search

CRSU-Apps / MetaInsight

An interactive web-based tool for analyzing, interrogating, and visualizing network meta-analyses using R-shiny
https://crsu.shinyapps.io/MetaInsight/
GNU General Public License v3.0
8 stars 6 forks source link

Mr 33 tab 4b #128

Closed tommorris168 closed 7 months ago

tommorris168 commented 7 months ago

There is currently no reusing of ui that already exists from the gemtc models. covariate_run_model_ui() could be reused with minor adaptations. I'm just submitting this for your review as it is.

I have added a warning message in case the model doesn't converge. Without the message the only indication that that has happened is that it takes longer than normal to run. The default data set with exchangeable covariate parameters demonstrates it.

The forest plot in bnma is not displaying very well. I have removed the ability to change the x-axis limits, since this option is not available in bnma::network.forest.plot. It does display well in the download.

I have added four functions to DataWrangling.R, three of which were added in the contribution matrix branch. None of them are actually necessary until the novel graph is added, so please ignore these for now.

JanionNevill commented 7 months ago

I was hoping that you'd keep the structure the same as the covariate tab for now, with the forest plot in the first tab. My plan was to rearrange both tabs at the same time once they were created.

tommorris168 commented 7 months ago

I was hoping that you'd keep the structure the same as the covariate tab for now, with the forest plot in the first tab. My plan was to rearrange both tabs at the same time once they were created.

I can change it to this, although on Jira you did say to put the forest plot on the second tab...

tommorris168 commented 7 months ago

What's the plan for remodularising this code?

Are you happy for that to be a separate task?

JanionNevill commented 7 months ago

What's the plan for remodularising this code?

Are you happy for that to be a separate task?

I am happy for it to be a separate Jira task, but I would like to see it done before this gets merged into the feature branch. It's mostly due to "technical debt"; if it is modularised straight away, then it shouldn't matter that it's not merged straight away, but if it doesn't get modularised straight away, then things could change elsewhere and make the process a lot more difficult.

You can continue on this branch and rerequest a review when it's done, or you could make another branch off this one, then merge back both at once. Does that sound reasonable? I'm happy to discuss further if you want.