counts for pkcg03 plot should be per cohorts

[Melkiades]

Discussed in #130

^{Originally posted by giscus[bot] September 22, 2023}

tlg-catalog/graphs/pharmacokinetic/pkcg03

Plot of Mean PK Concentration Over Time by Cohort

https://insightsengineering.github.io/tlg-catalog/graphs/pharmacokinetic/pkcg03.html

[legrasv]

A bit of code to help:

I had to hardcode the function g_lineplot it for one of my TLGs:

# df_N <- as.data.frame(table(alt_counts_df[[strata]], exclude = c(NA, NaN, Inf))) # nolint
df_N <- aggregate(USUBJID ~ eval(parse(text = strata)), data = alt_counts_df, FUN = function(x) length(unique(x))) # nolint

knowing that USUBJID has to be defined in the parameters of the function somehow

[Melkiades]

A bit of code to help:

I had to hardcode the function g_lineplot it for one of my TLGs:

# df_N <- as.data.frame(table(alt_counts_df[[strata]], exclude = c(NA, NaN, Inf))) # nolint
df_N <- aggregate(USUBJID ~ eval(parse(text = strata)), data = alt_counts_df, FUN = function(x) length(unique(x))) # nolint

knowing that USUBJID has to be defined in the parameters of the function somehow

@ayogasekaram could you take a look at this, please? Thank you!!! :)

[legrasv]

@ayogasekaram Actually on this plot, there is something else that is wrong:

It does not link all the points (because of the different nominal time that we have between the Treatment groups).

So I could also fix it myself by adding in g_lineplot :
df_stats <- df_stats[!is.na(df_stats$mean),] ## could be coded in dplyr as well

before the line of code:

• ##################################################

• Prepare certain plot's properties.

• ##################################################

[legrasv]

Result:

[ayogasekaram]

thank you @legrasv for your code suggestions! I believe your second suggestion is already implemented in g_lineplot here. the default for the mid variable is "mean" which is identical to the line of code you have above.

[legrasv]

Sounds good thank you @ayogasekaram

[legrasv]

Hello

@ayogasekaram @Melkiades
Could you please update the code of matches variable so that it works for this case:
It was show on the legend plot Day 1 Day 10 Day 5 :/

df_N = data.frame(AVISIT = factor(c("Day 1", "Day 10", "Day 5")))
levels(df_N$AVISIT) = c("Day 1", "Day 5", "Day 10")

group_var = "AVISIT"

strata_N <- paste0(group_var, "_N") # nolint
df_N[[strata_N]] <- paste0(df_N[[group_var]], " (N = 1)") # nolint

levels(df_N$AVISIT)
[1] "Day 1" "Day 5" "Day 10"

keep strata factor levels

matches <- sapply(unique(df_N[[group_var]]), function(x) {
    +   regex_pattern <- gsub("([][(){}^$.|*+?\\\\])", "\\\\\\1", x)
    +   unique(df_N[[paste0(group_var, "_N")]])[grepl(
      +     paste0("^", regex_pattern),
      +     unique(df_N[[paste0(group_var, "_N")]])
      +   )]
    + })

print(matches)

[[1]]
[1] "Day 1 (N = 1)" "Day 10 (N = 1)"

[[2]]
[1] "Day 5 (N = 1)"

[[3]]
[1] "Day 10 (N = 1)"

The code below works well

matches <- sapply(unique(df_N[[group_var]]), function(x) {
    +   regex_pattern <- gsub("([][(){}^$.|*+?\\\\])", "\\\\\\1", x)
    +   search_strings[grepl(
      +     paste0("^", regex_pattern, "\\b"),
      +     search_strings
      +   )]
    + })

print(matches)

[1] "Day 1 (N = 1)" "Day 5 (N = 1)" "Day 10 (N = 1)"

[ayogasekaram]

Hey @legrasv! thank you for letting me know - I can work on making the update in tern. Quick question: is search_strings in your preferred code a custom function?