httr instead of Rcurl

[sckott]

httr is best practice package now for HTTP requests in R - RCurl is okay, but will probably not be maintained very well moving forward

[pdatascience]

Hi @sckott !

I am resuming development on the package which was frozen for a while to other urgent projects. I decided to start with using httr instead of RCurl but ran into difficulties right away. I think it is an authentication problem.

For example, I was trying to replace the line











RCurl::getURL( paste( options$server_url, "/protocol", sep = "" ), verbose = FALSE, userpwd = options$userpwd, httpauth = 1L)

which is in repository_information.R with the appropriate httr call and came up with the following (I've hard-wired the password for testing purposes for now - I will change it as soon as I troubleshoot the problem):

httr::GET( paste( options$server_url, "/protocol", sep = "" ) , authenticate( "obkms", "1obkms", "basic"), verbose() )

However, this does not seem to work, as it produces the following output:

> r = httr::GET( request_url , authenticate( "obkms", "1obkms", "basic"), verbose() )
-> GET /graphdb/protocol HTTP/1.1
-> Host: 213.191.204.69:7777
-> Authorization: Basic b2JrbXM6MW9ia21z
-> User-Agent: libcurl/7.47.0 r-curl/0.9.7 httr/1.2.0
-> Accept-Encoding: gzip, deflate
-> Cookie: JSESSIONID=92BCB4C5F5004C8CC4FBF673A02CD280
-> Accept: application/json, text/xml, application/xml, */*
-> 
<- HTTP/1.1 200 OK
<- Server: Apache-Coyote/1.1
<- Content-Type: text/plain;charset=UTF-8
<- Content-Language: en-US
<- Content-Length: 1
<- Date: Fri, 07 Oct 2016 08:32:09 GMT
<- 

I played around and if I enter the wrong password, I pretty much the same empty result. I also remember that when I was using the getURL function from RCurl before, I had to add the option
httpauth = 1L for it to work. As far as I understand, this means "basic authentication." Just to make sure I tried all the other types of authentication from the httr:GET manpage to no avail.

Do you have any idea where the problem might be coming from?

Best Vic

[sckott]

I think all you need to do is content(r) to get the output. The output from a call to any of the http verb functions (e.g., GET) is a object of class response, with the output, headers, status code, etc. Then you call content on it to get back outout. Best practice to usually get back plain text like content(r, as = "text") then parse the text according to what type of data it is (e.g., for xml use xml2 package, for JSON data use jsonlite package)

[pdatascience]

@sckott many thanks for that! I guess I should have RTFM more carefully before asking! Sry :)

[sckott]

no problem!