clavellab / maldipickr Goto Github PK

Currently, the spectra and peaks are stored as RDS files containing MassSpectrum or MassPeaks R objects. These objects are metadata-rich which is fundamental.

There is indeed a couple of standard file format (e.g., mzML) to which these objects can be exported using MALDIquantForeign, but the metadata are loss in the process..

in the README quickstart example

Prepare for release:

• git pull
• Check current CRAN check results
• Check if any deprecation processes should be advanced, as described in Gradual deprecation
• Polish NEWS
• urlchecker::url_check()
• devtools::build_readme()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• revdepcheck::revdep_check(num_workers = 4) no rev deps
• Update cran-comments.md
• git push
• Draft blog post

Submit to CRAN:

• usethis::use_version('minor')
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• Add preemptive link to blog post in pkgdown news menu
• usethis::use_github_release()
• usethis::use_dev_version(push = TRUE)
• Finish blog post
• Tweet

needed before #4

[2023-10-24T15:35:03.710Z] ['error'] There was an error running the uploader:
Error uploading to https://codecov.io
Error: There was an error fetching the storage URL during POST:
404 - {'detail': ErrorDetail(string='Unable to locate build via Github Actions API.
Please upload with the Codecov repository upload token to resolve issue.', code='not_found')}

https://github.com/ClavelLab/maldipickr/actions/runs/6629129466/job/18007735293#step:8:39

will need to render all "no reliable identification" unique though.

this error can happen when multiple spectra have the same name, which should not happen with proper research data input but can arise during data analyses.

At the moment, the function stops. An improvement would be to exit gracefully at least, or provide a solution (would make.unique() be possible or too intrusive?)

The example tables in the vignette are limited in width and some relevant columns are masked.

Could be solved by adding a dplyr::relocate() to some of the examples

example_df %>% dplyr::relocate(name, to_pick)

A possible option to reduce the length of the vignettes is to split into 3 vignettes:

• Import data
• Process data
• Cherry-pick

While examples are part of the vignettes, additional examples could be added that would not make it to the examples section of the function but "just" in the vignettes.

Make sure to distinguish tutorials from how-to using https://diataxis.fr/

Improve coverage of functions: https://app.codecov.io/github/ClavelLab/maldipickr/tree/main/R
especially key functions:

• #25
• #26
• #27

Describe the bug
Current Github Actions uses Node.js 16 which is deprecated in favor of 20.

To Reproduce
Look up the Actions tab for warnings

Expected behavior
No warnings.

Solutions

Update:

• actions/checkout@v3
• JamesIves/[email protected]
• codecov/codecov-action@v3

Screenshots

File format conversion can be made with the compass software, see rformassspectrometry/RforMassSpectrometry.org#18 (comment)

https://common-changelog.org/

decide on a strategy to set a minimum R version. Currently, it implicitly uses the R version used for developing.

• https://blog.r-hub.io/2022/09/12/r-dependency/

Resources to read:

• https://r-pkgs.org/release.html
• https://github.com/ThinkR-open/prepare-for-cran
• https://github.com/DavisVaughan/extrachecks

Issues to solve:

• hide non standard files like codecov
• CITATION in inst (consider building the CITATION.cff from {cffr} https://docs.ropensci.org/cffr/ ?)
• solve the qpdf issue
• adjust proper email
• add contributors
• try out installation via rhub https://r-hub.github.io/rhub
• #20
• #21
• #22
• #24

Consider adding a remove_empty() function to clean up the spectra, the peaks and the metadata files from the object returned by process_spectra().

Could be coupled with check_spectra() to flag the elements of the list.

Is your feature request related to a problem? Please describe.
I'm always frustrated when I need to compute how many raw spectra were analyzed, how many valid, and why they were rejected.

Describe the solution you'd like
This function below is an untested, undocumented attempt:

gather_spectra_stats <- function(check_vectors){
  # check_vectors from maldipickr::check_spectra
  # src: https://stackoverflow.com/a/51140480/21085566
  aggregated_checks <- Reduce(`|`, check_vectors)
  check_stats <- vapply(check_vectors, sum, FUN.VALUE = integer(1)) %>%
    tibble::as_tibble_row()
  tibble::tibble(
    "n_spectra" = length(aggregated_checks),
    "n_valid_spectra" = n_spectra - sum(aggregated_checks)
  ) %>%
    dplyr::bind_cols(check_stats) %>% 
    return()
}

In the example of pick_spectra(), reproduced below:

# 4.2 Pick the spectra from clusters without spectra
#   labelled as `picked_before` (hard masking).
pick_spectra(clusters, metadata, "OD600",
  hard_mask_column = "picked_before"
)

A warning advocating for the replacement of summarise() by reframe() :

#> Warning: Returning more (or less) than 1 row per `summarise()` group was deprecated in
#> dplyr 1.1.0.
#> ℹ Please use `reframe()` instead.
#> ℹ When switching from `summarise()` to `reframe()`, remember that `reframe()`
#>   always returns an ungrouped data frame and adjust accordingly.
#> ℹ The deprecated feature was likely used in the maldipickr package.
#>   Please report the issue at <https://github.com/ClavelLab/maldipickr/issues>.

but current dplyr version used is 1.0.10

First release:

• usethis::use_cran_comments()
• Update (aspirational) install instructions in README
• Proofread Title: and Description:
• Check that all exported functions have @return and @examples
• Check that Authors@R: includes a copyright holder (role 'cph')
• Check licensing of included files
• Review https://github.com/DavisVaughan/extrachecks

Prepare for release:

• git pull
• urlchecker::url_check()
• devtools::build_readme()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• git push
• Draft blog post

Submit to CRAN:

• usethis::use_version('major')
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• Add preemptive link to blog post in pkgdown news menu
• usethis::use_github_release()
• usethis::use_dev_version(push = TRUE)
• Finish blog post
• Tweet

• MALDIquant and readBrukerFlexData both require a lot of maintenance
• the packages at RforMassSpectrometry have an improved backend
• implementing maldipickr with it could be relevant for long-term
• but still needs to import some of the MALDIquant routines rformassspectrometry/MsCoreUtils#119

I think this would improve readability. The README can become a simple .md file instead of Rmd

Source: https://pkgdown.r-lib.org/articles/customise.html?q=Get%20starte#navbar-heading

The clustering approach introduced in v1.1.0 implicitly use the single-linkage clustering (meaning friend-of-a-friends). This typically results in clusters in chain. The minimum similarity within the cluster in not controlled.

• does not necessary make sense if the workflow in handled by {targets}
• could be better file format alternative (e.g., qs)
• Look at: https://lifecycle.r-lib.org/articles/communicate.html to do it properly, (EDIT: more specifically for arguments: https://lifecycle.r-lib.org/articles/communicate.html#deprecating-an-argument-providing-a-new-default)
• will get rid of untested sections of the function (https://app.codecov.io/github/ClavelLab/maldipickr/blob/main/R%2Fprocess_spectra.R) as the CRAN is not happy when writing to disk during tests.

The packages {igraph} and {tidygraph} are conveniently imported for the dereplication but are only used to detect connected components.

There should be an easy way to recode an equivalent of the connected components algorithm, seeing as we only use limited options there.

This would certainly drastically reduce the number of dependencies!

Prepare for release:

• git pull
• Close v1.2.0 milestone
• Check current CRAN check results
• Polish NEWS
• urlchecker::url_check()
• devtools::build_readme()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• revdepcheck::revdep_check(num_workers = 4)
• Update cran-comments.md
• git push
• Draft blog post

Submit to CRAN:

• usethis::use_version('minor')
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• Add preemptive link to blog post in pkgdown news menu
• usethis::use_github_release()
• usethis::use_dev_version(push = TRUE)
• Finish blog post
• Tweet

two functions do not comply at the moment:

• similarity_to_clusters()
• identification_to_clusters()

could be unified to:

• delineate_clusters()

with two internal functions that are not exported?

pick_spectra

pick_spectra(clusters, discard_regex = "E11", only_show_discarded = TRUE)

#> Warning: Returning more (or less) than 1 row per `summarise()` group was deprecated in
#> dplyr 1.1.0.
#> ℹ Please use `reframe()` instead.
#> ℹ When switching from `summarise()` to `reframe()`, remember that `reframe()`
#>   always returns an ungrouped data frame and adjust accordingly.
#> ℹ The deprecated feature was likely used in the maldipickr package.
#>   Please report the issue at <https://github.com/ClavelLab/maldipickr/issues>.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.