Transferring Disqus comments (from export) to Hashover
q2apro opened this issue · comments
It would be interesting if somebody has written a converter that reads the disqus XML scheme and arranges the data to the hashover XML.
Example Disqus:
<thread dsq:id="568987623">
<id />
<forum>echteinfachtv</forum>
<category dsq:id="1298300" />
<link>http://www.test.com/trigonometrie/</link>
<title>Page Sinus und Kosinus </title>
<message />
<createdAt>2012-02-08T17:02:32Z</createdAt>
<author>
<email>mail@test.com</email>
<name>Super User V</name>
<isAnonymous>false</isAnonymous>
<username>superuserv</username>
</author>
<ipAddress>44.430.49.124</ipAddress>
<isClosed>false</isClosed>
<isDeleted>false</isDeleted>
</thread>
Example Hashover:
<?xml version="1.0" encoding="UTF-8"?>
<comment likes="0" notifications="yes" ipaddr="">
<name>Marko</name>
<passwd></passwd>
<email></email>
<encryption></encryption>
<website></website>
<date>01/02/2015 - 12:00am</date>
<body>Geniale Seite, danke dafür!</body>
</comment>
Note: I am still sure, that date tag of hashover should be changed to the international format, e.g. 2012-02-08T17:02:32Z.
Furthermore, comparing the data from disqus, maybe the IP should be saved as well in the xml?
This is planned.
IP addresses can be stored by setting the $ip_addrs variable in settings.php to 'yes'
I have imported comments from Disqus using the following (dirty) code:
https://gist.github.com/jacmoe/e9d1ffdbd10c6022f71e394458c46da6
I assumes that the body of the comments is clean - which it should be, since it is from Disqus.
The script saves the comments as xml files named after the timestamp - then it is a small matter of sorting them and renaming them 1, 2, 3, ..
I hope it will serve as a starting point :)
Also, a proper encryption key (from settings) should be inserted into it :)
I am running the script in the shell.
The script does not check if the thread url is wrong - which it can be, if you accidentally use the Disqus code on localhost or other sites (forgot to change the script) - I just went through the exported xml and removed the erroneous threads.
If a thread url is wrong, it usually means that there are duplicate thread id's, which will (maybe) create problems.
Anyway: I hope this helps.
Thank you for the code @jacmoe!
There will be a way to import Disqus comments into HashOver eventually, but the format of the HashOver comment files/database needs to be finalized before the import code can/should be written.
In HashOver 2.0 some progress has been made in allowing a script to "emulate" posting a comment, so as long as there is a way to sort the Disqus comments than they will automatically be imported in the proper 1, 2, 3 naming scheme. This work isn't done, though, it needs to be made easier to allow importing comments from any other kind of comment system.