tidyhydro — presubmission inquiry

[atsyplenkov]

Submitting Author Name: Anatoly Tsyplenkov
Submitting Author Github Handle: @atsyplenkov
Repository: https://github.com/atsyplenkov/tidyhydro
Submission type: Pre-submission
Language: en

---

• Paste the full DESCRIPTION file inside a code block below:

Package: tidyhydro
Type: Package
Title: Tidy Metrics for Assessing Hydrological Models Performance
Version: 0.1.1
Authors@R:
    person(given = "Anatoly", family = "Tsyplenkov", email = "atsyplenkov@fastmail.com", role = c("cre", "aut", "cph"), comment = c(ORCID = "0000-0003-4144-8402"))
Maintainer: Anatoly Tsyplenkov <atsyplenkov@fastmail.com>
Description: Provides tidy tools for comparing simulated and observed hydrological time series. Includes compatibility with the 'yardstick' package for model performance evaluation using commonly used metrics such as the Nash–Sutcliffe Efficiency (NSE), Kling–Gupta Efficiency (KGE), percent bias (pBIAS) and etc.
License: MIT + file LICENSE
Depends: R (>= 4.1.0)
Imports:
    Rcpp (>= 1.0.12),
    rlang (>= 1.1.0),
    yardstick (>= 1.3.1)
LinkingTo: Rcpp
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.3.2
Encoding: UTF-8
Language: en-US
Suggests:
    hydroGOF,
    testthat (>= 3.0.0),
    quickcheck (>= 0.1.3),
    quarto
Config/testthat/edition: 3
URL: https://github.com/atsyplenkov/tidyhydro, https://atsyplenkov.github.io/tidyhydro/
BugReports: https://github.com/atsyplenkov/tidyhydro/issues
LazyData: true
Config/Needs/website: bench, ggplot2, quarto, lubridate, dplyr

Scope

• Please indicate which category or categories from our package fit policies or statistical package categories this package falls under. (Please check one or more appropriate boxes below):

  Data Lifecycle Packages

  • data retrieval
  • data extraction
  • data munging
  • data deposition
  • data validation and testing
  • workflow automation
  • version control
  • citation management and bibliometrics
  • scientific software wrappers
  • field and lab reproducibility tools
  • database software bindings
  • geospatial data
  • text analysis

  Statistical Packages

  • Bayesian and Monte Carlo Routines
  • Dimensionality Reduction, Clustering, and Unsupervised Learning
  • Machine Learning
  • Regression and Supervised Learning
  • Exploratory Data Analysis (EDA) and Summary Statistics
  • Spatial Analyses
  • Time Series Analyses
  • Probability Distributions

• Explain how and why the package falls under these categories (briefly, 1-2 sentences). Please note any areas you are unsure of:

tidyhydro fits within the data validation and testing category / regression , as it provides automated, quantitative metrics (e.g., NSE, KGE, pBIAS) for evaluating the quality and performance of hydrological models compatible with the tidymodels infrastructure.

• If submitting a statistical package, have you already incorporated documentation of standards into your code via the srr package?

No, I just learned about the srr from the issue template. Feel a bit confused regarding that standards and not sure if it is applicable to my package.

• Who is the target audience and what are scientific applications of this package?

The target audience comprises the hydrological/ecological community and the general data science community who uses regression models within their workflows.

• Are there other R packages that accomplish the same thing? If so, how does yours differ or meet our criteria for best-in-category?

Yes, there is a package hydroGOF, which has some overlap with tidyhydro in terms of functionality. For example, metrics such as MSE, pBIAS, NSE, and KGE are implemented in both packages. However, tidyhydro provides a tidyverse-style interface, giving full compatibility and interoperability with the popular tidymodels framework (via yardstick). Additionally, tidyhydro uses C++ via Rcpp for performance, which helps in validating large time series efficiently (see benchmarks).

• (If applicable) Does your package comply with our guidance around Ethics, Data Privacy and Human Subjects Research?

Not Applicable

• Any other questions or issues we should be aware of?:

1. I raised this issue out of curiosity, as I have been thinking about submitting some package to rOpenSci for several years. However, I am not 100% sure that tidyhydro falls within the aims and scope of rOpenSci. I am interested to hear your thoughts about it.
2. The local pkgcheck fails to find continuous integration checks, which confuses me, as I do have a very comprehensive CI workflow set up. I am curious to see the results of the rOpenSci-hosted pkgcheck.
3. I do need some guidance regarding srr standards. As far as I can understand from the srr manual, it is not really applicable to my package, as tidyhydro is not introducing new methods but rather provides a re-implementation of a well-known ones in a tidy way.

[mpadge]

Thanks @atsyplenkov for your inquiry here. My first impressions are that the package seems like a good fit here indeed! Short responses to your specific questions:

1. I raised this issue out of curiosity, as I have been thinking about submitting some package to rOpenSci for several years. However, I am not 100% sure that `tidyhydro` falls within the aims and scope of rOpenSci. I am interested to hear your thoughts about it.

Already stated above: I'm very confident that this package will be a good fit (with proviso under 3rd point below).

2. The local `pkgcheck` fails to find continuous integration checks, which confuses me, as I do have a very comprehensive CI workflow set up. I am curious to see the results of the rOpenSci-hosted `pkgcheck`.

Sorry for confusion there. That confusing message has been updated this week. It won't be deployed in our bot here until next week, but you can also install https://github.com/ropensci-review-tools/pkgcheck-action in your repo to generate reports on every push.

3. I do need some guidance regarding `srr` standards. As far as I can understand from the `srr` manual, it is not really applicable to my package, as `tidyhydro` is not introducing new methods but rather provides a re-implementation of a well-known ones in a tidy way.

It's perfectly okay for a stats package to be a re-implementation. I think the biggest initial challenge we'll have here is deciding in which category the package best fits. For that, the guideline is that any category in which you think you'll be able to comply with at least half of standards is the right category. You've indicated Regression and Supervised Learning above, although it seems to me that the actual implementations are more summary estimators than regression models. Can you please have a look through our standards and think about how well you'll be able to meet these categories:

• Regression and Supervised Learning, because that's what you indicated above
• Exploratory Data Analysis and Summary Statistics which I think may be a good fit
• Time Series may be applicable, because as far as I understand your package, initial input data are presumed to be hydrological time series.
• Probability Distributions which I do not think will apply here, but I suspect just reading those may give you a different perspective on some of the implications of distributional assumptions currently hard-coded within your routines.

It's okay to tick multiple categories, but that just entails more work for you in documenting compliance standards. In short: Please look through, and report back on which category you think you'll be able to comply with the greatest number of standards. Thanks, and feel free to ask any questions here in the meantime 👍

[atsyplenkov]

Hi @mpadge 👋! I appreciate your quick and very detailed response. It has definitely shed a lot of light! I agree that the Exploratory Data Analysis and Summary Statistics category feels like the best fit.

That being said, I am happy to proceed with the full package review in the near future. I just need some time to check the documentation against the srr standards, as I wish to aim at the silver badge at least. I will raise an issue when I am ready.

One more question regarding issue templates. There is a significant difference between general software and statistical software templates. I noticed that the template for statistical packages doesn't include a Methods in Ecology and Evolution (MEE) article option. Does this mean statistical software is not welcome in the MEE journal, or is the template just outdated?

[mpadge]

Thanks @atsyplenkov, i'll put this issue on hold for the moment then. That just means the bot will add a "holding" label to this issue, which we can remove once things get going again. When you're ready, we'll remove that label, run our automated checks here, and hopefully proceed to a full submission from that point. The bot will send a reminder after 3 months.

As for MEE: That's a definite Yes, just that your package would be the first stats package to go on to MEE. Our understanding with them is that our peer review is acceptable as a phase of review for MEE, and there's no reason why stats peer review would be considered any different (especially given that the requirements are considerably more stringent than for non-stats review). If you were to follow that track, then we'd request your help in updating our templates and associated documentation once things get that far.

[mpadge]

@ropensci-review-bot put on hold

[ropensci-review-bot]

Submission on hold!

[atsyplenkov]

Awesome, thanks @mpadge. I will let you know when I am ready to proceed.