roshiro / openws_one Goto Github PK

Today the API returns an empty body and code 201.

Problems:

• How to apply GraphQL in schema-less collections?

The reason in migrating to heroku is because OpenShift doesn't allow Ruby + NodeJS running in the same instance.

Getting the JSON below when doing a POST:

{
    "msg": "not authorized for insert on openws.employees (16544)"
}

In the server:

[ 2016-10-19 15:11:51.1750 189390/7f7dc8384700 Pool2/Implementation.cpp:1274 ]: [App 204200 stdout] D, [2016-10-19T15:11:51.174906 #204200] DEBUG -- : MONGODB | 580787c47628e105a700006c-roshiro.rhcloud.com:37941 | openws.insert | FAILED | not authorized for insert on openws.employees (16544) | 0.003748378s

Investigate what are the best practices and what's the friendlier way to search documents passing more complex parameters.

Fix typo in ... to handle your data when prototyping you app anymore

Should be your instead of you.

Needs to convert from BSON id style:

{
  "_id": {"$oid":"52d3fdfd53656e5180020000"},
  "changed_at":"2014-02-15",
  "created_at":"2014-01-13T15:53:49.444+01:00",
  "updated_at":"2014-01-13T15:53:49.444+01:00"
}

to:

{
  "id": "52d3fdfd53656e5180020000",
  "changed_at":"2014-02-15",
  "created_at":"2014-01-13T15:53:49.444+01:00",
  "updated_at":"2014-01-13T15:53:49.444+01:00"
}

Solution:
According to this blog, we can user initialisers to achieve that:

# config/initializers/mongoid.rb
module Mongoid
  module Document
    def as_json(options={})
      attrs = super(options)
      attrs["id"] = attrs["_id"].to_s
      attrs
    end
  end
end

Landing page must:

• be simple;
• be clean;
• straight to the point;

Target audience:

• Developers prototyping web & mobile apps;

Message to convey:

• For testing;
• Save JSON in one step;
• No signup;
• Free;

Copy:
"Free, Open Source, Public JSON storage. Perfect for prototyping web & mobile apps"
"Save any JSON you want"
"No authentication, no signup, no credit card needed."

Resources:

• Get colour pattern ideas here

Widget with options to:
save

• textbox to type collection name (suggest one name);
• textarea with JSON to be saved;
• button to save;

get

delete

Add instructions on how to setup locally

Info on testing http://blog.hayleyanderson.us/2015/06/05/using-and-testing-rack-attack-to-improve-the-security-of-your-rails-app/

Still needs to decide on best storage. Memcached? Redis?

The idea here is only to setup the caching database, not cache responses for performance yet.
Rack::Attack depends on caching db.

Decide on what's the best approach to clean up old documents.

• Delayed jobs?
• Clean up by data created?

At first throttle by IP address, allowing each IP address X number of requests per minute;

• add localhost in safelist

• collection name size (beware namespace takes ~ 123 chars)
• Sanitise invalid characters
• Good article on the limitations of MongoDB http://tech.tulentsev.com/2014/02/limitations-of-mongodb/

An alternative would be to return:

{
  "total": "158",
  "pages": "4",
  "page": 1,
  "items": [{...}]
}

So the API should take page query string.

Applications CRUD