Opaque Pagination

[dtod]

I had not run into this before but was doing some development in Beta against the Enrollments REST API and it wasn't working because I was not using the opaque pagination headers. See Canvas API Change Log 

Should be easy(ish) to adapt, but just wanted to note it for the benefit of others.

[matthew_buckett]

Thanks for the heads up, I've been using the pagination on courses and find the parseable URLs useful so hope they don't move to opaque bookmarks there as well. For other's reference, this is what the headers now look like on beta:

$ ccurl -s -D -  'https://universityofoxford.beta.instructure.com/api/v1/courses/2598/enrollments?per_page=1'  | grep '^link'
link: <https://universityofoxford.beta.instructure.com/api/v1/courses/2598/enrollments?page=first&per_page=1>; rel="current",<https://universityofoxford.beta.instructure.com/api/v1/courses/2598/enrollments?page=bookmark:WyJTdHVkZW50RW5yb2xsbWVudCIsIkJ1Y2tldHQsIE1hdHRoZXciLDE0NzI5MF0&per_page=1>; rel="next",<https://universityofoxford.beta.instructure.com/api/v1/courses/2598/enrollments?page=first&per_page=1>; rel="first"‍‍

where as this is how they look on test at the moment:

$ ccurl -s -D -  'https://universityofoxford.test.instructure.com/api/v1/courses/2598/enrollments?per_page=1'  | grep '^link'
link: <https://universityofoxford.test.instructure.com/api/v1/courses/2598/enrollments?page=1&per_page=1>; rel="current",<https://universityofoxford.test.instructure.com/api/v1/courses/2598/enrollments?page=2&per_page=1>; rel="next",<https://universityofoxford.test.instructure.com/api/v1/courses/2598/enrollments?page=1&per_page=1>; rel="first",<https://universityofoxford.test.instructure.com/api/v1/courses/2598/enrollments?page=12&per_page=1>; rel="last"‍‍

[dtod]

I have no idea if it's any good, but it works. Here's my code for handling it (linux shell curl). The loop generates a $body and $header variable.

IFS_BAK=$IFS
IFS=$'\n'

head=true
#echo $result | while IFS= read -r line
for line in $result
do
    # echo $line
    if $head; then
        if [[ $line = $'\r' ]]; then
            head=false
        elif [[ "$line" == *"Link:"* ]]
        then
            header="$header"$'\n'"$line"
        fi
    else
        body="$body"$'\n'"$line"
    fi
done

# return delimiter to previous value
IFS=$IFS_BAK
IFS_BAK=

nextLink=$(echo $header | grep 'Link:' | sed -n -E 's/Link:.*<(.*?)>; rel="next".*/\1/p')
‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

[matthew_buckett]

Thanks David, if you output the headers to a temporary file:

curl -D tmpfile‍

the you can just read the temporary file and extract them out of there with this (my Link headers wasn't capitalised so I've adjusted sed for that):

nextLink=$(cat $headers | sed -n -E 's/^[Ll]ink:.*<(.*)>; rel="next".*/\1/p')‍