amcat.upload.articles - Error Handling Emojis & more

[mrwunderbar666]

I had issues uploading a large set to AmCAT via amcat.upload.articles.

Repeatedly got the error message: unable to escape string. String is not utf8

I isolated the bug to the package rjson which has certain difficulties handling correct encoding of utf-8 (e.g. see here: ramnathv/rCharts#446).

You can reproduce the bug rather easily with:

library(rjson)
emoji_string <- "I 💗 AmCAT 💥"
json_emoji <- rjson::toJSON(emoji_string)

But if you use the alternative package jsonlite then the string converts correctly:

emoji_string <- "I 💗 AmCAT 💥"
json_emoji <- jsonlite::toJSON(emoji_string)

I created an alternative function that uses jsonlite as backend.

A second minor problem is that the function amcat.upload.articles requires a headline parameter, but AmCAT requires a title field. I am not sure if this is specific to our implementation of AmCAT (autnes.amcat.nl) or just a naming issue of the fields.

A pull request will follow soon.

[vanatteveldt]

I merged the pull request for now, but we should probably merge the two functions later.

[mrwunderbar666]

Thanks,

this also raises the question whether emojis are desirable in AmCAT in the first place. I am not sure, if elasticsearch can index these correctly and thus become searchable with a query.

If I get around it, I try to implement a feature that would transform emojis to their alternative text versions (e.g. 👍 becomes :thumbsup: