Closed padilla410 closed 2 years ago
Thank you for the review! In addition to tackling your comments, I also made a few updates to 2_process
and to the description of this PR. I intend to talk about some of the issues outlined above under issues and questions
at our meeting later today.
This PR completes three tasks related to
1_fetch
and one task related to2_process
:1_fetch
fetch-process-viz
pattern from thetargets
pipeline training2_process
)2_process
.tif
) for the high and low pesticide estimatesThe next step will be using the
targets
generated inp1
to create the individual pesticide rasters. The next portion of the pipeline will use dynamic branching and will map over the filet targets created inp1
.I did a bit of manual testing to verify that modifying values in a file (e.g.,
1_fetch/in/maps/dbf19/H_24D.dbf
) will result in a target rebuild.Issues and questions
1_fetch
I am usingtar_files
to make sure that the 4 types of input files are tracked and can then be branched over to load the data. There is one issue with this: when runningtar_visnetwork
these targets always appear "outdated". This is a known issue. As suggested by the doc, I tried switching fromtar_files
totar_files_input
. That prevented the pipeline from working when it tried to produce the targets that depend onp1_pest_of_interest
(p1_pest_label_csv
andp1_pest_bin_csv
). I was not able to determine why it failed, outside of the ambiguous error message:I did verify that the internal components of
list_pest_csv()
work as intended.2_process
to the inputs of3_viz