Better error handling when specified raster size doesn't exist

[bschilder]

Describe the bug
get_phylopic gives an error about the URL length for certain species.

To Reproduce

s <- "mus musculus"
URL <- rphylopic::get_uuid(name = s,
                                       url = TRUE)  
png_data <- rphylopic::get_phylopic(uuid = names(URL),
                                                    format = "512")

Error in parse_url(url) : length(url) == 1 is not TRUE
12.
stop(simpleError(msg, call = if (p <- sys.parent(1L)) sys.call(p)))
11.
stopifnot(length(url) == 1)
10.
parse_url(url)
9.
is.url(url)
8.
stopifnot(is.url(url))
7.
build_url(parse_url(url)[c("scheme", "hostname", "port")])
6.
handle_name(url)
5.
handle_find(url)
4.
handle_url(handle, url, ...)
3.
GET(url = x)
2.
get_png(url)
1.
rphylopic::get_phylopic(uuid = names(URL), format = "512")

Expected behavior

If more than 1 URL is available, the function will automatically pick the first one.

If none are available, an error message will tell the user this.

Desktop (please complete the following information):

• OS: macOS Ventura 13.2.1
• rphylopic version: 1.0.0
• R Version: 4.2.1 (2022-06-23)

Session info

``` R version 4.2.1 (2022-06-23) Platform: x86_64-apple-darwin17.0 (64-bit) Running under: macOS Ventura 13.2.1

Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats graphics grDevices utils datasets methods base

other attached packages:
[1] shiny_1.7.4 orthogene_1.5.1 rphylopic_1.0.0 ggplot2_3.4.1 ggtree_3.6.2

loaded via a namespace (and not attached):
[1] utf8_1.2.3 gptstudio_0.1.0 devoptera_0.99.0 R.utils_2.12.2
[5] RUnit_0.4.32 tidyselect_1.2.0 RSQLite_2.3.0 AnnotationDbi_1.60.2
[9] htmlwidgets_1.6.2 grid_4.2.1 combinat_0.0-8 chatgpt_0.2.2
[13] devtools_2.4.5 munsell_0.5.0 codetools_0.2-19 ragg_1.2.5
[17] miniUI_0.1.1.1 withr_2.5.0 colorspace_2.1-0 Biobase_2.58.0
[21] TreeTools_1.9.1 knitr_1.42 rstudioapi_0.14 stats4_4.2.1
[25] ggsignif_0.6.4 MatrixGenerics_1.10.0 Rdpack_2.4 labeling_0.4.2
[29] GenomeInfoDbData_1.2.9 mnormt_2.1.1 optimParallel_1.0-2 topGO_2.50.0
[33] bit64_4.0.5 farver_2.1.1 rprojroot_2.0.3 coda_0.19-4
[37] vctrs_0.6.1 treeio_1.23.1 generics_0.1.3 clusterGeneration_1.3.7
[41] xfun_0.37 R6_2.5.1 doParallel_1.0.17 GenomeInfoDb_1.34.9
[45] rsvg_2.4.0 grImport2_0.2-0 bitops_1.0-7 cachem_1.0.7
[49] gridGraphics_0.5-1 DelayedArray_0.24.0 assertthat_0.2.1 promises_1.2.0.1
[53] scales_1.2.1 gtable_0.3.3 biocViews_1.66.3 processx_3.8.0
[57] phangorn_2.11.1 rlang_1.1.0 systemfonts_1.0.4 scatterplot3d_0.3-43
[61] rstatix_0.7.2 lazyeval_0.2.2 broom_1.0.4 BiocManager_1.30.20
[65] yaml_2.3.7 abind_1.4-5 ggimage_0.3.1.002 backports_1.4.1
[69] httpuv_1.6.9 RBGL_1.74.0 tools_4.2.1 usethis_2.1.6
[73] ggplotify_0.1.0 ellipsis_0.3.2 jquerylib_0.1.4 RColorBrewer_1.1-3
[77] BiocGenerics_0.44.0 sessioninfo_1.2.2 Rcpp_1.0.10 plyr_1.8.8
[81] base64enc_0.1-3 zlibbioc_1.44.0 purrr_1.0.1 RCurl_1.98-1.10
[85] ps_1.7.3 prettyunits_1.1.1 ggpubr_0.6.0 urlchecker_1.0.1
[89] fontawesome_0.5.0 S4Vectors_0.36.2 grr_0.9.5 fs_1.6.1
[93] here_1.0.1 magrittr_2.0.3 data.table_1.14.8 magick_2.7.4
[97] SparseM_1.81 R.cache_0.16.0 matrixStats_0.63.0 pkgload_1.3.2
[101] patchwork_1.1.2 mime_0.12 evaluate_0.20 xtable_1.8-4
[105] XML_3.99-0.14 jpeg_0.1-10 IRanges_2.32.0 compiler_4.2.1
[109] tibble_3.2.1 maps_3.4.1 crayon_1.5.2 R.oo_1.25.0
[113] htmltools_0.5.4 ggfun_0.0.9 later_1.3.0 tidyr_1.3.0
[117] aplot_0.1.10 expm_0.999-7 DBI_1.1.3 gprofiler2_0.2.1
[121] MASS_7.3-58.3 babelgene_22.9 Matrix_1.5-3 car_3.1-1
[125] piggyback_0.1.4 cli_3.6.1 quadprog_1.5-8 R.methodsS3_1.8.2
[129] rbibutils_2.2.13 parallel_4.2.1 igraph_1.4.1 GenomicRanges_1.50.2
[133] pkgconfig_2.0.3 numDeriv_2016.8-1.1 plotly_4.10.1 foreach_1.5.2
[137] bslib_0.4.2 stringdist_0.9.10 XVector_0.38.0 BiocCheck_1.34.3
[141] yulab.utils_0.0.6 stringr_1.5.0 callr_3.7.3 digest_0.6.31
[145] phytools_1.5-1 graph_1.76.0 Biostrings_2.66.0 rmarkdown_2.20.1
[149] fastmatch_1.1-3 tidytree_0.4.2 curl_5.0.0 lifecycle_1.0.3
[153] nlme_3.1-162 jsonlite_1.8.4 carData_3.0-5 OmaDB_2.14.0
[157] desc_1.4.2 viridisLite_0.4.1 fansi_1.0.4 pillar_1.9.0
[161] lattice_0.20-45 homologene_1.4.68.19.3.27 KEGGREST_1.38.0 fastmap_1.1.1
[165] httr_1.4.5 plotrix_3.8-2 pkgbuild_1.4.0 GO.db_3.16.0
[169] glue_1.6.2 remotes_2.4.2 png_0.1-8 iterators_1.0.14
[173] bit_4.0.5 sass_0.4.5 stringi_1.7.12 profvis_0.3.7
[177] blob_1.2.4 textshaping_0.3.6 memoise_2.0.1 dplyr_1.1.1
[181] ape_5.7-1

</details>

[willgearty]

It looks like the error here is due to there not being a 512pixel png available for that image: https://www.phylopic.org/images/c8f71c27-71db-4b34-ac2d-e97fea8762cf/mus-musculus.

[willgearty]

We should have better error handling in cases like these.

[LewisAJones]

Thanks for the catch @bschilder! As @willgearty has said, this is more to do with what is available on PhyloPic rather than the package itself. As a user, what would be your preferred handling in the event of these circumstances? An error being thrown or a warning and an svg (or other format) being returned by default? I can see value in both.

Thanks for being an active user!

[bschilder]

Thanks for being so responsive @willgearty @LewisAJones ! This makes sense to me.

Thanks for the catch @bschilder! As @willgearty has said, this is more to do with what is available on PhyloPic rather than the package itself. As a user, what would be your preferred handling in the event of these circumstances? An error being thrown or a warning and an svg (or other format) being returned by default? I can see value in both.

I can also see value in both approaches here. But in my particular case, I'd lean towards providing a warning message and returning NULL because I'm trying to gather images for many species within a parallelised loop.
https://github.com/neurogenomics/orthogene/blob/fa04f484de87f42dea5f9e69b1ee149b6f025c4f/R/get_silhouettes.R#L68

The warning + return NULL behavior would save me having to add a tryCatch (which is a rather sledgehammer-ish approach that I try to avoid bc it potentially obscures errors occurring for other reasons).

On that note, vectorizing the get_uuid and get_phylopic functions so that they can take multiple species at once and return the outputs as a list or data.frame would be a nice feature. But I'll leave it to you guys whether that's something you'd like to add.

[bschilder]

I should mention, I'd also be happy to make a PR with a version of my get_silhouettes function so it could be integrated directly into rphylopic, if you think that would be a useful feature.
https://github.com/neurogenomics/orthogene/blob/main/R/get_silhouettes.R

[keesey]

Given that images submitted in raster formats don't have a predictable set of sizes, I wonder if it would be better to pass minimum/maximum width/height/dimension? (Where dimension = max(width, height).)

(At some point I will be going back and vectorizing more silhouettes, but not for a while.)