IRC log for #koha, 2009-08-11

← Previous day | Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
12:05 gmcharlt bbiab
12:20 chris_n g'morning koha
12:25 g'morn gmcharlt
12:25 gmcharlt hi chris_n
12:26 I survived - that's about all I'll say ;)
13:44 jwagner Someone around who understands zebra & relevance ranking?
14:17 hdl_laptop jwagner: throw your question.
14:18 jwagner If you do a search for Woolf (as in Virginia Woolf), the relevancy ranking seems to always put entries with Wolf, Wolves, etc. in the hitlist ahead of entries with Woolf.  I can see that zebra is trying to be helpful, but why doesn't it return an exact match first in the list?  Fuzzy searching is off.
14:19 Same results for Reed -- it returns Red first in the list, along with things like Rede.
14:20 If you search for "Virginia Woolf" in quotes, it brings up a reasonable list, but Virginia Woolf without the quotes seems to search for anything with Virginia plus any variation on wolf, wolves, etc.
14:27 owen I thought zebra didn't understand quotes.
14:27 jwagner No, it uses quotes to treat as an exact phrase -- I've done that with several searches.
14:28 It does choke on certain symbols, like ? and :
14:31 hdl_laptop jwagner: which value is AutoTruncation ?
14:31 jwagner: what is the value of queryStemming ?
14:32 jwagner QueryAutoTruncate and QueryStemming are both on.  Should that make a difference?  From the descriptions they both seemed to operate at the end of terms, like making cat and cats equal.
14:32 hdl_laptop If QueryStemming is 1 Then you are trying first to "stem" your operands and then search relevancy.
14:33 So i guess in english wolf is stemmed wol and searches on wol* before  doing the search.
14:33 (and relevancy ranking)
14:33 which might not be what you want.
14:33 jwagner Hmmm.  I just turned them both off, but it didn't make a difference.  A search for Woolf still produces the same large result set, with wolf and wolves near the top.
14:36 (This is the general keyword search, incidentally -- which 99% of users are going to do.  A subject search for Woolf produces a more appropriate result set.)
14:37 owen jwagner: I don't find that putting search terms in quotes excludes results which don't have that exact phrase.
14:37 http://acpl.kohalibrary.com/cg[…]ch.pl?q=space+jam
14:38 http://acpl.kohalibrary.com/cg[…]q=%22space+jam%22
14:39 jwagner OK, now we have another variation, sigh.
14:40 owen Could it be a zebra configuration issue?
14:40 I *know* I was told previously that zebra didn't do exact phrase searching via quotes.
14:40 jwagner That's what I'm wondering, but I don't know enough about zebra's inner workings to even know what I'm looking for.
14:41 The only changes I've done to record.abs on these systems are to add the 035 field (it was in record.abs but commented out) and add the 856 field.
14:43 owen, with regard to quotes, I've found that searching time and again and "time and again" produce different result lists, and the one with quotes is more accurate. Same for time again and "time again" (different title, again the one with quotes is more accurate).  I've seen it work the same way on other titles.
14:47 atz owen: zebra does it, but Koha pre-parsing strips it out
14:47 so in yaz client you can specify a search that is an exact string, or a truncated string, possibly even a left-truncated string, etc.
14:48 hdl_laptop jwagner: in fact, zebra has some relevancy ranking on the fly
14:48 jwagner atz, would the pre-parsing account for my problem?  Why variations on wolf, wolves show up in the hitlist before the term that was actually searched (Woolf)?
14:49 hdl_laptop And koha is  building the query.  
14:49 jwagner: maybe relevancy suppose that you are also using stemmed operands.
14:49 (in koha)
14:50 atz jwagner: i'm not sure... i would bet wol* shows up in more fields for the "wolf" records than just the one author field for Virgina's books
14:50 so that probably affects relevance
14:50 jwagner That might be the case.  Any way to tweak zebra in that regard?
14:50 hdl_laptop jwagner: it is in C4/Search.pm
14:51 not in zebra that you should look.
14:51 atz jwagner: like i said, you *could* search zebra that way.  koha doesn't
14:51 mostly because of having a fairly clunky pre-parser
14:56 jwagner hdl_laptop, I'm looking at C4/Search.pm.  Not quite sure what I'm looking for, but it seems to give 1st priority to title-cover, followed by exact title, then phrase title, then exact any, then word list any.  So since I'm searching one word (Woolf) which may show up in a title but isn't a phrase, is that part of the problem?
14:57 Sorry, exact any, then word list any are commented out, never mind those.
15:06 hdl_laptop jwagner: you are searching for QueryWeightFields
15:06 and its options.
15:10 jwagner hdl_laptop, thanks.  I think I'm going to have to turn that over to one of the programmers to research.  Sigh....
17:16 kf hi there, testing requests atm, I m able to put a item on hold, that is already checked out to me (current HEAD) - can somebody try this and tell me if its a bug in my configuration or a real bug?
18:01 chris morning
18:01 sekjal good morning, chris
18:03 chris heya sekjal, hows it looking for the migration?
18:04 sekjal I think it's going alright; discovered that I can't import ALL the records first, then run rebuild_zebra.pl.  Run out of space about 80% through the merge
18:04 once this latest load finishes, I've just got to get the patrons over, then the current circulation transactions, and I'm set
18:05 chris sweet, sounds like its going fine
18:05 (apart from the hiccup with diskspace)
18:06 sekjal yeah.  since some of my bibs have over 1000 attached items, I've got to run rebuild_zebra.pl  with the -x flag
18:06 chris ahh
18:06 those are serials?
18:06 sekjal yes
18:07 we've got a couple that go back to the 1800s, with several volumes per year
18:07 and multiple copies at different locations
18:07 chris medical journals eh?
18:07 sekjal yes.  Annals of the New York Academy of Sciences was the first one I noticed
18:08 chris my sister is a 'research fellow' in public health at the school of medicine here, she would be jealous :)
18:09 sekjal we do have a pretty good medical collection.  sadly, a lot of it is in storage now due to space constraints
18:09 atz sekjal: so as soon as you get done w/ the migration, you start a digitization project...
18:10 sekjal atz: sounds good to me.  our Digital Project's librarian is deciding on institutional repository software now
18:11 chris http://www.wnmeds.ac.nz/academ[…]ck%20profile.html  <--- thats her (she has finished her Phd now)
18:12 sekjal chris: very cool
18:12 atz PhD, eh?  cormacks not known for their slouching, apparently....
18:13 chris http://www.uow.otago.ac.nz/aca[…]onna_Cormack.html
18:13 the internet .. you can find anything
18:13 atz: yeah im the black sheep of the family hehe, i only have a BA, and a BSc
18:14 owen chris, you forgot to add "international fame"
18:15 chris infamy maybe?
18:18 http://www.fishpond.co.nz/Book[…]hor=Ian%20Cormack
18:18 my dad is more famous
18:20 owen "Getting the Buggers to Turn Up" Awesome title!
18:22 chris heh
18:27 He Pa Auroa is in Library of Congress's catalogue, i use it to test z3950 on koha usually :-)
18:29 find the router Sharon ?
18:29 Sharon yep, it was unboxed, but safely stored here in the office
18:31 chris :)
18:48 sekjal Zebra is really frustrating me.  Even doing it in two batches, I still get the same memory error halfway through the second
18:50 atz sekjal: further bifurcation is in order... or you can try to target the problem record, if it is a precise one
19:02 sekjal it seems to be a problem with the register's allocated memory
19:02 do I really have that big an index?
19:17 atz guess i can turn off the sprinkler now that it's pouring rain!
19:22 tomascohen hi, i sent an email to koha-devel asking about
19:22 setting up a dev environment for koha
19:22 atz eclipse, right?
19:22 tomascohen yeap
19:22 atz I'm replying right now...
19:22 tomascohen ah, ok
19:23 I'll wait for your reply
19:23 atz haven't used eclipse, but my main point was to get familiar w/ git
19:23 (if you haven't already)
19:24 tomascohen yeap, I've been using it
19:24 the main purpose of my question
19:24 was to kknow what pwoplw use for debugging
19:24 koha development
19:24 atz it depends what is being developed
19:25 i.e., web interface or AJAX features... or the installer, or reports, etc.
19:26 tomascohen i've been developing extra modules (perl + html::template)
19:26 and reports with google viz api
19:26 atz interesting
19:26 the most powerful tool I've used for perl debugging is NYTprof
19:27 (New York Time profiler)
19:27 *Times
19:28 it generates a ton of performance analysis down to how many picoseconds a given line takes to execute
19:28 tomascohen i see, interesting
19:29 is there a way to introduce breakpoints and inspect variables?
19:29 that's why i've been trying with EPIC
19:29 atz ah... that is more basic kind of debugging...
19:30 you can send data to the error log with statements like:
19:30 warn "var1: $var1";
19:30 to make it smooth, do:
19:30 use C4::Debug;
19:30 tomascohen yesm that's what i've seen commented out in some files
19:30 and used it
19:30 atz $debug and warn "var1: $var1";
19:31 tomascohen i just  thought koha devs might use more interesting methods
19:31 atz C4::Debug provides the variable $debug
19:32 and you can set that from the command-line like:
19:32 export DEBUG=1
19:32 tomascohen in eclipse, using EPIC you can run a Perl/CGI project
19:33 atz or in Apache VirtualHost like:  SetEnv DEBUG 1
19:33 tomascohen setting several environment variables
19:33 is not difficult to have a full development environment for this kind of projects
19:34 my doubt has to do with "the proper way" of working for this project
19:34 as I intend that our efforts
19:34 can be usefull for some anytime
19:34 atz the project does not necessarily constrain the types of tools you should use
19:35 you are welcome to use newer or more advanced tools (and perhaps demo them to the rest of us!)
19:36 tomascohen I'll try to share my settings for eclipse if I reach an usable config
19:37 well, have to leave or i'll be locked in my office till tomorrow!
19:37 thanks for your time
19:37 bye (atz)
19:37 atz np, ttyl
19:39 johnson hi all
19:43 owen Hi johnson
19:43 johnson a few months back I upgraded Koha for the small library I was working for at the time
19:45 i'm looking at taking a job now with a university in the US, and i'm hoping to be a shill for koha within the organization
19:45 +i'm looking at getting involved in development
19:45 so i'm just introducing myself
19:45 i might be around a lot in the future
19:46 gmcharlt johnson: great!
19:50 chris yay !!
19:51 sekjal: you can control the index size in the zebra config
19:52 memMax or something
19:52 sekjal chris:  there is that, and register
19:52 register seems to be what's throwing the error for me
19:53 chris ah right
19:53 sekjal though I probably should have increased memMax, too...
19:53 chris cant hurt to try
19:54 sekjal each failure costs about an hour right now... and I'm running low on those (tomorrow is my last day for two weeks)
19:55 chris :(
19:56 hdl_laptop hi johnson
19:56 welcome
19:56 chris how big are your 2 files, in megabyte size?
19:57 or you 2 directories .. if are using rebuild_zebra with -k
19:57 sekjal each somewhere between 200 and 300 MB
19:57 chris i have 12 hours of daylight coming up, if you wanted to email me them
19:57 johnson thanks hdl_laptop
19:58 chris so if it gets time to knock off work, and its still jamming
19:58 feel free to email them, and i can have a crack with my zebra
19:59 and now i have to go catch my bus to work
19:59 sekjal thanks, chris.  I really appreciate that.
19:59 chris bb in about 40 mins
19:59 sekjal ride safely!
20:09 richard hi
20:12 hdl_laptop good nite koha
20:15 Wizzyrea_ owen++ for posting all of his jquery hackery.
20:32 chris_n I think there's light at the end of the tunnel now
20:44 sekjal chris: I got my bibliographic records to index.  it was just a matter of adjusting the size of the register directory
20:44 still can't get my authority records to be searchable, though
20:45 but that's considered a lower priority right now
20:46 chris yay! for biblios boo! for authority
20:46 authorities are always keeping us down
20:46 sekjal its true
20:47 chris why is the man always messing with us
20:48 sekjal apparently, my problem is the field 001's aren't being properly recognized, and are resulting in undefined objects
20:49 chris ahh
20:49 sekjal doesn't matter if I bring them in as MARCXML from our old catalog, or fresh binary from NLM
20:50 gmcharlt sekjal: pure MeSH?
20:50 sekjal yes
20:51 gmcharlt sekjal: just loading using bulkauthimport?
20:51 sekjal yes
20:52 the database shows that authids have been assigned, and inserted into the MARCXML (and I think the binary MARC, too)
20:53 I just can't seem to retrieve the values from the field when it comes time
21:17 Jo morning all
21:17 chris hi Jo
21:22 sekjal day's end.  time to catch a train.  cheers, all!
21:35 Wizzyrea_ question: what order, if any, are the holdings tables sorted by
21:35 that is to say, what is the sort order of the items tables
21:35 we can't seem to figure it out
21:36 chris url ?
21:44 Wizzyrea_ cgi-bin/koha/catalogue/detail.pl
21:46 cgi-bin/koha/cataloguing/additem.pl
22:19 atz could be just the order they appear in the items table... not sure
22:22 gmcharlt I think atz has it
22:22 Wizzyrea_ yea,it makes more sense in the OPAC so I may just take a look at how that works and duplicate that on the staff side.
22:25 atz a few different things make sense... table order is actually pretty useful in additem, since it will be chronological order
22:25 and that makes sense to the adder of items
22:26 Wizzyrea_ yea, if you only have one library
22:26 it gets confusing, for example, with items like serials, where you have zillions of items
22:27 atz yeah, a lot of things are exploded by serials
22:27 Wizzyrea_ and 26 libraries
22:27 it's merely annoying with most bibs, but it's really obnoxious with a serial title
22:28 atz Might throw in a jquery control box for filtering the table display.  Like:
22:28 Items per branch:
22:28 MAIN - 3 (show only, hide)
22:29 AUX - 2 (show only, hide)
22:29 etc
22:29 Wizzyrea_ hmmm yes that idea has merit, especially since what they really want is to sort by logged in branch first
22:29 atz the nice thing about that jquery approach is that it doesn't have to build any kind of tablesorter index at load time
22:30 Wizzyrea_ that one would take me longer to figure out, but I'm willing to give it a go
22:30 (seems like an opportunity to learn something :P)
02:14 chris ack well that might be a hurdle, or more a brick wall
02:14 i wonder if anyone has tried to get biblios working with unimarc before
02:39 johnson chris: where do you recommend I start with getting familiar with Koha's architecture?
02:40 chris easiest thing, is pick something you want to work on, and jump in
02:41 johnson that's what i'll do, then
02:41 chris http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=3516
02:41 munin 04Bug 3516: enhancement, P5, ---, oleonard@myacpl.org, NEW, Display picture of patorns in OPAC
02:41 chris looks like a fun one :)
02:41 johnson break the thing a few times and see what happens
02:41 lamiette learn by doing :) always fun
02:42 chris bugs.koha.org has a ton of things people would like to see, or little (or big) bugs
02:42 and if you have a new feature you want to work on
02:42 its a good idea to register an bug, marked enhancement there
02:43 stops people doubling up
02:47 johnson yes, i've seen the bug reports
02:47 and the 3.2 roadmap
02:47 i'll just dig in
02:47 chris cool
03:21 Amit hi chris, brendan
03:21 good morning koha
03:21 hi jo, pie
03:43 brendan hey amit
04:17 harry potter
04:17 whoops -- wrong screen there
04:17 chris heh
04:17 brendan hi #koha
04:17 it's funny what amuses me these days :)
04:19 pie sometimes you gotta laugh at yourself, best way to get through the day :D
04:19 brendan :P
04:21 chris oh laugh at squirrels
04:21 http://www.urlesque.com/2009/0[…]steals-the-scene/
04:22 brendan priceless :)
05:58 anybody got a favorite online resource for RegEx (find and replace) ?
06:00 hdl_laptop which kind vi ? perl ?
06:01 brendan vi
06:01 hdl_laptop http://delicious.com/laurenthdl/vim
06:01 brendan thanks -- taking a look now
06:02 :)
07:14 chris evening
07:14 kf morning chris
07:14 can you test something for me? I m not sure if its a bug or a configuration problem
07:15 I cant place holds on titles with one item and allowOnShelfHolds OFF and OpacItemHolds OFF
07:15 I can place a hold, when I add a second item and check it out, so that both items are checked out
07:15 hdl_laptop hi
07:16 kf its current head - well, patched on friday
07:16 hi hdl
07:16 check in second item, delete it - no hold can be placed :(
07:21 chris hmm
07:21 kf hmm?
07:22 chris sure sounds like a bug
07:22 thats in the librarian interface eh?
07:23 kf no in Opac, sorry
07:23 librarian interface works fine
07:23 chris so if you only havfe one item, and its checked out, you cant place a hold
07:23 ?
07:24 kf yes
07:24 chris ok, ill check waht the code is doing
07:25 but i have to put kahu to bed first
07:26 kf thx chris - just thought you could test it (because you sure have a current head installation) - so that I can file a bug if you can repeat it
07:27 I dont want to steal too much from your time this late
07:50 chris right he's asleep
07:51 kf I have a short talk about dokuwiki in ten minutes - will be back soon (I hope)
07:54 ok, dont look at it
07:54 I tried everything I could think of yesterday to put this item on hold :( and now it works
07:54 with another user
07:54 I m sorry
07:56 seems to be a problem with current location and permanent location
07:56 will do more tests later
07:59 current location is borrowers location, but permanent isn't - so it may be a configuration problem
07:59 chris ah
07:59 :)
07:59 kf have to leave no for the meeting - but thx chris!
08:00 chris hehe i didnt do anything :)
08:01 hdl_laptop: you about?
08:02 hdl_laptop yes
08:02 chris have you ever tried biblios?
08:03 hdl_laptop tried for usmarc it is quite helpfull
08:03 chris yep
08:04 im doing some work for a UNIMARC library, and they wanted it integrated with koha, but i dont think it will work with unimarc without a lot of work
08:04 hdl_laptop but adapting to unimarc would be quite ressource demanding.
08:05 chris *nod*
08:05 hdl_laptop yes... many xsl are hardcoded.
08:06 And taking all the values from UNIMARC format would require much more time than simply processing UNIMARC to MARC21 back and forth.
08:06 chris yeah
08:06 hdl_laptop (nonetheless, MARC212UNIMARC is not roundtrip)
08:07 chris i think i will recommend that they dont at this time, unless they have the funds for the work to add unimarc to biblios
08:09 thanks, i just wanted to check i wasnt missing something simple
08:22 hdl_laptop or they could use opencataloger which is not as powerfull and friendly as biblios
08:23 but was developped in xul and well adapted for UNIMARC.
08:23 but still alpha and no devs.
08:24 chris that was what toins was working on eh?
08:33 hdl_laptop chris yes
08:40 kf back
08:49 chris: I just tested with borrower location = permanent location = current location - it works :)
08:50 chris yay :)
08:52 hdl_laptop tell gmcharlt SOPAC_integration branch updated
08:52 munin : tell gmcharlt SOPAC_integration branch updated
08:52 munin hdl_laptop: I suck
08:52 hdl_laptop @later gmcharlt SOPAC_integration branch updated
08:52 munin hdl_laptop: downloading the Perl source
08:52 hdl_laptop @tell gmcharlt SOPAC_integration branch updated
08:53 chris that will do it straight away
08:53 @later tell
08:53 munin chris: (later tell <nick> <text>) -- Tells <nick> <text> the next time <nick> is in seen. <nick> can contain wildcard characters, and the first matching nick will be given the note.
08:53 chris will do it next time the bot sees him do something
09:33 hdl_laptop later tell gmcharlt SOPAC_integration branch updated
09:37 chris heh, not having much luck :)
09:38 hands too :)
10:15 gmcharlt hdl_laptop: thanks
11:22 hdl_laptop: about?
11:34 @later tell hdl_laptop got a biblibre-sopac integration branch at git://git.librarypolice.com/git/koha-rm.git (also: http://git.librarypolice.com/?[…]s/biblibre-sopac)
11:34 munin gmcharlt: The operation succeeded.
11:34 hdl_laptop here.
11:34 gmcharlt hdl_laptop: hi
11:34 hdl_laptop hi gmcharlt
11:35 gmcharlt unless there's more pending in the queue for SOPAC, probably ready to push soon
11:35 and may be easier for Jean-Andr? to develop against that branch
11:35 hdl_laptop nothing more.
11:35 gmcharlt ok
11:36 main issues are - syspref (I just need the SQL script updates, I do the updatedatabase and slot it in)
11:36 the ilsdi.tmpl is still missing (did it ever exist?)
11:36 hdl_laptop do you want me to send you ?
11:36 gmcharlt yes
11:37 need documentation of Koha's implementation of the API
11:37 and lastly, documentation of how to hook up SOPAC to Koha
11:37 hdl_laptop I don't have any help  ilsdi.tmpl.
11:37 gmcharlt (whew!)
11:38 hdl_laptop but Kivutar wrote some page on how to connect to SOPAC.
11:39 Kivutar I did, but in french
11:39 gmcharlt Kivutar: that's OK - I'd like the link anyway
11:40 if nothing else, can running it through babelfish
11:41 Kivutar It is on our private wiki, so I should send you by mail
11:41 gmcharlt Kivutar: ok, thanks
11:48 Kivutar gmcharlt: I forgot I had it on my blog http://blog.kivutar.me/2009/06/10/0/
11:49 ask me if you need help
11:51 hdl_laptop gmcharlt: sysprefs do you need them to be sent via email ?
11:51 gmcharlt hdl_laptop: please
11:52 hdl_laptop doing this straight away
11:57 done.

← Previous day | Today | Next day → | Search | Index

koha1