Aargh!!
okay, so I've been wrestling with this script... it's blown up a couple of times after many many hours of work, and I've done a number of not-very-pristine-code bits to get it back up and running. I'm not good at babysitting a script.
Anyway, after getting home and seeing it dead and fixing it and rebooting it with toes crossed... I just discovered a logic failure in one of my four primary categories. "Former" indicates a user with plenty of activity on one or more Wiktionary languages, but on which the user has not edited within the past year. However, the script does not value recentness of editing activity over number of edits - so someone who has been editing (albeit rarely) on a language but was formerly active on a different language will be listed as formerly.
That's more of a value judgement than a logic problem, I guess. It's just that I personally would value recentness over number of edits.
en.WN
Showing posts with label PHP. Show all posts
Showing posts with label PHP. Show all posts
Sunday, February 14, 2010
Saturday, February 13, 2010
Nothing quite like working the hell out of a new piece of code.
I've been casually working on a php mediawiki API class for, uhm, months. But got motivated this past week and rebuilt it a lot. Decided to hammer out a quick proof-of-concepts today and ended up with some junk code which will need to be rewritten.
On the other hand, right now it's chewing through a 1013 long list of names, checking the contributions of each on up to 171 different Mediawiki website, and for each site it's creating an instance of my class, connecting and logging in, requesting the name's contributions, parsing them if any, logging out, and moving on to the next site (assuming the user has not made enough contributions.) With so many website connections and communications the script is not quick, and each user can take several minutes.
I think I can safely say the log-in and user contributions routines are good.
On the other hand, the editing routines haven't been touched yet. They don't get any use until after working through the list of users the third time (first time builds the list of unique voters, the second time builds the list of users ordered by how active they are on Wiktionary projects, third time builds the wiki syntax for writing to the wiki.) By my estimations, three and a half days from now it will (finally) get done. That's a pessimistic guess though.
So I'm just watching the terminal as it ever-so-slowly scrolls along... we're all the way up to the An* usernames.
I've been casually working on a php mediawiki API class for, uhm, months. But got motivated this past week and rebuilt it a lot. Decided to hammer out a quick proof-of-concepts today and ended up with some junk code which will need to be rewritten.
On the other hand, right now it's chewing through a 1013 long list of names, checking the contributions of each on up to 171 different Mediawiki website, and for each site it's creating an instance of my class, connecting and logging in, requesting the name's contributions, parsing them if any, logging out, and moving on to the next site (assuming the user has not made enough contributions.) With so many website connections and communications the script is not quick, and each user can take several minutes.
I think I can safely say the log-in and user contributions routines are good.
On the other hand, the editing routines haven't been touched yet. They don't get any use until after working through the list of users the third time (first time builds the list of unique voters, the second time builds the list of users ordered by how active they are on Wiktionary projects, third time builds the wiki syntax for writing to the wiki.) By my estimations, three and a half days from now it will (finally) get done. That's a pessimistic guess though.
So I'm just watching the terminal as it ever-so-slowly scrolls along... we're all the way up to the An* usernames.
Blog roll
-
-
Review: In My Father's Country13 years ago
-
A piece of cake15 years ago
