bug: replace_publications_in_pub_index_md() behavior changed after cleaning

[mrustl]

Cleaning (#6) unfortunately introduced a bug. The function replace_publications_in_pub_index_md() now produces wrong results (content is linked to wrong rec_id and formatting is also not correct, see screenshot for rec_id=1045):

remotes::install_github("kwb-r/kwb.pubs@4ce47443f1812a96e033d5a927a4b307d2d5617b", upgrade = "always")
kwb.pubs::replace_publications_in_pub_index_md(md_paths = pub_md_paths, 
                                               endnote_db_refs = en_refs)

Data for bug-fixing:

• pub_md_paths: https://github.com/KWB-R/pubs/tree/improve-algolia/content/de/publication
• en_refs.zip

Please fix and verify that publication stays the same for all records!

[hsonne]

The problem was not in replace_publications_in_pub_index_md() but in my new function format_given() where I defined the arguments in the wrong order. This affected the content of en_refs when being modified by add_columns_to_endnote_db(). Thus, your uploaded version of en_refs (in en_refs.zip) was already "corrupt". This is fixed now. I used the following script to recreate the md files. However, I get differences due to the fact that I do not have the latest version of the Endnote db (I assume).

remotes::install_github("kwb-r/kwb.pubs@clean", upgrade = "never")

path_en_db <- "<path-to-my-local-endnote-dbs>/KWB-documents_20191205.Data/sdb/sdb.eni"

contents <- kwb.pubs::read_endnote_db(path_en_db)

en_refs <- kwb.pubs::add_columns_to_endnote_db(contents$refs)

en_refs$publication <- stringr::str_replace_all(en_refs$publication, pattern = '"', '\\\\"')
en_refs$publication <- sprintf("\"%s\"", en_refs$publication)

path <- "content/de/publication"
pub_md_paths <- dir(path, "md$", full.names = TRUE, recursive = TRUE)

#kwb.pubs::replace_dates_in_pub_index_md(pub_md_paths, en_refs)
#kwb.pubs::replace_publishdates_in_pub_index_md(pub_md_paths, en_refs)
kwb.pubs::replace_publications_in_pub_index_md(pub_md_paths, en_refs)

Can you do the verification test?

[mrustl]

I can verify that the issue is fixed now

library(kwb.pubs)
### using en_refs.Rds from zip file (see above)
en_refs <- readRDS("en_refs.Rds") %>% dplyr::select(- publish_datetime, - publication)
en_refs <- kwb.pubs::add_columns_to_endnote_db(en_refs)


en_refs$publication <- stringr::str_replace_all(en_refs$publication, pattern = '"', '\\\\"')
en_refs$publication <- sprintf("\"%s\"", en_refs$publication)

path <- "content/de/publication"
pub_md_paths <- dir(path, "md$", full.names = TRUE, recursive = TRUE)

#kwb.pubs::replace_dates_in_pub_index_md(pub_md_paths, en_refs)
#kwb.pubs::replace_publishdates_in_pub_index_md(pub_md_paths, en_refs)
kwb.pubs::replace_publications_in_pub_index_md(pub_md_paths, en_refs)

Could you add this step to the function kwb.pubs::add_kwb_style() ?

en_refs$publication <- stringr::str_replace_all(en_refs$publication, pattern = '"', '\\\\"')
en_refs$publication <- sprintf("\"%s\"", en_refs$publication)

[mrustl]

Hi @hsonne ,

seems we have one more issue to solve (EN ID: 675, 855,1163, 1164) -> https://github.com/KWB-R/kwb.pubs/files/5145961/en_refs.zip):

Maybe just a different behavior in case of missing data? I think your approach is there better than my simple one before!