Exclude fields using config

[jonathon-m]

Hey,

I'd like to be able to mark fields for exclusion from the generated schema.
Some of my models have sensitive data which is not sent to end users via my API. Adding excludeFromJSONSchema: true or something similar in my model definitions would cover this use case 👍

If you can provide a hint about how you'd want this implemented I'm happy to make a pull request.

Cheers!

[DScheglov]

Hey, @jonathon-m

Interesting idea. But I suggest don't do it in so direct way.
The first reason is that your api can have a several consumers some of them could have an access to the sensitive data.

The Library allows to build schema for each single query:

• https://github.com/DScheglov/mongoose-schema-jsonschema#mongoosemodeljsonschema
• https://github.com/DScheglov/mongoose-schema-jsonschema#mongoosequeryprototypejsonschema

See examples:

• https://github.com/DScheglov/mongoose-schema-jsonschema/blob/master/samples/2.js (model schema)
• https://github.com/DScheglov/mongoose-schema-jsonschema/blob/master/samples/3.js (query schema)

In many cases it is more correct approach to generate a schema.

The second reason is that "to exclude from the JSONSchema" is an effect, but is not a cause. So it is more correct approach to mark sensitive fields with something like: private: true.

So, what do you think?

[jonathon-m]

Hey @DScheglov thanks for the response 😄
The query schema builder doesn't work for my use case, unfortunately. I'm using the library to generate OpenAPI documentation and the relevant queries are only available at runtime.

I agree with your approach of private: true over an explicit exclude field. Would it make sense to have a a config option to toggle the behavior of excluding private fields (similar to the populate parameter to jsonSchema())?

[DScheglov]

@jonathon-m

I'm going to add required functionality with some "post-processing" utilities.

const { Schema } = require('mongoose-schema-jsonschema')(require('mongoose'));
const { postProcess } = require('mongoose-schema-jsonschema/post-process');

const BookSchema = new Schema({
  title: { type: String, required: true }, 
  year: Number,
  author: { type: Schema.Types.ObjectId, required: true, ref: 'Person' },
});

const jsonSchema = postProcess(BookSchema.jsonSchema())
  .exclude('title', 'author')
  .select('year')
  .map((path, schema) => replaceSchema(path, schema))
  .done();

// also will be available the curried api:
// see: https://babeljs.io/docs/en/babel-plugin-proposal-pipeline-operator

const { select, exclude, map } = require('mongoose-schema-jsonschema/post-process');

const jsonSchema2 = BookSchema.jsonSchema()
  |> exclude('title', 'author')
  |> select('year')
  |> map((path, schema) => replaceSchema(path, schema));

Have any thoughts around that?

[jonathon-m]

This is perfect.
To add more context, I'm building a plugin for serverless that builds documentation from mongoose schemas & api templates. A post processor will add heaps of useful flexibility!