Maintain feature parity with model-res-avm
#19
Conversation
| test_data_card <- read_parquet(paths$output$test_card$local) | ||
| # recent 10% of sales and already includes predictions. | ||
| test_data_card <- read_parquet(paths$output$test_card$local) %>% | ||
| filter(!is.na(loc_census_puma_geoid)) |
There was a problem hiding this comment.
This should not be here, but I cannot get this stage to run without it.
There was a problem hiding this comment.
I don't think we actually need to include these additional workflow scripts in the condo model, since the res model workflow can work on both sets of output.
There was a problem hiding this comment.
@wrridgeway I'm sure there are lots of little kinks to work out, but overall this looks good. Only changes are to boot the files/chunks that we don't need to duplicate across the repos. Then this should be good to go.
There was a problem hiding this comment.
Likewise for this workflow, we can keep it res model only.
| dict <- read.csv( | ||
| here::here("misc", "file_dict.csv"), | ||
| col_types = readr::cols() | ||
| colClasses = c("character", "character", "numeric", rep("character", 9)), | ||
| na.strings = "" |
There was a problem hiding this comment.
Why the switch here? readr should be in the dependencies IIRC
There was a problem hiding this comment.
This is a syntax copied from the res repo
See #18.
This PR is rather extensive. It primarily refactors the pipeline and surrounding infrastructure in order to bring it in line with the res model pipeline.
It's not up-to-date with the last week or so of work in the res pipeline, but this PR is getting pretty unwieldy.