Clean up evaluation error tooltips in agent report

[yjunechoe]

Summary

This is a minor QoL visual change to the contents of the tooltip text for evaluation errors in the agent report.

df <- data.frame(date = "invalid date")
agent <- create_agent(df) |> 
  col_vals_equal(date, Sys.Date()) |> 
  interrogate(progress = TRUE)
#> 
#> ── Interrogation Started - there is a single validation step ───────────────────
#> ℹ Step 1: an evaluation issue requires attention (ERROR).
#> 
#> ── Interrogation Completed ─────────────────────────────────────────────────────

Before:

After:

Note that the error object with the full context is still available in $capture_stack$error, if one wanted to diagnose it more seriously.

---

Implementational details:

1. validation_set$capture_stack now tracks an additional information about the call passed to pointblank_try_catch(), in $pb_call:

   agent$validation_set$capture_stack[[1]]$pb_call
   #> tbl_val_comparison(table = table, column = column, operator = operator, 
   #>     value = value, na_pass = na_pass)

   Two unit tests were updated to reflect this change to the contents of $capture_stack.

2. This information about the pointblank-internal call lays dormant in the agent, and is only consumed at get_agent_report() as part of formatting the error objects to string before rendering to html <title>. This is handled by a new internal function pointblank_cnd_to_string().

Note that warning objects turned out to be are somewhat more complicated to work with, so the reformatting currently only applies to errors.

[rich-iannone]

This is great, thanks!

[rich-iannone]

@yjunechoe Feel free to merge this in at your leisure!