Time |
S |
Nick |
Message |
11:13 |
|
hdl |
kados : where is the database name entered in your koha.xml ? |
11:14 |
|
hdl |
With Zconnauth it was supposed to be authorities. |
11:14 |
|
hdl |
But now. I cannot find where to get name from with new Zconn. |
11:15 |
|
hdl |
sorry. |
11:15 |
|
hdl |
I found it. |
11:58 |
|
paul |
CVS sucks... |
17:36 |
|
hdl |
kados around ? |
17:36 |
|
kados |
hdl: yep |
17:37 |
|
hdl |
in AUTH count Usage. |
17:37 |
|
hdl |
Tumer did great, as usual. |
17:38 |
|
hdl |
But I cannot find out why Controlled-Subject-Index seems not to be used. |
17:39 |
|
hdl |
Even when I try and create a new att such as : Koha-Auth-Number in bib1.att and use it in record.abs, it is not used. |
17:39 |
|
kados |
hmmm |
17:39 |
|
hdl |
Should I do something ? |
17:39 |
|
kados |
yes |
17:39 |
|
kados |
you need to define it in the cfg file |
17:40 |
|
kados |
zebra-auth.cfg |
17:41 |
|
kados |
hdl: in zebra-biblios.cfg we have: |
17:41 |
|
kados |
recordId: (bib1,Local-number) |
17:41 |
|
hdl |
You know : it is to count the number of biblios attached to one Authority. |
17:41 |
|
kados |
hdl: does it make sense? |
17:41 |
|
hdl |
I don't think so. |
17:41 |
|
kados |
hmmm |
17:41 |
|
kados |
I dont' understand then |
17:41 |
|
kados |
what is Koha-Auth-Number? |
17:41 |
|
kados |
is it the authid? |
17:41 |
|
hdl |
You have biblios |
17:42 |
|
hdl |
In biblios, you use authorities. and $9 stores authid for fields linked to authorities. |
17:42 |
|
hdl |
right ? |
17:43 |
|
hdl |
I want to count how many biblios use authid number 1234 |
17:43 |
|
hdl |
So I try and do find @attr 1=Koha-Auth-Number 1234 |
17:44 |
|
hdl |
But it always returns 0 |
17:44 |
|
kados |
that syntax is wrong |
17:44 |
|
kados |
in bib1.att, what attribute is Koha-Auth-number linked to? |
17:47 |
|
kados |
hdl: ? |
17:47 |
|
hdl |
8910 |
17:47 |
|
hdl |
sorry |
17:48 |
|
hdl |
not too wrong. |
17:48 |
|
hdl |
I managed som searches with this syntax. |
17:48 |
|
hdl |
But anyway. |
17:48 |
|
kados |
f @attr 1=8910 1234 |
17:48 |
|
kados |
try that |
17:49 |
|
hdl |
find @attr 1=8910 1234 returns 0 count. |
17:49 |
|
hdl |
(I already did) :) |
17:50 |
|
hdl |
I am OK for a 0 result but not for used authorities. |
17:56 |
|
hdl |
hi tumer. |
17:56 |
|
dewey |
well, hi tumer is still strugling |
17:57 |
|
tumer |
hdl still awake? |
17:57 |
|
hdl |
tumer : yes. |
17:57 |
|
hdl |
I was waiting for you :) |
17:57 |
|
tumer |
did you sort authorities? |
17:57 |
|
hdl |
Quite. |
17:57 |
|
hdl |
remains some "minor" problems. |
17:57 |
|
tumer |
any questions? |
17:58 |
|
tumer |
shoot |
17:58 |
|
hdl |
Yes AUTHcountUsage : |
17:58 |
|
hdl |
You use GILS attribute 2057 |
17:58 |
|
hdl |
ControlledSubjectIndex. |
17:58 |
|
tumer |
at devweek yes |
17:58 |
|
hdl |
I tried this on my authorities and it always returns 0 |
17:59 |
|
hdl |
on HEAD : what do you use ? |
17:59 |
|
hdl |
I explain what I did : |
17:59 |
|
tumer |
have to check but that does not matter |
18:00 |
|
hdl |
In record.abs I declared 686$9 as a Controlled-Subject-Index |
18:00 |
|
hdl |
Controlled-Subject-Index:p and :s |
18:01 |
|
hdl |
But then trying to make a search on one : it miserably fails. |
18:01 |
|
tumer |
i am just checking the code |
18:02 |
|
hdl |
I checked in GILS and it was OK. |
18:02 |
|
hdl |
gils.att I mean |
18:03 |
|
tumer |
you have to have the index (whatever it is) where the authid resides in MARC record to work |
18:04 |
|
tumer |
so 686$9 is authid is it? |
18:05 |
|
tumer |
the logic is in record.abs you have defined an index for authid to reside for biblio |
18:06 |
|
tumer |
than AUTHcountusage just counts the authid in biblio |
18:06 |
|
hdl |
how can I define an index for authid ? |
18:06 |
|
tumer |
just a sec |
18:07 |
|
hdl |
686$9 for instance. |
18:07 |
|
hdl |
But there can be manymore : |
18:07 |
|
hdl |
600$9 631$9 or 701$9 702$9 710$9 |
18:08 |
|
tumer |
in record abs i use melm 686$9 Authority-format-id (not gils by the way) |
18:09 |
|
tumer |
i also use 001 Authority-format-id in authorities |
18:09 |
|
hdl |
I saw Authority/format-id |
18:09 |
|
hdl |
Is Authority-format-id a new attribute YOU defined ? |
18:10 |
|
tumer |
but whatever you use gilss etc does not matter as long as you search on that -- yes Authority/format-id is correct |
18:11 |
|
hdl |
what if authority-format-id is in many $9 subfields in biblios. Is it ok for index ? |
18:11 |
|
tumer |
yes i do 650$9 100$9 etc its ok |
18:11 |
|
hdl |
I can now well search on strings but it seems that cannot search on int. |
18:11 |
|
hdl |
Is there something to do ? |
18:12 |
|
tumer |
yes you have tou use n instead of w or p |
18:12 |
|
hdl |
Do I have to reindex the whole file before making a search ? |
18:12 |
|
tumer |
when indexing numeric fields u say Authorityformat-id:n |
18:13 |
|
tumer |
and yes have to reindex if anything changed in record.abs |
18:14 |
|
hdl |
does adm-reindex biblios do the job ? |
18:14 |
|
hdl |
(in yaz-client) |
18:15 |
|
tumer |
you have to use zebraidx update .... |
18:15 |
|
tumer |
i do not know adm-reindex |
18:16 |
|
tumer |
do you have biblios indexed with authid in them? |
18:17 |
|
hdl |
normally, yes |
18:18 |
|
hdl |
how can we know how many non-null given subfields there is in our database ? |
18:18 |
|
tumer |
in yaz you can search find @attr 1=Authority/format-id "authid" where you put some value to authid |
18:20 |
|
hdl |
if I want to count biblios which have not-null 686$9, is there a way ? |
18:20 |
|
tumer |
if you say find @and @attr 1=_ALLRECORDS @attr 1=title "" it should give you all those with no title subfield |
18:21 |
|
tumer |
check the zebra manual for _ALLRECORS for correct syntax and you will get it |
18:24 |
|
tumer |
find @not @attr 1=_ALLRECORDS @attr 2=103 "" @attr 1=Title @attr 2=103 "" |
18:25 |
|
kados |
hdl: you need zebra 2 to use ALLRECORDS |
18:25 |
|
tumer |
find @not @attr 1=_ALLRECORDS @attr 2=103 "" @attr 1=Controlled-subject-index @attr 2=103 "" should give you all with null values (since you used controlled-subject-index) |
18:25 |
|
tumer |
aha yes |
18:26 |
|
tumer |
i suggest you upgrade to 2 |
18:26 |
|
tumer |
its better and more stable |
18:26 |
|
hdl |
Is it seamless ? |
18:26 |
|
tumer |
seamles?? |
18:26 |
|
tumer |
xcues my english |
18:27 |
|
hdl |
Shall I have to throw all abs att files away ? |
18:27 |
|
tumer |
no all stays the same |
18:27 |
|
hdl |
is there a change in API ? |
18:27 |
|
tumer |
nope more features |
18:28 |
|
hdl |
yaz searches for authid still fails. |
18:29 |
|
tumer |
what is your index name? |
18:29 |
|
tumer |
controlled-subject is gils so you have to give that |
18:29 |
|
tumer |
ie. @attrset GILS |
18:31 |
|
tumer |
kados how did going live is going? |
18:32 |
|
kados |
tumer: so far so good |
18:32 |
|
kados |
tumer: I've discovered some holes in my API |
18:32 |
|
kados |
tumer: that I'm patching up now |
18:32 |
|
kados |
tumer: but on the whole, people are very happy |
18:32 |
|
tumer |
best of luck |
18:32 |
|
kados |
thanks |
18:32 |
|
hdl |
I did that |
18:33 |
|
tumer |
now and again check head as i may have filled those gaps foe head |
18:33 |
|
hdl |
find @attr GILS 1=2057 160545 |
18:33 |
|
hdl |
OK tumer : thanks. |
18:33 |
|
hdl |
I will give it a try. |
18:34 |
|
hdl |
Is your new model useable ? |
18:34 |
|
tumer |
hdl:dows it say blooming failure or just 0 results? |
18:34 |
|
hdl |
simple 0 results. |
18:35 |
|
tumer |
and you sure you have 160545 in records? |
18:36 |
|
hdl |
yes. |
18:36 |
|
tumer |
new model meaning <koharecord>? its in production with us |
18:36 |
|
tumer |
very happy with it |
18:37 |
|
tumer |
we now have a union catalogue with biblios only where no items reside as well |
18:37 |
|
hdl |
I have at least 1 biblio in my catalogue |
18:37 |
|
hdl |
items are in an other zebra base ? |
18:37 |
|
tumer |
what does record.abs say ? |
18:38 |
|
hdl |
Sorry ? |
18:38 |
|
hdl |
what does record.abs say about what ? |
18:38 |
|
tumer |
no one <koharecord> zebra for internal and opac use another for union cataliogue biblios only |
18:39 |
|
tumer |
record.abs for biblios: melm 686$9 |
18:39 |
|
hdl |
melm 686$9Controlled-Subject-Index,Controlled-Subject-Index:n |
18:39 |
|
hdl |
melm 686$9Koha-Auth-Number,Koha-Auth-Number:n |
18:39 |
|
hdl |
|
18:40 |
|
tumer |
ok try @attr 1-2057 @attr 4=109 160545 |
18:40 |
|
tumer |
@attr 1=2057 that is |
18:41 |
|
hdl |
unsupported use of attribute. |
18:41 |
|
hdl |
I definitely need to update to zebra 2.0 |
18:41 |
|
tumer |
which attribute? |
18:41 |
|
hdl |
4=109 |
18:42 |
|
hdl |
no 2057 |
18:42 |
|
tumer |
so are you sure Controlled- is 2057? |
18:42 |
|
hdl |
it did not tell such a thing before. |
18:43 |
|
hdl |
I tried many times @attr 1=2057 and it silently failed. |
18:43 |
|
hdl |
No Blloming failure |
18:44 |
|
tumer |
use @attr 1=Koha-auth-Number @attr 4=109 |
18:44 |
|
hdl |
idzebra/tab/gils.att:att 2057 Controlled-Subject-Index |
18:44 |
|
tumer |
forget numbers use names |
18:46 |
|
hdl |
No success. |
18:47 |
|
tumer |
definitely not indexed than |
18:47 |
|
hdl |
I cannot do good job any longer. |
18:47 |
|
hdl |
I will re-index tommorrow |
18:47 |
|
tumer |
is it open for me to connect to zebra with yaz? |
18:48 |
|
hdl |
no. |
18:48 |
|
tumer |
k |
18:48 |
|
hdl |
Not at the moment. |
18:48 |
|
hdl |
High time to go to bed. |
18:48 |
|
tumer |
and you can search other things on it? |
18:48 |
|
tumer |
good nigth |
18:48 |
|
hdl |
yes. |
18:49 |
|
hdl |
titles and Authors are searchable. |
18:49 |
|
hdl |
THANK YOU. |
18:49 |
|
tumer |
email me yor record.abs and one record at tgaripneu.edu.tr |
19:05 |
|
hdl |
tumer : I sent you my whole catalogue, 8.6 Mo. Sorry |
19:06 |
|
tumer |
probably wont be able to get it due to proxy |
02:55 |
|
hdl |
hi |
03:46 |
|
slef |
hi all, hi alex |
03:46 |
|
alex |
hi there |
03:46 |
|
slef |
I want my broadband back :-/ |
03:47 |
|
alex |
dialup? |
03:47 |
|
slef |
aye |
03:47 |
|
slef |
at least I have a telephone line since yesterday... only took 3 weeks :-/ |
03:48 |
|
slef |
telephone network monopolies, doncha luv em? |
03:48 |
|
alex |
mmmmm |
03:53 |
|
alex |
so I'm considering my debian packaging... |
03:54 |
|
alex |
thinking I'll just dump the bulk of the install in /usr/share/koha |
03:55 |
|
slef |
should we use /usr/lib/cgi-bin/koha? |
03:55 |
|
slef |
and templates and stuff into /usr/share/koha |
03:59 |
|
alex |
presumably we'd need /usr/lib/cgi-bin/koha/opac and /usr/lib/cgi-bin/koha/intranet |
03:59 |
|
slef |
sounds right |
04:00 |
|
alex |
I don't know, I suppose putting cgis under /usr/lib/cgi-bin is more correct? |
04:00 |
|
alex |
Is there debian policy about this? |
04:00 |
|
slef |
there's a webapp policy, but I think it's only draft... think I emailed you or list with a link to it |
04:01 |
|
slef |
there's also a perl policy linked from www.debian.org/devel/ |
04:01 |
|
slef |
which I suppose we should follow but IIRC it's pretty liberal |
04:01 |
|
alex |
I guess we should really put the /usr/local/koha/intranet/modules/C4/ somewhere appropriate |
04:01 |
|
slef |
in summary: webapp policy is draft, so we could do whatever for now |
04:02 |
|
alex |
yeah I looked at the webapp policy but it was fairly barbones. |
04:02 |
|
slef |
ah yeah, C4 should follow perl policy |
04:02 |
|
btoumi |
hi all |
04:02 |
|
slef |
hi btoumi |
04:02 |
|
btoumi |
hi slef |
04:04 |
|
alex |
just looking through install.pl |
04:04 |
|
alex |
perl stuff |
04:04 |
|
alex |
find my domain name |
04:05 |
|
slef |
aye, there's probably no good reason to do that in perl for the deb |
04:05 |
|
alex |
yes I was thinking that |
04:05 |
|
alex |
shell should be fine |
04:05 |
|
alex |
set etcdir |
04:05 |
|
alex |
find an auto install file |
04:06 |
|
alex |
check not cvs layout |
04:06 |
|
alex |
RC warning |
04:06 |
|
alex |
checkabortedinstall |
04:07 |
|
alex |
I really need to be reading Install.pm at the same time |
04:07 |
|
alex |
I wonder what checkabortedinstall does? |
04:07 |
|
alex |
obvoiusly checks for an aborted install, |
04:07 |
|
alex |
but what does it do if it discovers such? |
04:08 |
|
alex |
ok, there's a description in Install.pm |
04:09 |
|
alex |
then if (-e "$etcdir/koha.conf") { |
04:09 |
|
slef |
craps out, probably |
04:09 |
|
slef |
maybe able to get summaries from the command: perldoc Install.pm |
04:10 |
|
alex |
ok this section is modifying koha.conf and putting the new versonnumber in? |
04:10 |
|
alex |
maybe not |
04:11 |
|
alex |
just doing some check of a previous version? |
04:11 |
|
slef |
should be IIRC |
04:11 |
|
alex |
and exiting if already installed? |
04:11 |
|
slef |
koha.conf rewrite is later if at all... (might be in upgrader instead) |
04:11 |
|
slef |
koha.upgrade even |
04:12 |
|
alex |
ok, then does a welcome message to the user |
04:12 |
|
alex |
check perl modules |
04:13 |
|
alex |
some get.... functions, perhaps asking for user input for stuff |
04:13 |
|
alex |
getapacheinfo |
04:13 |
|
alex |
yep, this is all given if we assume debian/apache2 |
04:14 |
|
alex |
btw, I think I am going to target to apache2 and tell the user to sort it out manually if they want something different |
04:14 |
|
slef |
I think we may be able to handle debian/apache and debian/apache-ssl in one framework, but maybe not in first package |
04:14 |
|
alex |
yes |
04:15 |
|
alex |
very much I want to get something that works even though it is not polished or flexable as it coul be |
04:15 |
|
alex |
for a first stab |
04:15 |
|
alex |
then let bus/requests from others determine where it shoudl be improved |
04:16 |
|
alex |
bugs/requests I should say |
04:16 |
|
alex |
getinstallationdirectories |
04:16 |
|
alex |
this should all be a given for a debian package |
04:16 |
|
alex |
getdatabaseinfo |
04:17 |
|
alex |
mmm, this needs to be handled post install of course |
04:17 |
|
alex |
getapachevhostinfo |
04:17 |
|
slef |
that can be debconf'd |
04:18 |
|
alex |
yes |
04:18 |
|
alex |
webmaster email address, virtual hostname, and the ports |
04:18 |
|
alex |
that the OPAC and INTRANET modules run on. |
04:19 |
|
alex |
do we need an email address? |
04:19 |
|
alex |
does the standard apache2 install have one we could use? |
04:19 |
|
slef |
probably not.. it's only used for ServerAdmin |
04:19 |
|
slef |
IIRC |
04:19 |
|
slef |
Yes, apache2 should have one already |
04:20 |
|
alex |
my vanilla apache2 install has - ServerAdmin webmasterlocalhost |
04:21 |
|
alex |
I've been wondering about the setup of the virtualhosts |
04:22 |
|
alex |
I'd like to get a "safe" "static" config that works as a base to go from |
04:22 |
|
alex |
so i'm thinking |
04:23 |
|
alex |
localhost:6722 for opac |
04:23 |
|
alex |
and localhost:5642 for intranet |
04:23 |
|
alex |
or *:6722 for opac |
04:24 |
|
alex |
ocalhost:5642 for intranet |
04:24 |
|
slef |
putting intranet on localhost seems a good idea |
04:24 |
|
alex |
and of course telling people to change it. |
04:24 |
|
alex |
if they need |
04:26 |
|
alex |
I like having names like opac.mydomain.com or library.mydomain.com and koha.mydomain.com |
04:27 |
|
alex |
as the virtualhosts, but this relies on dns/name resolution being set up correctly |
04:27 |
|
alex |
and is liable to be problematic/difficult or impossible depending on the individual situations |
04:28 |
|
alex |
Also I don't know if I would be allowed to include listen directives before the virtualhost directives to tell apache to listen on that port |
04:29 |
|
alex |
and whether to automatically enable the virtualhosts |
04:29 |
|
alex |
or tell the user to run a2ensite.... |
04:30 |
|
slef |
probably a final debconf question, defaulting to no |
04:30 |
|
alex |
there is a ports.conf with a listen directive |
04:30 |
|
alex |
I wonder if there is policy/procedure to modify that file |
04:32 |
|
slef |
I think I'd put it in the vhost file |
04:32 |
|
alex |
yes, seems sensible |
04:33 |
|
alex |
updateapacheconf is next in installer.pl |
04:34 |
|
alex |
which creates the virtualhost conf after getting info from the user |
04:34 |
|
alex |
presumably |
04:34 |
|
alex |
so this would be largely static in my setup |
04:35 |
|
alex |
installfiles is next |
04:35 |
|
alex |
which installs the files and creates koha.conf |
04:36 |
|
alex |
I want to move files at package build of course and create a koha.conf if necessary post install |
04:37 |
|
alex |
BTW, is this script used on upgrade, or does it only generally cater for installs> |
04:37 |
|
alex |
? |
04:37 |
|
slef |
installs |
04:37 |
|
slef |
there's a koha.upgrade or similar nearby? |
04:38 |
|
alex |
yes, there is a koha.upgrade |
04:38 |
|
alex |
ok, but they share Install.pm |
04:38 |
|
alex |
makes sense |
04:39 |
|
alex |
databasesetup |
04:40 |
|
alex |
do we requre cmd line mysql client libs? |
04:41 |
|
slef |
yes, at the moment |
04:41 |
|
alex |
it looks like Install.pm uses them rather than using perl db libs |
04:41 |
|
slef |
this is one bit that I'm unsure about |
04:41 |
|
slef |
should this be done from postinst or from a new web setup screen |
04:42 |
|
alex |
so we'll requre a dependancy on mysql-clientxxx |
04:42 |
|
slef |
probably postinst for the mysql connection |
04:42 |
|
alex |
yes |
04:42 |
|
slef |
then web setup for loading the database tables |
04:42 |
|
slef |
as there are librarian questions like which MARC tables to load |
04:42 |
|
slef |
s/tables/datasets/ |
04:42 |
|
alex |
this is where i am vague |
04:43 |
|
alex |
is it ok to leave that stuff to the web-ui |
04:43 |
|
alex |
ie it is possible to do that |
04:44 |
|
slef |
it's possible, I think |
04:44 |
|
slef |
maybe even a good idea |
04:44 |
|
slef |
keep the deb install for sysadmin questions |
04:44 |
|
slef |
and try to do all librarian stuff from the web |
04:44 |
|
alex |
yes |
04:44 |
|
slef |
but we're writing new code |
04:44 |
|
slef |
hopefully it will be welcome ;-) |
04:44 |
|
alex |
I'll just go through database setup |
04:44 |
|
slef |
I guess kados is still asleep |
04:45 |
|
alex |
this will be done in postinst (probably) |
04:45 |
|
slef |
database setup is a bit screwy |
04:45 |
|
slef |
at least last I looked |
04:45 |
|
slef |
it loads an old SQL layout |
04:45 |
|
slef |
then runs scripts from updater/updatedatabase to modify it |
04:45 |
|
alex |
:) |
04:46 |
|
slef |
this is something that has been on my 'I should tidy that up' desk for some time |
04:46 |
|
alex |
why tidy up when we can load another layer of hacking? |
04:47 |
|
alex |
my motto is why do today what I can put off till tomorrow :) |
04:47 |
|
alex |
just kidding |
04:47 |
|
slef |
it's actually avoiding some hacking |
04:47 |
|
alex |
yes |
04:47 |
|
slef |
as the updating will be needed for koha.upgrade |
04:48 |
|
slef |
but it's not the way I'd choose to do it |
04:48 |
|
alex |
no |
04:49 |
|
alex |
I wonder if we should run this postinst database setup as a perl script which calls Install.pm? |
04:49 |
|
alex |
just a small part of installer.pl / koha.upgrade? |
04:50 |
|
alex |
for the db setup? |
04:51 |
|
slef |
not sure |
04:51 |
|
slef |
I guess the sysadmin will need to fix problems with the schema |
04:52 |
|
alex |
databasesetup itself only creates the db, user, access control |
04:52 |
|
slef |
but picking which MARC data to load is a question for a librarian |
04:52 |
|
alex |
next in install.pl comes updatedatabase, then populatedatabase, paradoxically |
04:53 |
|
alex |
updatedatabase "Updates the Koha database structure, including the addition of |
04:53 |
|
alex |
MARC tables" |
04:53 |
|
tumer |
hdl:?? |
04:54 |
|
hdl |
yes |
04:54 |
|
alex |
but this asks librarian questions: MARC21/UNIMARC.... |
04:54 |
|
slef |
yeah, I said it was screwy |
04:56 |
|
alex |
then populatedatabase "Populate the non-MARC tables." |
04:58 |
|
alex |
including setting up branches and printers? |
04:59 |
|
alex |
then load more sql stuff |
04:59 |
|
alex |
including the "This is a VERY important feature. By selecting the proper options..." |
05:00 |
|
alex |
by this stage in the install my eyes had glazed over and I had no idea what I was supposed to be selecting |
05:01 |
|
alex |
ok, so of databasesetup, updatedatabase and populatedatabase |
05:01 |
|
alex |
databasesetup is the base db setup, no sql loaded |
05:01 |
|
slef |
ITYM no data loaded |
05:02 |
|
alex |
and it would be ideal if the rest could be done through the web ui |
05:02 |
|
alex |
ITYM? |
05:02 |
|
slef |
I think you mean |
05:02 |
|
slef |
I think the tables and so on should be created in postinst |
05:02 |
|
slef |
as the sysadmin will have to handle that if it fails |
05:02 |
|
slef |
(do librarians often hack SQL?) |
05:03 |
|
slef |
picking MARC, data, branches and printers should be web tasks |
05:03 |
|
alex |
I don't know where the table structure gets loaded, I'll try and find it |
05:03 |
|
slef |
I think we'll need to mysql_dump a few installer creations |
05:04 |
|
alex |
that's another easy way out of course. |
05:04 |
|
alex |
just mysqldump an install which is as vanilla as possible and do all further |
05:05 |
|
alex |
tweaking by the user through the web-ui |
05:06 |
|
slef |
Adding a "load this SQL file" function to the system screens is probably a good idea anyway |
05:06 |
|
slef |
Is there a backup/restore tool already? I forget. |
05:06 |
|
alex |
ahh yes, you are right, system("$mysqldir/bin/mysql '-u$user' $database < koha.mysql"); |
05:06 |
|
alex |
appears in databasesetup |
05:07 |
|
alex |
the first part of the db stuff before any q's are asked |
05:07 |
|
alex |
so the table structure is there early on |
05:08 |
|
alex |
do you mean mysqldump during postinstall? |
05:08 |
|
alex |
or do you mean a koha dump script of some kind? |
05:09 |
|
alex |
mysql cmd line client should be available in postinstall, we'll make it a dep. |
05:09 |
|
alex |
however, the db server may not be online at this stage |
05:10 |
|
alex |
so it would be good to have the db initial setup script as a utility external to postinst |
05:10 |
|
alex |
so the admin has the ability to run it later if needed |
05:11 |
|
alex |
I've found from practice that apt-get install mysql-server random-mysql-webapp |
05:11 |
|
alex |
usually fails because the server is not running when webapp does it postinst |
05:12 |
|
alex |
my big question is is koha able to run and present an intranet screen with the tables created but no data in it? |
05:13 |
|
alex |
I guess I need to try it. |
05:14 |
|
alex |
OK, I think I've got plenty to work on, I'll see what I get done in the next few days |
05:14 |
|
alex |
Unfortunately I don't actually have any time for it.....:) |
05:15 |
|
alex |
I'll post my postinstall/debian dir workings when I get something. |
05:27 |
|
hdl |
tumer ? |
05:27 |
|
dewey |
tumer is here for a few seconds ;-) or debugging it very fast as it it in use |
06:04 |
|
tumer |
hdl:? |
06:04 |
|
hdl |
yes |
06:04 |
|
tumer |
i chaeched your records it works with me |
06:04 |
|
hdl |
I installed zebra-2.0.2, reindexed. |
06:05 |
|
hdl |
search for 686$9 works. |
06:05 |
|
tumer |
and stop using att 89.. atrributes they are not portable anway |
06:05 |
|
hdl |
But now, for authority search, I get a [warn] bib1.att:85: Local attributes not supported |
06:06 |
|
tumer |
instead use names for anything not Bib-1 |
06:06 |
|
tumer |
so thats all use the same names |
06:06 |
|
hdl |
tumer: but then, how can I give names ? |
06:07 |
|
tumer |
ie.authid, linkid etc |
06:07 |
|
hdl |
in record.abs ? |
06:07 |
|
tumer |
yes search with @attr 1=koha-auth-number |
06:07 |
|
hdl |
and find @attr 1=myname will work ? |
06:07 |
|
tumer |
and do not define them in bib-1 |
06:07 |
|
tumer |
as long as record.abs have it yes |
06:08 |
|
hdl |
COOL. |
06:08 |
|
tumer |
and its the prefered method for zebra |
06:08 |
|
hdl |
I thought it would be a nightmare to maintain 3 different files for search definitions. |
06:09 |
|
hdl |
But then for ccl.properties, does it work all the same ? |
06:09 |
|
tumer |
i think we have to publish a list of what to use for indexing and al use the same |
06:09 |
|
hdl |
DEFINITELY. |
06:09 |
|
tumer |
yes ccl.properties will work if you define names in it as well |
06:10 |
|
hdl |
but LC and exotic classification would be out of reach. |
06:10 |
|
tumer |
ie. auth=koha-auth-number |
06:10 |
|
tumer |
no |
06:10 |
|
hdl |
(not LC but exotic ones.) |
06:10 |
|
tumer |
no check head for koha2index.xml see the names i used there |
06:11 |
|
tumer |
you can use the same in record.abs |
06:11 |
|
tumer |
i suggest we use global bib-1 for all bibliographic data and koha-fieldnames for anything else |
06:12 |
|
hdl |
You removed a big pain in my neck with saying use names.... |
06:12 |
|
hdl |
Some important information to index cannot be found in Kohafields. |
06:12 |
|
tumer |
also check koha_attr table which will remove some more pain |
06:13 |
|
tumer |
hdl give me an example of what you mean? |
06:14 |
|
tumer |
i search subjects with three differnt attributes lc-subject-heading, subject-heading, subject_subdivision |
06:14 |
|
hdl |
imagine I am a multiple branches library |
06:14 |
|
hdl |
And I want to trace modification and adding. |
06:15 |
|
hdl |
Which library created that biblio. |
06:15 |
|
hdl |
So I would use 801$b for instance. |
06:15 |
|
hdl |
or 831$b |
06:16 |
|
tumer |
i use similar for keeping catalogures id , circulation id, modifiers id |
06:16 |
|
hdl |
I definitely will check your work out. |
06:16 |
|
tumer |
so all you do is melm 831$b cataloguing-branch |
06:16 |
|
tumer |
and serach on it |
06:18 |
|
tumer |
koha_attr helps to define catalogures-branch to UNIMARC 831$b and also defines @attr 1=cataloguers branch also defines whether this field is searchable by OPACS as well etc.. |
06:20 |
|
tumer |
i have commitetd koha_attr table for MARC21 you can modify it for UNIMARC |
06:29 |
|
hdl |
thx : will check after lunch |
09:06 |
|
btoumi |
kados: are u around |
09:10 |
|
kados |
btoumi: yep |
09:10 |
|
btoumi |
sorry kados |
09:10 |
|
btoumi |
hdl help me |
09:11 |
|
kados |
k |
09:11 |
|
btoumi |
perahps later i have question about zebra ;=) |
09:11 |
|
kados |
:-) |
09:35 |
|
paul |
kados : hello. did you get some sleep yesterday ? |
09:35 |
|
paul |
how is NPL going ? |
09:36 |
|
kados |
paul: so far very well |
09:37 |
|
kados |
there are some outstanding bugs |
09:37 |
|
kados |
some code was changed between 2.2.4 and 2.2.6/dev_week ... in borrower management, etc. |
09:37 |
|
kados |
and the behavior is different than expected |
09:38 |
|
kados |
but all in all, things have gone very well so far |
09:44 |
|
paul |
kados : |
09:44 |
|
paul |
it seems there are 2 new columns in items table : |
09:44 |
|
paul |
cutterextra & onloan. |
09:45 |
|
paul |
can you explain when & why you introduced them ? |
09:45 |
|
kados |
paul: those were added by tumer |
09:45 |
|
kados |
onloan keeps track of the date_due of an item |
09:46 |
|
kados |
if it's not on loan it should be 0000-00-00 |
09:46 |
|
paul |
ok, so it's a "date" column |
09:46 |
|
kados |
cutterextra has something to do with how NEU does call numbers |
09:46 |
|
kados |
paul: it's not a date column |
09:46 |
|
kados |
paul: as in mysql date |
09:46 |
|
kados |
paul: because it needs to be readable by zebra |
09:47 |
|
paul |
what fills it ? |
09:47 |
|
kados |
it is filled with the cron job |
09:47 |
|
kados |
update_items.pl |
09:47 |
|
kados |
wait ... |
09:48 |
|
kados |
it's updated on circ too |
09:48 |
|
kados |
so ... I will explain |
09:49 |
|
kados |
in dev_week, the authoritative items data is in the items table, not in the MARC record |
09:49 |
|
kados |
so in circ, biblioitems.marc isnt' updated at all |
09:49 |
|
kados |
only issues,reserves,items tables |
09:50 |
|
kados |
then ... as often as you want, you can synchronize the items, biblioitems.marc and zebra |
09:50 |
|
kados |
with update_items.pl |
09:50 |
|
kados |
(which, I need to commit my new copy of) |
09:50 |
|
kados |
(I'll do so now) |
09:51 |
|
kados |
ok, committed |
09:51 |
|
kados |
paul: misc/update_items.pl |
09:52 |
|
kados |
it looks like -itemnumber is not implemented yet |
09:52 |
|
kados |
I will do so soon |
09:52 |
|
paul |
ok, understood. |
09:53 |
|
paul |
what is onloan column in your SQL ? |
09:53 |
|
paul |
(same question for cutterextra) |
09:53 |
|
paul |
i'll add it to updater/updatedatabase |
09:53 |
|
kados |
ahh ... |
09:53 |
|
kados |
onloan IS a date ;-) |
09:53 |
|
kados |
Cutterextra varchar(45) |
09:54 |
|
paul |
kados : SAN-OP plans to have a database for tests at the end of the month. I hope we will score the goal, even if with a lot of stress... |
09:54 |
|
kados |
my feeling is that it's not impossible |
09:55 |
|
kados |
do they plan to use authorities in zebra? |
09:55 |
|
paul |
mine too. |
09:55 |
|
paul |
yep & hdl is working on them |
09:55 |
|
kados |
I have more table changes that I can explain |
09:56 |
|
kados |
can I paste them in? |
09:56 |
|
kados |
to upgrade NPL I did: |
09:56 |
|
kados |
alter table kohanew.biblio add frameworkcode char(4); |
09:56 |
|
kados |
alter table kohanew.deletedbiblio add frameworkcode char(4); |
09:56 |
|
kados |
update kohanew.biblio,kohanew.marc_biblio set biblio.frameworkcode=marc_biblio.frameworkcode where marc_biblio.biblionumber=biblio.biblionumber; |
09:57 |
|
kados |
alter table kohanew.biblioitems add marcxml text; |
09:57 |
|
kados |
(though not used I think) |
09:57 |
|
kados |
alter table kohanew.biblioitems add lcsort varchar(25); |
09:57 |
|
kados |
alter table kohanew.biblioitems add ccode varchar(4); |
09:57 |
|
kados |
alter table kohanew.items add onloan date; |
09:57 |
|
kados |
alter table kohanew.items add Cutterextra varchar(45); |
09:57 |
|
kados |
alter table kohanew.issues add issue_date date; |
09:57 |
|
kados |
drop table if exists kohanew.marc_word |
09:57 |
|
kados |
alter table kohanew.items modify binding tinyint(1) default NULL |
09:58 |
|
kados |
(this was done with chris's approval) |
09:58 |
|
kados |
alter table kohanew.systempreferences change value value text |
09:58 |
|
kados |
alter table kohanew.systempreferences change explanation explanation text |
09:59 |
|
kados |
NOOO! :-) |
09:59 |
|
kados |
I will explain it in SQL |
09:59 |
|
kados |
UPDATE biblioitems set ccode = itemtype; |
09:59 |
|
kados |
UPDATE biblioitems set itemtype = "CIRC"; |
09:59 |
|
kados |
UPDATE biblioitems set itemtype = "NOCI" WHERE ccode IN ("REF","LH","JREF"); |
09:59 |
|
kados |
UPDATE biblioitems set itemtype = "AVID" WHERE ccode IN ("AV","AVJ","AVJN","AVNF","DVD","DVDJ","DVDN","DVJN"); |
09:59 |
|
kados |
UPDATE biblioitems set itemtype = "AVBK" WHERE ccode IN ("AB","ABN","AC","ACN"); |
10:00 |
|
kados |
UPDATE biblioitems set itemtype = "NOCI" WHERE ccode IN ("REF","LH","JREF"); |
10:00 |
|
kados |
in many libraries, the 'material type' is not the same as 'circ rules' |
10:01 |
|
kados |
and the only thing material type is used for is searching |
10:01 |
|
kados |
so ... it's better to have ccode and itemtype, ccode is 'material type' and itemtype is 'circ rules' |
10:01 |
|
kados |
paul: does that make sense? |
10:03 |
|
kados |
hehe |
10:03 |
|
paul |
yes, that really makes sense & seems highly interesting. |
10:03 |
|
hdl |
already here. |
10:04 |
|
paul |
read previous explanations about CCODE, that seems really interesting to me. |
10:04 |
|
kados |
in fact, later this month, I must develop a solution for another library that needs two levels of 'circulation rules' ... one at record level and one at item level ... so I will be adding items.itemtype as well |
10:04 |
|
paul |
couldn't you use ccode another way ? |
10:04 |
|
kados |
and I may move biblioitems.itemtype to biblioitems.biblioitemtype (as it should have been already) |
10:05 |
|
kados |
do you have an idea? |
10:05 |
|
paul |
because if CCODE = itemtype, then you've got what you need. |
10:05 |
|
paul |
if I don't mind |
10:05 |
|
paul |
(i mean : the list of available Ccode = the list of itemtypes) |
10:06 |
|
kados |
I don't follow |
10:06 |
|
kados |
ccode != itemtype |
10:07 |
|
paul |
differents fields, but if you set them to have the same values in them. |
10:07 |
|
paul |
mmm... no I understand your problem |
10:07 |
|
kados |
do you mean biblioitems.ccode and biblioitems.itemtype? |
10:07 |
|
kados |
what we need is: |
10:07 |
|
kados |
biblioitems.biblioitemtype (circ rules) |
10:07 |
|
kados |
biblioitems.ccode (collection code) |
10:07 |
|
kados |
biblioitems.mt (material type) |
10:08 |
|
kados |
items.itemtype (circ rules) |
10:08 |
|
kados |
items.itemtype overrides biblioitems.biblioitemtype |
10:08 |
|
kados |
so you can set circ rules on the whole record ... and also on individual items |
10:09 |
|
kados |
regardless of the collection code or material type |
10:09 |
|
kados |
I consider what I did for NPL to be a poor hack |
10:10 |
|
kados |
that I will fix in dev_week in late October or early november |
10:10 |
|
kados |
paul: also, you may enjoy this: http://kados.org/apache/remove[…]meters-from-uris/ |
10:11 |
|
paul |
the other difficulty being to have something quick & easy to set for libraries that want something basic... |
10:11 |
|
kados |
for such a library, they can just use biblioitem.itemtype |
10:11 |
|
kados |
well ... |
10:11 |
|
kados |
biblioitems.biblioitemtype |
10:11 |
|
kados |
the rest are optional |
10:14 |
|
kados |
paul: does that make sense? |
10:14 |
|
paul |
yes, really |
10:15 |
|
kados |
if I'm able to comprehend the code, I'll be working on this in a week or so |
10:15 |
|
kados |
otherwise, I may ask for help ;-) |
10:30 |
|
kados |
by the way, NPL's modest one-processor machine, with only 4 gigs of ram: |
10:30 |
|
kados |
load average: 0.60, 0.41, 0.35 |
10:30 |
|
kados |
compared to rel_2_2 which runs at average of 6-7 throughout the day |
10:30 |
|
kados |
and more during busy days ;-) |
10:30 |
|
kados |
zebra++ |
10:31 |
|
kados |
removing_date_manip++ |
10:31 |
|
kados |
paul: it seems you undid what changes I made by removing date_manip |
10:31 |
|
paul |
??? |
10:32 |
|
kados |
I saw commit where date::manip was put back into rel_3_0 |
10:32 |
|
paul |
not from me, but from toins maybe. |
10:32 |
|
kados |
I would prefer it be removed and any function that you need be reimplemneted |
10:33 |
|
paul |
I agree we want to get rid of Date::Manip. |
10:33 |
|
paul |
toins will investigate when he has a chance |
10:33 |
|
kados |
Date::Calc is what I used in Circ2.pm |
10:33 |
|
toins |
kados, there are a lot of script which use ParseDate from Date::Manip |
10:33 |
|
kados |
it has many nice features |
10:33 |
|
kados |
toins: parsedate could be reimplemented in Date::Calc quite easily |
10:33 |
|
toins |
yep sure |
10:33 |
|
kados |
without all the overhead |
10:34 |
|
kados |
even adding 'use Date::Manip' to your code means about 80% more proc for circ |
10:34 |
|
kados |
based on my testes |
10:34 |
|
toins |
i have put back Date::Manip only for test in circulation |
10:34 |
|
kados |
tests even :-) |
10:34 |
|
toins |
wow... |
10:35 |
|
toins |
kados, do you still used C4::Date.pm in dev_week ? |
10:35 |
|
kados |
toins: not for circ |
10:35 |
|
toins |
ok |
10:45 |
|
paul |
kados ? |
10:45 |
|
dewey |
i heard kados was becoming a true Perl Monger... |
10:46 |
|
paul |
why did you write : |
10:46 |
|
paul |
alter table kohanew.biblioitems add marcxml text; |
10:46 |
|
paul |
(though not used I think) |
10:46 |
|
paul |
you don't use marcxml in dev_week ? |
10:46 |
|
paul |
because we use it a lot in rel_3_0 |
10:46 |
|
paul |
(& we don't use raw marc) |
10:47 |
|
kados |
I use biblioitems.marc |
10:47 |
|
paul |
mmm... we will have to decide which one we use... |
10:47 |
|
paul |
marcxml is nice, because it's easy to edit a wrong XML record |
10:47 |
|
paul |
iso is nice because it's small |
10:48 |
|
kados |
yep |
10:48 |
|
kados |
I think in rel_3_0 marcxml is the right choice |
10:49 |
|
paul |
that sound nice to my ears ;-) |