john-harrold / ubiquity Goto Github PK

After parameter estimation compare the estimates to the bounds of those parameters. For parameters near their bounds notify the user.

Speed up generate_subject: calling mvrnorm for each subject seems to be slower than calling it for all subjects. Also look into sparse block diagonal vs normal matrices 

When running pso and GA check to make sure the parameters being estimated have realistic bounds. If they don't throw and error and exit gracefully.

Specifically the Reporting vignette.
`
checking re-building of vignette outputs ... [36s/43s] ERROR
Error(s) in re-building vignettes:
...
--- re-building ‘Deployment.Rmd’ using rmarkdown
--- finished re-building ‘Deployment.Rmd’

--- re-building ‘Estimation.Rmd’ using rmarkdown
--- finished re-building ‘Estimation.Rmd’

--- re-building ‘Language.Rmd’ using rmarkdown
--- finished re-building ‘Language.Rmd’

--- re-building ‘NCA.Rmd’ using rmarkdown

Quitting from lines 358-359 [unnamed-chunk-26] (NCA.Rmd)
Error: processing vignette 'NCA.Rmd' failed with diagnostics:
argument must be coercible to non-negative integer
--- failed re-building ‘NCA.Rmd’

--- re-building ‘Reporting.Rmd’ using rmarkdown

Quitting from lines 523-524 [unnamed-chunk-39] (Reporting.Rmd)
Error: processing vignette 'Reporting.Rmd' failed with diagnostics:
argument must be coercible to non-negative integer
--- failed re-building ‘Reporting.Rmd’

--- re-building ‘Simulation.Rmd’ using rmarkdown
--- finished re-building ‘Simulation.Rmd’

--- re-building ‘Titration.Rmd’ using rmarkdown
--- finished re-building ‘Titration.Rmd’

SUMMARY: processing the following files failed:
‘NCA.Rmd’ ‘Reporting.Rmd’

Error: Vignette re-building failed.
Execution halted
`

The fortranify_line function hangs when lines cannot be wrapped. This should just fail, throw an error, and display the line for the user to see.

Add footers to slides
Add masters:
2 rows x columns

• 1 column 2 rows
• column 1: 2 rows; column 2: 1 row

Add the ability to specify time intervals and output partial AUCs for those intervals

Add the ability to use table and figure keys as references in text and have them automatically linked to the correct sequence.

Place an obvious message after termination or before parameter estimates indicating the criteria for stopping estimation.

Abstract out rcomponents functions by adding model short name as a prefix and fixing all of the function calls to include those.

When running simulate_subjects() in parallel the progress indicator is broken.

When dumping secondary parameters to r_components the conditional statements may not be being dumped. 

Add a system file notation to add lags to bolus and infusion compartments.

Tlag can be dependent on: system parameters, secondary parameters, covariates.

Check auto_rcomponent.R. infusion rates are evaluated before covariates. Check Matlab and see if this is consistent. 

To prevent breaking office reports check to see if there are slip_in calls that can make things more robust.

davidgohel/officer#343

Vignette/documentation for dsfilter, dsinc, NCA_options, system_nca_summary() also include Markdown examples/description. If you filter out predose data it won’t be used in back extrapolation 

It seems that system_plot_cohorts() isn't pulling the width and the height from the specified plot options.

Add option to control individual specific data to be used when performing half-life calculations

Check the rxode2 output to ensure it's correctly creating the IC values for the systems

Add warning when you have a bolus or infusion rate that starts before the first output time and add that time.