feathers generate service vs sequelize model:create

[mbt-olivier-o]

In the context of using sequelize with feathers:
to create a model you can run

 sequelize model:create --name post --attributes title:string,content:text

a model file is created accordingly in the migration folder specified by a .sequelizerc

const path = require('path');

module.exports = {
  'config': path.resolve('migrations/config/config.js'),
  'migrations-path': path.resolve('migrations'),
  'seeders-path': path.resolve('migrations/seeders'),
  'models-path': path.resolve('migrations/models')
};

in this example you get : /migrations/models/post

Once the model is generated next step is to create a service

feathersjs generate service

another model is created under /src/models/post.model.js

Ideally, it would be nice when the generator is running it detects that a model is already there and either make a require of it or copy the table creation part and add any feathers specific.

[DesignByOnyx]

You are mixing and matching generators. You are also using the migrations code from the feathers docs (I know, I wrote it ;)). You should not be using the sequelize generators for anything but migrations, and you should not be creating models in the migrations folder.

I recommend not mixing generators. Here's why:

• Generators are merely a convenience - they're not perfect, and you are supposed to change the generated code to suit your needs. Generators cannot (and should not) check for code/files created by other generators.
• Feathers models should all be stored in one place: src/models
• The models path used by the migrations is path.resolve('migrations/models'). Have you looked in that folder? There shouldn't be any models in that folder. It loads all of the models from your application.
• The feathers generators will generate a model for you and register it with your app. Your migrations script will automatically see this without you needing to do anything.

[mbt-olivier-o]

@DesignByOnyx , thanks for the clarification.
My goal is:

• to create a migration file to be able to run any database update independently from the feathersjs server process
• have a model file generated by feathers generate service command that matches the model described in the migration

According to the sequelize doc it doesn't seem to be possible to create a migration without generating a model. I'm trying to come up with a best practice solution.
Should one run

sequelize model:create --name post --attributes title:string,content:text 

then delete the model file generated by sequelize and then run the feathers generate service to get the model?
or is there a better way to achieve the described goal.

[daffl]

Sounds about right to me but maybe @DesignByOnyx can weight in again. I'm going to close this issue though since the generator will stay the way it is for the moment.

[mkozjak]

Maybe I'm missing something, but when we create a service via feathersjs' cli, we only get a text column in the database. Can we define columns when using the generator?

[daffl]

The generator is supposed to get you set up quickly but other than the initial model template and connection setup it will not do anything specific to the ORM/database you chose. In most ORMs adding columns is a line or two of code.

[mkozjak]

Ok, makes sense!

…

On 18 Dec 2017, at 01:53, David Luecke ***@***.***> wrote: The generator is supposed to get you set up quickly but other than the initial model template and connection setup it will not do anything specific to the ORM/database you chose. In most ORMs adding columns is a line or two of code. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[huesoamz]

But there is a way to to import all columns to the model? instead of only gettin a "text" column? I have a lot of tables with a lots of columns, there will be a mess write all on hand, instead of having a cli command like loopback have as "discover" there is something similar in feathersjs?