Dynamically Passing params to session

[cycomachead]

I'm working on an app where I'd like to be able to log custom parameters passed by the consumer.

However, the catch is that I won't know what those parameters are. These show as custom_*.

Essentially I'd like to do something like:

copy_launch_attributes_to_session = %w {
      custom_*
     lis_*
}

In this version, there'd need to be some defined type of syntax, or better yet, you could allow copy_launch_attributes_to_session to be an array including regex's which would call match for the keys.

Or possibly:

copy_lanch_attributes_to_session = -> (params) {
     params.select { |k, v| /x/.match k } 
}

Then you could really do whatever you'd like inside the lambda, which seems like a more flexible approach.

I'm fairly new to ruby, but I'd be happy to take a stab at implementing one option (probably the second).

Or is there a way to do this already? :)

[cycomachead]

I was able to do the second part by modifying DceLti::SessionHelpers::capture_attributes_from

def captured_attributes_from(tool_provider)
      attributes_to_copy = Engine.config.copy_launch_attributes_to_session
      if attributes_to_copy.respond_to?(:call)
        attributes_to_copy = attributes_to_copy.call(params)
      end
      attributes_to_copy.inject({}) do |attributes, att|
        attributes.merge(att => params[att]) #tool_provider.send(att)
      end
    end

So far things seem to work, but I'm unsure about replacing tool_provider.send. The problem is, the ims_lti module separates params on its own, so you can't directly access a custom param, because you need to go through tool_provider.custom_params[:custom_...]. It seems like tool_provider is mere aggregating, and not modifying the params here, so it doesn't look like this is causing issues but I'm not sure.

[djcp]

@cycomachead I am in a completely different problem space right now and can't context switch, but I will look through and think about this (probably tomorrow). Thanks for trying to push through it yourself!

[cycomachead]

Thanks for the reply! No rush though. :)

I realize my path breaks the current rspec tests, so I'm going to look into that, and I'll update this if I find anything.