Update ggcoxdiagnostics.R, fix for ox.scale = "time" and non unique eventtimes

[Istalan]

I hope this is the correct place for this, as this is my first time really using Github.

Pleasantries first: This is a wonderful package and i'm very thankful for having it.

The issue

The following Code produces empty plots:

library("survival")
library("survminer")
fit <- coxph(Surv(time, status) ~sex + age, data = lung)
ggcoxdiagnostics(fit, type = "schoenfeld", ox.scale = "time")

also it generates the following warning:

 Warning messages:
1: In ggcoxdiagnostics(fit, type = "schoenfeld", ox.scale = "time") :`
  NAs introduced by coercion
2: Removed 330 rows containing non-finite values (stat_smooth). 
3: Removed 330 rows containing missing values (geom_point).

I could not find an issue for this problem, but it was also described here:
https://community.rstudio.com/t/ggcoxdiagnostics-warning-message/51867

My understanding of the problem and my solution

The introduction of NAs by Coercion happens in current line 92

xval <- as.numeric(rownames(res))

with res being defined by

res <- as.data.frame(resid(fit, type = type))

While resid(fit, type = type) returns an object with rownames that correspond to the sorted event-times, if the rownames are not unique the casting to data.frame will make them:

raw_res <- resid(fit, type = "schoenfeld")
res <- as.data.frame(raw_res)
rownames(raw_res)[1:4]
rownames(res)[1:4]

The rownames of the data.frame then can't be converted to numeric. My Solution is to just keep the initial result raw_res <- resid(fit, type = "schoenfeld") to initiate xval.