Serialize into a data.frame?

[karthik]

It would be nice to provide some helper functions to serialize the results into a data.frame, especially since the fields returned are often * (but not always) standard. For e.g.

Arrrr> library(aRxiv)
Arrrr> z <- arxiv_search(id_list = "1403.3048,1402.2633,1309.1192")
sapply(z, length)
Arrrr> sapply(z, length)
entry entry entry
   17    17    15

and fields returned are also not super consistent.

Arrrr> sapply(z, names)
$entry
 [1] "id"               "updated"          "published"        "title"
 [5] "summary"          "author"           "author"           "author"
 [9] "author"           "doi"              "link"             "comment"
[13] "journal_ref"      "link"             "link"             "primary_category"
[17] "category"

$entry
 [1] "id"               "updated"          "published"        "title"
 [5] "summary"          "author"           "author"           "author"
 [9] "author"           "author"           "author"           "author"
[13] "comment"          "link"             "link"             "primary_category"
[17] "category"

$entry
 [1] "id"               "updated"          "published"        "title"
 [5] "summary"          "author"           "doi"              "link"
 [9] "comment"          "journal_ref"      "link"             "link"
[13] "primary_category" "category"         "category"

This helper function could take a rbind.fill approach to get an even data.frame returned, or you could consult the API and get a complete list of field names and construct a standard data.frame into which search results can be coerced. Feel free to discard the idea -- just throwing out a suggestion.

[karthik]

The link fields could use some tidyr and get collapsed into a list within a data.frame (otherwise you'll repeat the other information across multiple rows). 🐙

[kbroman]

This is what I'm working on next. My plan was to combine things like the multiple author names into a single field, with | separators and then put it all into a data frame.
The links are particularly weird...there are abstract, pdf, and DOI links differentiated with different attributes. But I think I'm on top of this.

[kbroman]

With PR #7, I have arxiv_search() returning a data frame.

Maybe I should have had the columns like authors be lists within the data frame; instead, I pasted together to make a single string with | separators.