Remove complete logging, update readme, set v0.1

DESCRIPTION

@@ -2,7 +2,7 @@ Type: Package
 Package: netidmtpreg
 Title: Net Survival Binomial Regression for Progressive Illness Death
     Models
-Version: 0.0.0.9000
+Version: 0.1.0
 Authors@R: 
     person("Quentin", "Marcou", , "[email protected]", role = c("aut", "cre"),
            comment = c(ORCID = "0000-0001-7074-2761"))

R/renewnetTPreg.R

@@ -1,6 +1,7 @@
-# Code adapted from idmTPreg package v1.1 under GNU-GPLv2
+# Code initially inspired from idmTPreg package v1.1 under GNU-GPLv2
 # Original Authors: Leyla Azarang and Manuel Oviedo de la Fuente
-# Adapted for net survival setting by Quentin Marcou, based on <ref>
+# Largely rewritten and adapted for net survival setting by Quentin Marcou,
+# based on <ref>
 
 
 #' @title Wraps the call to mod.glm.fit to handle convergence errors
@@ -32,7 +33,7 @@ mod.glm.fit.errorwrapper <-
            maxmaxit = 1000,
            warning_str = "",
            ...) {
-    result <- tryCatchLog::tryCatchLog({
+    result <- tryCatch({
       # Try
       withCallingHandlers({
         mod.glm.fit2(

README.Rmd

@@ -23,7 +23,7 @@ knitr::opts_chunk$set(
 
 The goal of netidmtpreg is to enable net survival estimation through direct
 binomial regression, allowing modeling continuous covariate effects that could
-not be handled through stratified Pohar-Perme estimation, all this in a 
+not be handled through e.g stratified Pohar-Perme estimation, all this in a 
 multistate Illness-Death setting.   
 
 ## Installation
@@ -59,7 +59,7 @@ n_ind <- 1e2 # number of simulated individuals
 synth_idm_data <- generate_uncensored_ind_exp_idm_data(
   n_individuals = n_ind,
   lambda_illness = 1.0,
-  lambda_death = 0.1
+  lambda_death = 1.0
 )
 
 # Generate random age and sex labels
@@ -77,7 +77,7 @@ synth_idm_data <-
   ))
 
 # Generate population mortality assuming equal constant population rate
-l_pop_death <- 0.1 # extra disease mortality doubles the population mortality
+l_pop_death <- 1.0 # extra disease mortality doubles the population mortality
 population_death_times <- generate_exponential_time_to_event(
   n_individuals = n_ind,
   lambda = l_pop_death
@@ -92,13 +92,13 @@ crude_synth_idm_data <- netidmtpreg:::apply_iddata_death(
 )
 ```
 
-Now let's carry the of net survival:
+Now let's carry the of net survival estimation:
 ```{r estimation, echo=TRUE, warning=FALSE, message=FALSE}
 # Estimation can be sped up and carried in parrallel using futures:
 future::plan("multisession") # will work on any OS
 # future::plan("multicore") # more efficient but only works on UNIX systems
 net_estimate <- renewnetTPreg(
-  formula = ~ sex,  # intercept + binary covariate
+  formula = ~1, # intercept + binary covariate
   data = crude_synth_idm_data,
   # Use a standard ratetable
   ratetable = const_ratetable,
@@ -108,8 +108,9 @@ net_estimate <- renewnetTPreg(
     year = start_date
   ),
   time_dep_popvars = list("age", "year"),
-  s = 0.5,
-  trans = "13",
+  s = 0.2,
+  by = n_ind / 10,
+  trans = "11",
   link = "logit",
   R = 100 # Number of bootstraps
 )
@@ -118,8 +119,9 @@ future::plan("sequential") # close the multisession, see future's documentation
 
 We obtain a `TPreg` object with a dedicated plotting method using ggplot:
 
-```{r plot, echo = FALSE}
-plot(net_estimate)
+```{r plot}
+plot(net_estimate) + ggplot2::ylim(-5, 5)
 ```
 
+Let's make sure we a
 More examples to come!

README.md

@@ -13,7 +13,7 @@ Status](https://img.shields.io/badge/lifecycle-maturing-blue.svg)](https://lifec
 
 The goal of netidmtpreg is to enable net survival estimation through
 direct binomial regression, allowing modeling continuous covariate
-effects that could not be handled through stratified Pohar-Perme
+effects that could not be handled through e.g stratified Pohar-Perme
 estimation, all this in a multistate Illness-Death setting.
 
 ## Installation
@@ -61,7 +61,7 @@ n_ind <- 1e2 # number of simulated individuals
 synth_idm_data <- generate_uncensored_ind_exp_idm_data(
   n_individuals = n_ind,
   lambda_illness = 1.0,
-  lambda_death = 0.1
+  lambda_death = 1.0
 )
 
 # Generate random age and sex labels
@@ -79,7 +79,7 @@ synth_idm_data <-
   ))
 
 # Generate population mortality assuming equal constant population rate
-l_pop_death <- 0.1 # extra disease mortality doubles the population mortality
+l_pop_death <- 1.0 # extra disease mortality doubles the population mortality
 population_death_times <- generate_exponential_time_to_event(
   n_individuals = n_ind,
   lambda = l_pop_death
@@ -101,7 +101,7 @@ Now let’s carry the of net survival:
 future::plan("multisession") # will work on any OS
 # future::plan("multicore") # more efficient but only works on UNIX systems
 net_estimate <- renewnetTPreg(
-  formula = ~ sex,  # intercept + binary covariate
+  formula = ~ 1,  # intercept + binary covariate
   data = crude_synth_idm_data,
   # Use a standard ratetable
   ratetable = const_ratetable,
@@ -111,21 +111,25 @@ net_estimate <- renewnetTPreg(
     year = start_date
   ),
   time_dep_popvars = list("age", "year"),
-  s = 0.5,
-  trans = "13",
+  s = 0.2,
+  by=n_ind/10,
+  trans = "11",
   link = "logit",
   R = 100 # Number of bootstraps
 )
+#> [1] "estimate"
+#> [1] "bootstrap"
 future::plan("sequential") # close the multisession, see future's documentation
 ```
 
 We obtain a `TPreg` object with a dedicated plotting method using
 ggplot:
 
-    #> Warning: Removed 1 row containing missing values or values outside the scale range
-    #> (`geom_line()`).
-    #> `geom_line()`: Each group consists of only one observation.
-    #> ℹ Do you need to adjust the group aesthetic?
+``` r
+plot(net_estimate) + ggplot2::ylim(-5,5)
+#> Warning: Removed 1 row containing missing values or values outside the scale range
+#> (`geom_line()`).
+```
 
 <img src="man/figures/README-plot-1.png" width="100%" />