Consolidation/harmonization TODOs

[csoneson]

To consider at some point:

• [x] Harmonize variable names (., _, camelCase)
• [x] Harmonize messaging levels (verbose usage) across functions
• [x] Harmonize use of ::. Let's try the following conventions:
  • make sure all used external functions are imported above the function using @importFrom
  • prefer to @importFrom the specific method (e.g. SparseArray::rowSums) instead of the generic (e.g. BiocGenerics::rowSums), to make it easier to see the exact method used
  • don't use the namespace:: prefixes in the function body
• [x] Consider removing obsolete(?) functions (sumOverRegions?)
• [x] Consider removing dependency on scuttle::aggregateAcrossCells (only used once in readBedMethyl)
• [x] Consider removing dependency on stringr::str_extract (only used once in getAnchorRegions)
• [x] Consider removing dependency on purr::map (only used once in plotRegion)
• [x] Add a NEWS file
• [x] Add examples to all exported functions
• [x] Run BiocCheck() and pkgdown::build_site() (to check that all functions are included in _pkgdown.yml)
• [x] Add the sticker 🙂
• [x] Revisit on what value to use for the modification probability of implicit non-modified bases (0 vs 0.02)
• [x] Review the use of >0 vs >=0 (depending on how implicitly given modification probabilities are represented)
• [x] in read-level-data.Rmd set eval=TRUE for chunk with (currently non-functional) rowMeans(NaArray)

[mbstadler]

Added the sticker in 1685845

[mbstadler]

Removed sumOverRegions() (superseded by getAnchorRegions()) in 2404a9a

[mbstadler]

added a NEWS.md file in a2616d7e

[mbstadler]

The use of :: has been harmonised according to the rules above in c38f3aec

[mbstadler]

Some first steps towards harmonising the messaging (verbose usage) in 161d23dd:

• a new utility function .message(...) that is mostly a drop-in replacement for message (although it doesn't support multi-part messages like in message("this is ", "a long ", "message in multiple parts"). It has the following advantages:
  • it automatically looks for a logical scalar verbose in the calling environment and only generates a message if verbose == TRUE
  • it is implemented via cli::cli_inform (see e.g. here) and thus supports pluralisation and inline-markup, for example .message("reading from {.file fname}") which adds a link to a file, or .message("processed {length(unique(x))} element{?s} of a total of {length(x)}")
• replace the use of if (verbose) { message(...) } by .message(...), possibly augmented using inline-markup

Having the messaging done by a central .message() would make it easy if we decide to, for example, change the styling of messages, or remove the verbose arguments in all functions in favour of a central config option.

Many of our dependencies already depend on cli (e.g. ggplot2), so this did not add any further dependencies. Furthermore, it could easily be used to also support inline-formatting in errors and warnings by using cli_abort or cli_warn instead of stop and warning, or any of the other features that cli offers (emojis, progress bars, etc.).

[mbstadler]

Removed some dependencies:

• purrr::map (replaced by dplyr::group_modify) in 5dd70bfa
• stringr::str_extract (replace by sub) in 54fae054

[mbstadler]

read-level-data.Rmd now evaluates all code chunks (currently it uses mean instead of rowMeans which is not yet implemented and probably will not be implemented in BioC 3.20), done in 8a8a0299.

[csoneson]

Revisit on what value to use for the modification probability of implicit non-modified bases (0 vs 0.02)

This is now set to 0 (no need to use an artificial non-zero value anymore with the change from SparseArray to NaArray). For readModkitExtract, the values provided in the input file (0) are used as they are.

[csoneson]

Harmonization of variable names done in 71a61d79 and ec0482f4