Look inside the formula environment first when materializing ~ tbl

[yjunechoe]

Summary

Fixes an edge case where tbl is passed as formula and the variable is defined in an environment not available at interrogate(). The following reprex from #598 (comment) now works as intended:

agent <- local({
  df <- data.frame(x = 1)
  create_agent(tbl = ~ df)
})
agent %>% 
  rows_distinct(x) %>% 
  interrogate()

Previously, cases like this would error from not searching for df in the formula environment (in this case, it would scope the stats::df() function instead).

Related GitHub Issues and PRs

• Ref: Error in rows_distinct(): must be a vector, not a function #598

Checklist

• I understand and agree to the Code of Conduct.
• I have listed any major changes in the NEWS.
• I have added testthat unit tests to tests/testthat for any new functionality.

[yjunechoe]

2afe74d should also fix #602:

informant <- local({
  df <- data.frame(x = 1)
  create_informant(tbl = ~ df)
})
informant <- informant %>% 
  incorporate()
#> 
#> ── Incorporation Started ────────────────────────────────────────────────────────
#> ✔ Information gathered.
#> ✔ Information built.
#> 
#> ── Incorporation Completed ──────────────────────────────────────────────────────

[rich-iannone]

LGTM!

[rich-iannone]

@yjunechoe ! You’re on a tear lately, and this package is becoming super stable because of all that :)

[yjunechoe]

thanks! It's a pleasure working on this codebase! :)