[aappling-usgs]

This MR could be merged after #264 because this one contains a version update 1.9.3->1.9.4. If we do that, we should add NEWS entries for #264 in one of the PRs- I think this one would be easier.

But this MR contains its own stuff, too:

• resolve EGRET populateDaily now merges to 1 row per day #265: updates for changes in EGRET to the behavior of populateDaily. I've written about the thinking & changes in this PR within that issue.
• resolve preds_lm has inconsistent col name #258: loadLm was returning predictions in a column named "fit", which should be "conc" or "flux" to be consistent with other models. That issue is so simple there's not much to discuss, except that I can confirm no regression (or improvement) in tests related to loadLm.

Test results: attaching a bundle of test scripts and pre-and-post change scripts, pretty much. The trick here is that I needed different EGRET versions installed to understand the regression and then the fix. The test files:

1. 'test-output-261005-1736.rds' - representative of the state before EGRET changed
2. 'test-output-261013-1102.rds' - representative of the new problems due to the EGRET::populateDaily update. This test might also sit on other loadflex updates as in Address issues 259 260 261 #264, but as noted in Address issues 259 260 261 #264 (comment), the new test errors relative to 261005-1736 appear to be due entirely to the change to EGRET
3. 'test-output-261013-1447.rds' - after the changes in this PR, with tests fixed relative to BOTH of the preceding files

The test scripts and outputs are bundled here:
tests-PR266.zip

To summarize the changes, this MR fixes two old issues (already present in 261005-1736) and the six issues that popped up as soon as I installed EGRET updates this week (new in 261013-1102).

Details on errors and test comparisons

Changes from 1 to 2 (the new errors to fix):

r$> compare_edits('temp/test-output-261005-1736.rds', 'temp/test-output-261013-1102.rds')
OLD: temp/test-output-261005-1736.rds (Phase 2 complete — tidyverse-drift fixes applied)
    Tests: 89 | Passed: 41 | Failed: 61 | Errors: 16 | Warnings: 21 | Skipped: 24
NEW: temp/test-output-261013-1102.rds (Edits for PR #264, "Address issues 259 260 261")
    Tests: 89 | Passed: 35 | Failed: 61 | Errors: 22 | Warnings: 21 | Skipped: 24

========================================================
  ERRORS (structured, matched by file + test)
========================================================
Totals: OLD = 16 across 16 tests | NEW = 22 across 22 tests

DEGRADED - ERRORS newly present in these tests
----------------------------------------------
test-44-plots.R - plotEGRET works for fitdat and meta (now 1)

test-51-EGRET.R - convertToEGRET allows metadata and data or newdata without model (now 1)

test-51-EGRET.R - convertToEGRET complains about redundant metadata or data (now 1)

test-51-EGRET.R - convertToEGRETDaily returns partial info without preds (now 1)

test-51-EGRET.R - convertToEGRETSample correctly converts (now 1)

test-51-EGRET.R - expandFlowForEGRET returns correct columns (now 1)

Changes from 1 to 3 (this PR fixes two old EGRET-related errors from before the EGRET change - a bonus):

r$> compare_edits('temp/test-output-261005-1736.rds', 'temp/test-output-261013-1447.rds')
OLD: temp/test-output-261005-1736.rds (Phase 2 complete — tidyverse-drift fixes applied)
    Tests: 89 | Passed: 41 | Failed: 61 | Errors: 16 | Warnings: 21 | Skipped: 24
NEW: temp/test-output-261013-1447.rds (v1.9.4: adapt to new EGRET::populateDaily behavior; fix #258)
    Tests: 91 | Passed: 45 | Failed: 61 | Errors: 14 | Warnings: 21 | Skipped: 24

========================================================
  ERRORS (structured, matched by file + test)
========================================================
Totals: OLD = 16 across 16 tests | NEW = 14 across 14 tests

IMPROVED - ERRORS fully gone from these tests
---------------------------------------------
test-51-EGRET.R - convertToEGRET returns prediction-free Daily when predictions are missing (had 1)

test-51-EGRET.R - convertToEGRETDaily correctly converts (had 1)

Changes from 2 to 3 (this PR was focused on fixing the six errors introduced between 1 and 2, and it succeeds as demonstrated here):

r$> compare_edits('temp/test-output-261013-1102.rds', 'temp/test-output-261013-1447.rds')
OLD: temp/test-output-261013-1102.rds (Edits for PR #264, "Address issues 259 260 261")
    Tests: 89 | Passed: 35 | Failed: 61 | Errors: 22 | Warnings: 21 | Skipped: 24
NEW: temp/test-output-261013-1447.rds (v1.9.4: adapt to new EGRET::populateDaily behavior; fix #258)
    Tests: 91 | Passed: 45 | Failed: 61 | Errors: 14 | Warnings: 21 | Skipped: 24

========================================================
  ERRORS (structured, matched by file + test)
========================================================
Totals: OLD = 22 across 22 tests | NEW = 14 across 14 tests

IMPROVED - ERRORS fully gone from these tests
---------------------------------------------
test-44-plots.R - plotEGRET works for fitdat and meta (had 1)

test-51-EGRET.R - convertToEGRET allows metadata and data or newdata without model (had 1)

test-51-EGRET.R - convertToEGRET complains about redundant metadata or data (had 1)

test-51-EGRET.R - convertToEGRET returns prediction-free Daily when predictions are missing (had 1)

test-51-EGRET.R - convertToEGRETDaily correctly converts (had 1)

test-51-EGRET.R - convertToEGRETDaily returns partial info without preds (had 1)

test-51-EGRET.R - convertToEGRETSample correctly converts (had 1)

test-51-EGRET.R - expandFlowForEGRET returns correct columns (had 1)

[github-actions]

📖 pkgdown preview for this PR: https://doi-usgs.github.io/loadflex/preview/pr266/

[gfkltn]

Admin review completed. No issues found.