IRC log for #koha, 2006-03-23

All times shown according to UTC.

Time S Nick Message
12:47 pierrick kados, are you there ? did you see Tumer proposal to solve Zebra/UTF-8 issue ?
12:47 kados pierrick: yes, reading it now
12:50 kados wow, that's interesting
12:50 kados so we definitely need to add utf-8 instructions to our zebra files
13:07 pierrick see you tomorrow, good afternoon to USA, good morning to kiwis, good night to France :-)
14:12 kados hi cindy
14:13 cm hi!  I finally made it.
14:13 kados cool :-)
14:13 kados firewall keep you out?
14:13 kados so here's a snip from your log:
14:13 kados QUERY : MPL at /usr/local/koha/intranet/modules/C4/SearchMarc.pm line 249.
14:13 kados no records found
14:13 cm no, me figuring out how to log in did.  :/
14:13 kados sorted at /usr/local/koha/intranet/modules/C4/SearchMarc.pm line 265.
14:13 kados hehe
14:14 kados so it looks to me like zebra's not finding any records
14:14 kados did you run bulkmarcimport.pl from HEAD or from rel_2_2?
14:14 cm I already had records in there.  So do I need to dump the db and do it again?
14:15 kados in that case you can run rebuild_zebra.pl
14:15 cm ah...I'll try that.  :D
14:15 kados with zebra, every change you make to the configuration files requires a reindexing
14:16 kados we're still working on making that less painful
14:16 kados under normal circumstances (ie over the command line) zebra can index 200K records in about 10 minutes
14:16 kados but when we do it one at a time with perl-zoom it takes much much longer
14:17 cm um, it died on line 28.  I'll try it a different way.
14:17 kados make sure the C4 for HEAD (not 2.2) is in your perl path
14:18 kados easiest way to do ensure that is to navigate to /path/to/cvs/head/koha
14:18 cm that would do it.  :)  
14:18 kados and run it from there
14:18 kados or you can go perl -I /path/to/cvs/head/koha /path/to/rebuild_zebra.pl
14:19 cm I'll try it that way.
14:21 cm it <I>still</I> did it.
14:22 kados huh
14:22 kados what's the exact error?
14:23 cm Died at /usr/local/koha-cvs-repository/HEAD/koh​a/misc/migration_tools/rebuild_zebra.pl line 28.
14:23 kados right ...
14:23 kados it didn't print anything about needint to specify 'confirm'?
14:23 kados perl -I /path/to/cvs/head/koha /path/to/rebuild_zebra.pl -c
14:23 kados try that
14:23 cm ok.
14:24 kados I see, someone forgot to print the options out when the script is run without -c
14:24 cm it said "Syntax OK"
14:24 kados cool, is it running?
14:25 cm how long should it take?  it went back to the command prompt immediately.
14:25 kados hmmm ...
14:26 kados try rebuild_zebra_idx.pl
14:26 kados same options
14:26 cm ok
14:26 kados wait, nevermind
14:26 cm same thing.
14:26 kados yea
14:27 kados we've been mainly testing with bulkmarcimport
14:27 kados systems that don't have records yet
14:27 cm i have 98,000 (so far)  :)
14:27 kados well, I mean don't have records in Koha
14:28 cm silly ;)
14:28 kados :-)
14:28 kados hdl: are you around?
14:31 kados cm: looking through the code it looks like rebuild_zebra.pl needs some updating to use our new modules
14:31 kados cm: lots of hard-coded stuff in there
14:31 cm oh, boy...
14:31 kados cm: I had assumed you'd be using bulkmarcimport.pl to get your data in to the plugin
14:32 kados cm: updating rebuild_zebra.pl should'nt take much more than an hour or so
14:32 kados cm: I'll see if I can fit it in tonight
14:32 cm cool!  we have all of mpl's circ data in there.
14:32 kados gotcha
14:33 cm thanks!
14:33 kados there's another procedure that Tumer's been testing that involves dumping out your records and then indexing manually
14:33 kados since you've already got your records in rel_2_2 that might make the most sense here
14:33 cm yeah, I saw his post.
14:34 cm possibly
14:34 kados it would certainly be the fastest route to take
14:34 cm maybe I'll try it this afternoon.
14:37 kados cm: make sure you back up your rel_2_2 database before you do anything :-)
14:37 kados rel_2_2 is very stable, but the plugin hasn't been fully tested
14:40 cm i am the queen of backups!  We're not using it in production yet anyway.  later!
14:43 kados heh
14:43 kados ciao
14:44 kados cm: I'll email you as soon as the rebuild_zebra.pl script's up to date
14:44 cm excellent!
17:03 shedges hi Fujitsu
17:04 Fujitsu Hi.
19:29 thd hdl: If you are still awake kados gave you a good answer to your question but I can amplify.
19:31 thd hdl: I had been working overnight yesterday so I needed to sleep and did not see your question until late this afternoon New York time.
19:32 thd hdl: If you are not about presently, I expect to be awake during your time zone tomorrow.
19:44 thd chris: are you about I have a simple question?
19:44 chris yep
19:46 thd chris: a value list called branches or something is automatically created when a user starts to fill branch information.
19:46 chris huh?
19:46 chris whats the context?
19:46 chris a user filling in branch information where?
19:47 thd chris: What is the name of the corresponding value list for items.booksellerid?
19:47 chris im sorry i dont know what you are asking me
19:48 chris is this in the marc editor?
19:48 chris adding items?
19:48 chris if so, i dont know, i didnt write that code
19:48 thd chris: In the admin module if I add a branch name in the form of BRANCH that name fills a value list which I can call from the MARC editor
19:48 chris right
19:48 chris thats pulled from the database
19:48 chris from the branches table
19:49 chris there is a table in koha called aqbooksellers
19:50 chris which stores supplier information, its used by the full acquisitions module (with does ordering and receiving)
19:50 chris you can add suppliers to it from full acquisitions
19:50 chris BUT afaik the MARC editor doesnt touch it
19:50 thd chris: I have identified that value list in the MARC framework for branches that can Identify the branch name
19:50 thd chris this is read only
19:51 chris yes
19:51 chris but, the marc editor doesnt look at that table to get the list of suppliers
19:51 chris it looks at the branches table to get the list of branches
19:51 thd chris: The editor can use a selection box to read branch names.
19:51 chris yes
19:51 chris it gets those names from the branches table
19:51 chris in the koha database
19:52 thd chris: so maybe if I called the aqbooksellers table I would get booksellers that had been entered?
19:53 chris yep
19:53 chris but there wont be any entered, unless you had gone to full acquisitions and entered some
19:53 chris heres how it gets the branches
19:53 chris my $sth=$dbh->prepare("select branchcode,branchname from branches order by branchname");
19:54 thd chris: for branches the curious thing is that the value list does not specify a column from the branches table just the table name in the bibliographic framework.
19:55 chris yep
19:55 chris cos the query is coded in additem.pl
19:56 chris additem.pl would have to be edited to know about aqbooksellers
19:56 chris and to fetch the data from it
19:56 thd chris: but not coded for aqbooksellers so adding the table name would not magically query the table without adding similar supporting code to additem.pl
19:57 chris exactly
19:58 chris unfortunately
19:59 thd thank you chris so it is as I would have expected.  The behaviour seemed magical when only examining the bibliographic framework code.
19:59 thd chris: It is good to know that there is code working hard to support the apparent magic
20:00 thd and where it could be changed to create more apparent magic
20:02 chris :)
05:14 hdl thd: are you there ?
05:14 thd yes
05:14 hdl about kados response ?
05:15 thd hdl: Did you have a specific question about it?
05:15 thd hdl: let me review it again in logbot
05:16 hdl my custommer agree to make 100$a linked with biblio.author.
05:17 hdl Is this what you usually do ?
05:19 hdl kados ? you rtried to catch me for rebuild zebra ?
05:20 hdl When you wake up, I'll be there.
05:20 hdl The fact is that usually, it lacks the db parameter.
05:21 hdl to the erbraidx commands.
05:21 hdl s/erbaidx/zebraidx/
05:21 thd hdl: yes that would be the most correct
05:22 thd hdl: 200 $f should not be the main entry for author in UNIMARC.  That is the nonstandard practise of one of your libraries.
05:23 hdl Is it ?
05:24 thd hdl: 100 is the main entry for personal author in MARC 21 which is equivalent to 700 in UNIMARC.
05:24 hdl thd: Pb : UNIMARC splits firstname and surname into two fields.
05:24 hdl thd: How would we gather information then ?
05:24 thd hdl: yes the UNIMARC framework may need correcting on that point
05:25 hdl And how would we be sure that ppl enter both 200$f and 700 field information. Since you know  the more ppl are used to a tool, the more they are lazy.
05:25 thd hdl: kados fixed the record editor so that plugins can fill several subfields in the same field.
05:25 hdl :)
05:25 hdl DID he ?
05:26 hdl It is for 3.0 or is it backported ?
05:26 thd hdl: yes there have been very many bug fixes
05:26 hdl Great !
05:27 thd hdl: he fixed it in rel_2_2.  I do not know if all the fixes have been applied to 3.0 yet
05:27 hdl Great... again :)
05:27 thd hdl: I have been too busy with 2.2 matters to actually set up 3.0 for testing yet
05:28 hdl about 008 marc-21 plugin.
05:28 thd hdl: we have a new MARC 21 default framework which has yet to be committed I think
05:28 hdl kados pointed a file in Attic directory. But I cannot find it on CVS.
05:29 hdl where is it ?
05:29 thd hdl: I can email it to you as an SQL script it is not perfectly perfected but is much better than the previous one
05:31 hdl thd: can you explain how  it works in your email ? henridamien at koha-fr dot org
05:31 thd hdl: it is well commented
05:33 thd hdl: I intend it to be still updated further and there were a few elements left of last but it should be perfectly functional
05:33 thd hdl it is 3 times the former size.
05:34 hdl good.
05:34 hdl Thx!
05:40 thd hdl: you should have it now
05:42 thd hdl: you do need a current rel_2_2 for it to work properly
05:45 thd hdl: aside from the few things left to do within the framework itself there are yet no subsequent frameworks for different media and the work on authorities support for MARC 21 is not yet complete or even working yet for building authorities.
05:46 hdl thd: Did you have a look at the build_authorities I sent you ?
05:47 hdl thd: It is usefull for a first import.
05:47 hdl s/import/creation/
05:47 thd hdl: yes but we did something wrong
05:48 thd hdl: is there something special about the first authority building?
05:49 thd hdl: how is first authority building different from subsequent authority building?
05:51 hdl thd: I said first import, because ppl could ask for an authority building "on-the-fly" (which is BAD and could not get implemented simply) When inputing for instance a 100 field, seeking in authorities automatically for the value, and if not seen, creating a new authority.
05:52 thd hdl: so the script is different for building authorities the first time?
05:57 thd hdl: or do you mean to say simply that authorities must be built from a script before they can be useful which I understand to be necessarily true?   At least, as long as we do not yet have the importation of a complete set of authorities records from BNF or LC.
05:59 hdl authorities can be built from biblio data (that is the build_authorities script aim)
05:59 hdl Or from a list that is provided. For instance MSC, on which I worked.
06:00 hdl But for each new authority list, things have to be designed.
06:00 thd hdl: yes I understand the building need currently.
06:00 hdl s/MSC/MSC2000/
06:00 thd hdl: what is MSC?
06:00 hdl It is AMS (American Mathematical Society) thesaurus
06:01 thd hdl: very nice
06:01 hdl in extenso : Mathematical Science Classification
06:03 thd hdl: really we need to import the as complete files with tracings and references for them to be more useful for finding the correct authorised form from searching an unauthorised form
06:05 hdl thd: I miss one word.
06:05 thd hdl: building from the biblios does not obtain the references and tracings when the biblios have only the authorised form
06:05 thd which one?
06:05 hdl thd: the as complete files, the what ?
06:06 thd hdl: I did not say that well but i meant the complete file of authority records from BNF, LC, etc.
06:07 thd If true authority importing were working
06:07 hdl Then it is not Zebra but google that hsould be used to dig into authorities :))))
06:07 thd as in bulkauthimport.pl
06:08 hdl Usually, ppl only need a very little part of these huge authorities.
06:09 thd hdl: the authorities concept is more intelligent than Google.  Google does not understand bibliographic records or even fielded search well.
06:09 hdl Then we should provide a service that allow them to import authorities notice, as Z3950 biblios.
06:09 thd hdl: yes
06:10 hdl maybe that with a zebra authority base, there could be a way for such things. Don't you think so ?
06:12 thd hdl: I helped kados: run a test for MARC 21 authorities building and it built the authorities but for some reason the links to the biblios did not function.  We made better progress with a previous test so I am not certain what had gone wrong.
06:13 thd hdl: we could get authorities import to work fine in rel_2_2 even but we needed to understand building first.
06:14 hdl sometimes, links fails because authoritise are not reinitalised first.
06:15 hdl did you remove all authorities in your base?
06:15 thd hdl: what is reinitialised?
06:16 thd hdl: I do not know if kados deleted the old ones.
06:16 hdl both authorities and links in biblios to authorities.
06:17 hdl I happened to get some link problem when rebuilding without removieng properly all authority link.
06:17 thd hdl: that is an option for the building script?
06:17 hdl -d normally.
06:18 thd hdl: maybe kados had not noticed the option or forgot it as he was the one doing the execution
06:19 thd hdl: he had decided to concentrate on the record editor problems which he had more confidence about fixing in time for release 2.2.6
06:23 thd hdl: UNIMARC 200 and MARC 21 245 are not authority controlled including not 200 $f and 245 $c respectively.
06:25 thd hdl: UNIMARC 200 and MARC 21 245 are for the title exactly as it appears on the book
06:26 thd hdl: UNIMARC 200 $f and MARC 21 245 $c are for the author's name exactly as printed on the title page
06:26 hdl thd: yes, but it can be a base for author authorities.
06:26 thd s/on the book/ion the title page/
06:28 thd hdl: the complete authorised form of the names is supposed to be in UNIMARC 7XX and MARC 21 1XX and 7XX
06:31 thd hdl: UNIMARC 200 $f and MARC 21 245 $c is for the unauthorised form exactly as it appears in the title page and may be different from the authorised form which is the most complete name obtained from copyright registration records and should usually include the authors dates to distinguish authors with the same name
06:32 thd hdl: nothing should stop you from building authorities from 200 $f if you want to but it is not standard practise
06:33 thd hdl: I understand that you have libraries which are not using 7XX
06:35 hdl thd: And we have libraries that donot use 7XX and then wnat to change to get 7XX work.
06:35 thd hdl: they may think that it is double work to put the name twice but there is supposed to be a difference in how the name appears between the authorised and unauthorised form
06:35 hdl thd: that's why we build authorities form what they input.
06:36 thd hdl: you have matched their records from BNF for upgrading when they wanted to use 7XX later have you not?
06:37 hdl thd: No.
06:37 thd hdl: you have upgraded records for other purposes?
06:37 hdl Sometimes, use of BNF file is more confusing than helping.
06:38 thd hdl: if Koha knows what to do with all the information well then it will be more helping :)
06:38 hdl thd: When I migrate data, we donot (unless required) propose to get data from BNF.
06:39 thd hdl: we will fix the confusion for all the data elements in bibliographic files for 3.X
06:40 thd hdl: why does osmoze want MARC 21 records in France?
06:41 thd hdl: Is it for his library that you have been asking about MARC 21 systems yesterday?
06:43 hdl Why do you think it is osmoze demand ?
06:44 hdl thd: It is for a library in New Caledonia.
06:44 thd hdl: did he not ask about converting records from UNIMARC to MARC 21?
06:45 hdl Not that I reckon.
06:46 hdl or maybe becaus of a data conversion from a Z3950 server that served MARC21 instead of UNIMARC.
06:47 hdl Oh yes, but not yesterday, was it last week or the week before.
06:48 thd hdl: New Caledonia has too much English influence to use UNIMARC?
06:49 hdl they use Reference Manager, which should be MARC-21.
06:50 hdl or I am mistaken.
06:50 hdl and most of their titles are english.
06:52 thd hdl: What is the prevalent language in New Caledonia?
06:53 hdl Normally French, since it is TOM
06:54 thd TOM?
06:55 hdl TOM : Territoire d'Outre Mer. Far away territories.
06:55 osmoze hello
06:56 thd yes I almost guessed with a little thought
06:56 hdl Two kinds of foreign countries are considered french : DOM and TOM DOM stands for Department, T for Territories.
06:56 thd hello osmoze
06:56 hdl hi osmoze.
06:57 hdl DOM are considered as part of France.
06:57 hdl TOM as possibily independant.
06:57 osmoze and TOM ?
06:57 osmoze ok
06:58 hdl Their relationship with France is weaker than DOM
06:58 hdl kados : I cannot see value_builder/marc21_008.tmpl file.
06:59 hdl Could I get it from somewhere ?
06:59 thd hdl: that should be in rel_2_2 unless I misspelt it
07:03 thd hdl: I have it on my system from maybe two weeks ago on CVS
07:04 thd hdl: It should be in the same directory as other plugins for rel_2_2
07:09 thd hdl: I see that you asked about the template.  Not the plugin.
07:13 thd hdl: /koha-tmpl/intranet-tmpl/npl/en/va​lue_builder/marc21_field_008.tmpl
07:14 thd hdl: you should use the NPL intranet templates as not all of the changes have been incorporated to default
07:16 thd hdl: and the things which are broken like auto-filling the call number in the item are broken in all templates now :)
07:38 thd hdl: I sent you the values for marc_subfield_structure_hidden
07:39 thd s/_hidden/.hidden/
08:22 pierrick hi #koha :-)
08:34 pierrick hello thd
09:30 hdl hi pierrick
09:31 pierrick hello hdl

| Channels | #koha index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary