of files with glob pattern matching:
{
"overrides": [
{
"files": ["*.test.js", "*.spec.js"],
"rules": {
"no-unused-expressions": "off"
}
}
]
}
This also allows you to disable other rules which become troublesome in testing, such as no-underscore-dangle when using rewire.

Just found another option using Relative Glob Patterns:
In your .eslintrc file:
"overrides": [
{
"files": "*.test.js",
"rules": {
"no-unused-expressions": "off"
}
}
]

I've made a small plugin called eslint-plugin-chai-friendly that overrides the default no-unused-expressions rule and makes it friendly towards chai. The modified rule ignores the expect and should statements while keeping default behavior for everything else.

Combining jonalvarezz's answer with Ihor Diachenko's answer gave me exactly what I wanted:
npm install --save-dev eslint-plugin-chai-friendly
// .eslintrc.js
module.exports = {
// ...
plugins: ['chai-friendly'],
overrides: [{
files: '*.test.js',
rules: {
'no-unused-expressions': 'off',
'chai-friendly/no-unused-expressions': 'error',
},
}],
// ...
}
This way, the no-unused-expression rule will only be overridden in *.test.js files
AND
a no-unused-expression rule will still be in place to catch any unused expressions in the test files that are unrelated to chai.

In case anyone is stumbling upon this today, I had the same issue and found this solution on eslint documentation. In your eslint configuration file, you can specify one or several environments, which will predefine global variables for this environment. For us, it'd be mocha, and you'd configure like this in your .eslintrc.json:
{
"env": {
"mocha": true
},
...
...
...
}
As a result, it will remove all false positive about mocha describe, it, beforeEach, etc. without needing to completely disable eslint or completely disable any specific rule.
Tested with ESLint v.4.11 and mocha 5.0

I had this issue with tslint and solved it by simply moving the rule for unused expressions down one level. My ./tslint.json has all the other rules I care about, then I made ./src/tslint.json that just looks like
{
"rules": {
"no-unused-expression": true
},
"extends": "../tslint.json"
}
tslint automatically checks for a config file in every level as it descends the tree (with --project or using the VSCode extension) so this means that my tests (under ./test/) have all the other rules applied, but no-unused-expression only applies to files under ./src/.

using flowtype to statically check mocha test code

I have some complex Mocha code which I would like to statically check with FlowType because why not?
Below is a minimal repro:
/* #flow */
describe('it', function () {
it('fails', function() {
const s: number = 'flow spots this error';
});
});
When I run Flow on this, Flow does indeed spot the problem with the assignment of string to number which shows that the approach is working to some extend.
However, I also get:
test/test.js:4
4: describe('it', function () {
^^^^^^^^ identifier `describe`. Could not resolve name
test/test.js:5
5: it('fails', function() {
^^ identifier `it`. Could not resolve name
… apparently the Mocha test definitions run in an environment where these functions are globally available but looking at the test file there's nothing that would allow Flow to detect that.
I am not sure these problems are specific to Mocha but I don't feel I can confidently frame the question in broader terms, so my questions are:
how can I have Flow type check Mocha test code without suppressing every line that contains describe or it ?
is this is an instance of a broader class of situations and, if so, what would the latter be?

Third-party libraries usually need definition files, i.e. files containing all the type information for a given library.
In this case, you need a definition file for mocha, which fortunately is provided by flow-typed.
Install it with
npm install -g flow-typed
then run
flow-typed install
It will automatically install all the available definition files for your dependencies, including mocha.

You can simply declare the flow describe, it variables.
/* #flow */
declare var describe: any;
declare var it: any;
describe('it', function () {
it('fails', function() {
const s: number = 'flow spots this error';
});
});

xo lint error: `document` is not defined

I've been a long time user of Standard, and now that I'm working on a new project, I've been asked to start writing semicolons.
I'm trying to use both xo, Babel and React, but I keep getting an error when I try to lint my code:
document is not defined. no-undef
I've tried adding an env option to the xo field in my package.json file, but no success.
My xo config:
"xo": {
"esnext": true,
"extends": "xo-react",
"space": true,
"rules": {
"react/jsx-space-before-closing": 0
}
}

It is cumbersome to specify linting options such as /** global document **/ and edit a configuration file every time you use a global.
This error can be suppressed by using --env=browser option:
xo --env=browser [<file|glob> ...]
Note: Same problem comes with Node.js, where the linter will complain that require and friends are not defined. The switch has to change to --env=node in that case.
However, XO defaults the env to node, therefore this will not be a problem in most cases. You will need multiple environments if your project contains both client and server files. in that case, --env switch can be set multiple times:
xo --env=browser --env=node [<file|glob> ...]

You have to define globals in ESLint. There are two ways to accomplish this, firstly as a comment in your code:
/* global document */
Or you can configure in configuration file like so:
{
"globals": {
"var1": true,
"var2": false
}
}
See the ESLint docs for more