Outputs revamp #921
Outputs revamp #921
Conversation
|
@lucassz I left |
| def normalize(x): | ||
| ans = [i.split('#') for i in x] | ||
| return ans | ||
|
|
There was a problem hiding this comment.
Do we still need these functions (normalize and denormalize)?
|
@lucassz do we still need the |
webapp/apps/core/views.py
Outdated
| url = OutputUrl.objects.get(pk=pk) | ||
| except OutputUrl.DoesNotExist: | ||
| raise Http404 | ||
|
|
There was a problem hiding this comment.
@lucassz what do you think about doing the URL routing through the new CoreRun table using the uuid variable?
webapp/apps/core/views.py
Outdated
| return render( | ||
| request, | ||
| 'taxbrain/not_avail.html', | ||
| not_avail_context) |
There was a problem hiding this comment.
This logic is aimed at data with a backwards incompatible format. Ideally, we won't need to accommodate backwards incompatible data. I think it's safe to remove it. Hopefully, we won't need to add it back.
webapp/apps/core/views.py
Outdated
| error.text = error_contents | ||
| error.save() | ||
| model.error_text = error | ||
| model.save() |
There was a problem hiding this comment.
@lucassz what do you think about moving the error messages into the CoreRun table?
|
@hdoupe I only temporarily kept I will also remove the I agree with moving the URL routing and error storage into |
Ok, I see that makes sense. Nope, I'm not sure how to do that.
Great, thanks. |
|
All tests are passing in PR #921. @lucassz thank you for your hard work in developing and implementing the new approach to displaying outputs. @martinholmer thank you for your feedback as this feature was developed. The changes made in this PR open up many possibilities for how outputs can be rendered in the future. Further, many patterns developed here such as using a All of the TaxBrain apps (static, behavioral, and macro-elasticity) are now running with the new outputs implementation. Once it has been polished up, there should not be any noticeable difference in how the outputs look. This fall, I plan to transition the Cost-of-Capital calculator to this new approach. A lot of work is still required to clean up after this major change. There needs to be some of editorial work to indicate that the behavioral app parameters are now on the main TaxBrain page and that the macro elasticity app is available if no behavioral parameters are specified on the main taxbrain page. A comprehensive solution will be applied to this problem once the outputs work started in this PR is complete. I don't anticipate finishing up this work taking more than a week or two. |
Implements the prototype in #916. A Tax-Calculator PR will be opened and worked on in tandem with this PR in order to prepare Tax-Calculator for the new outputs standards.
Goals:
compute.pylogic corresponding to @lucassz's work in Initial commit pb_deploy#1 *core*** Once this is merged, there will only be one AWS endpoint per worker node cluster to query instead of one for each worker node. Thus, there is no need to implement our own round robin logic--this is done by celery. When a job is posted, it is placed on celery's task queue (which is run via redis) and celery pulls those jobs in a LIFO manner. I think we should implement this simplifying
compute.pylogic and the ability to display the new tables simultaneously because these sections of the code base are tightly coupled. This means that changing one requires significant changes to the other. So, changing them at the same time will be a lot of work but not much more work than changing them sequentially. I'm happy to discuss further if others have differing opinions.** To accomplish the goal of turning PolicyBrain into a platform (#906), more code will need to be shared between apps. Moving code that can be used by multiple apps into this new
coreapp is a step in that direction. Also, see #816.