Re: [delicious-discuss] Py-Delicious - get_urlposts
That API is screen-scraping. We don't support that. Joshua On Aug 10, 2005, at 3:52 AM, Michael Foord wrote: Hello, Any of you Delicious'ers using the Python interface to the API ? (The mailing list over at belios.de is pretty quiet). It looks like the ``get_urlposts`` function has stopped working. The facility still exists with del.icio.us : http://del.icio.us/url/ + md5.md5(the_url).hexdigest() still returns the page of posts - but HtmlToPosts isn't extracting them anymore. Anyone got any ideas ? (short of fetching the rss version and parsing it myself :-) Best Regards, Fuzzy http://www.voidspace.org.uk/python ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss -- joshua schachter [EMAIL PROTECTED] ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss
Re: [delicious-discuss] Py-Delicious - get_urlposts
Not yet. My main worry here is that if we provide the API, people will just hammer away at it for every URL they know about. (the same problem exists on /url itself) I'm still looking for a good way to throttle these requests. Joshua On Aug 10, 2005, at 8:06 AM, Michael Foord wrote: joshua schachter wrote: That API is screen-scraping. We don't support that. Is there a way to obtain this information via the REST API ? I would like to know what categories users have put specific URLs in. Regards, Fuzzyman Joshua On Aug 10, 2005, at 3:52 AM, Michael Foord wrote: Hello, Any of you Delicious'ers using the Python interface to the API ? (The mailing list over at belios.de is pretty quiet). It looks like the ``get_urlposts`` function has stopped working. The facility still exists with del.icio.us : http://del.icio.us/url/ + md5.md5(the_url).hexdigest() still returns the page of posts - but HtmlToPosts isn't extracting them anymore. Anyone got any ideas ? (short of fetching the rss version and parsing it myself :-) Best Regards, Fuzzy http://www.voidspace.org.uk/python ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss -- joshua schachter [EMAIL PROTECTED] -- joshua schachter [EMAIL PROTECTED] ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss
[delicious-discuss] IT Conversations Podcast: Folksonomy - How I Learned to Stop Worrying and Love the Mess
Hi All, I just drop a quick note to tell the mailing list that IT conversation has just published a podcast with a panel on Folksonomy: In this dynamic panel from ETech 2005, Joshua Schachter (del.icio.us), Stewart Butterfield (Flickr), Jimmy Wales (Wikipedia) and Clay Shirky discuss several topics important to folksonomies. Surprising aspects of the implementation of tagging in various environments and approaches to balancing the needs of the system to the desires of the user are discussed from various viewpoints. Here's the link: http://feeds.feedburner.com/ITConversations-EverythingMP3?m=349 ciao, -- Nick ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss
Re: [delicious-discuss] Py-Delicious - get_urlposts
On 10 Aug 2005, at 16:17, Pete Freitag wrote: Another way to implement it would be to allow X number of connections from per IP per day (Yahoo!'s API's typically allow 5,000 requests per day). Then just keep a database table with the IP and number of connections for the day. If they excede the connections delay, and 503. This would probably perform ok because you could wipe the table clean every day. That would still entail a fairly hefty performance overhead of making one update to a database table per API request. My guess is that this contention would rapidly become a performance bottleneck. An alternative way of implementing the same idea might be to use LiveJournal's memcached software http://www.danga.com/memcached/. It acts as a very large in-memory hash table which can be queried by sending request over a socket. It supports expiration times on data, and has an atomic incr operation which could be used to keep a running total of the number of requests made in a finite time window. There's a Perl client, so it should be reasonably straightforward to hack some code into the start of each API request which incrs the count for that username (and/or IP address), and conditionally throws a 503 response code, with an appropriate Retry-After header if the client is being unreasonable. Richard. ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss
Re: [delicious-discuss] Py-Delicious - get_urlposts
Or you could setup some/one dedicated server for the api that had a throttled connection to the del.icio.us database. Can be done with: http://sqlrelay.sourceforge.net/ This seems neat. Any actual war stories? Joshua -- joshua schachter [EMAIL PROTECTED] http://del.icio.us/joshua ___ discuss mailing list discuss@del.icio.us http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss