Evergreen ILS Website

IRC log for #evergreen, 2016-12-21

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
03:04 tsbere joined #evergreen
03:23 Bmagic joined #evergreen
03:48 tsbere joined #evergreen
04:16 Bmagic joined #evergreen
05:00 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
07:10 agoben joined #evergreen
07:15 rjackson_isl joined #evergreen
08:39 mdriscoll joined #evergreen
08:52 bos20k joined #evergreen
08:54 collum joined #evergreen
08:58 Dyrcona joined #evergreen
09:05 rlefaive joined #evergreen
09:15 Bmagic jeff: I think I can handle the rest. I just wondered if the reporter could find bibs that lacked a tag. I use nullable?
09:20 yboston joined #evergreen
09:36 Dyrcona Looking for something that isn't there is a logical contradiction. :)
09:38 maryj joined #evergreen
09:38 Dyrcona Bmagic: I would not try that in the reporter, but then I never try anything in the reporter... Every time I do, I choke the db server nearly to death.
09:50 Dyrcona Here's a question: Why does the actor.org_address table have an org_unit column, when the org_unit table links to org_address via id?
09:51 Dyrcona And, the prompt for this question is that someone came into my office this morning wondering why the org_units in the org_adress table do not match up to the org_units that link to the addresses.
09:52 jvwoolf joined #evergreen
09:52 Dyrcona That wasn't there question in so many words, but that was the cause of their question.
09:53 Dyrcona We also don't seem to have any issues because the aoa.org_unit != aou.id of the aou that links to aoa via one of its address ids.
09:54 Dyrcona I'm guessing that aoa.org_unit isn't used anywhere, but haven't looked.
09:54 csharp Bmagic: theoretically, you could do a left join ("Parent" in nullability-speak) from Bib Record to the metabib.full_rec table (don't remember the reporter label for that) others may have a better idea
09:55 csharp our cataloger has wanted something similar before to identify "incomplete" bibs
09:56 * Dyrcona has recently found that left joins to mrfr can be slow, but that's basically how I'd do it.
09:57 Dyrcona Oh, and upshot of my Perl threads experiment yesterday is, Don't do it, kids.
09:58 Dyrcona MARC::File::XML is not thread safe because Encode.pm is not thread safe.
09:58 Dyrcona @blame 9 Encode.pm
09:58 pinesol_green Dyrcona: Encode.pm is why we can never have nice things!
10:20 berick can anyone point me at the docs that form the basis for the authority control set mappings, specifically LoC?
10:24 jeff "You map that way so often. I wonder what your basis for comparison is."
10:30 jvwoolf joined #evergreen
10:42 Dyrcona berick: You could start here: http://www.loc.gov/marc/authority/
10:48 berick Dyrcona: yeah I started there...  did not find what I was looking for.
10:48 Christineb joined #evergreen
10:48 berick i'm sure it's in there somewhere
10:51 abneiman berick: one of these maybe? https://www.loc.gov/standard​s/mods/mods-conversions.html
10:55 Dyrcona berick: I don't think there is a simple document for those. I think it comes right out of the authority field specifications.
10:57 berick Dyrcona: i'm happy to dig if I'm digging in the right place.  and i'm probably looking right past it, but I can't seem to find an example of bib field X is controlled by authority field Y (for LoC thesaurus) in the LC docs.
10:58 Dyrcona berick: Yeah. I thought that is what you're looking for and I don't see it, either.
10:59 Dyrcona berick: Mayb this: http://www.loc.gov/marc/authority/ecadcntf.html
11:00 Dyrcona Eh, maybe not. That appears to just explain what to do when you're making a link in a controlled subfield.
11:03 Bmagic Dyrcona: years ago, I programmed a "multi" thread perl program that used Encode.pm, however, to get around the multi thread issue, I spawned my seperate instances that all wrote back to their own pid file
11:04 Bmagic then the master thread periodically checked up on the "threads" by reading that file
11:04 Dyrcona Bmagic: I'm thinking about either 1) rewriting in Java to use threads or 2) using fork and pipes to communicate.
11:05 Dyrcona I have a case where I want long-running processes to get information from a master process.
11:05 Bmagic right, so it can be done in perl
11:05 rlefaive joined #evergreen
11:06 Dyrcona I didn't say it couldn't be done in Perl, just not with Perl ithreads on Perl 5.14. (I didn't get that specific before, tbh.)
11:06 berick abneiman: thanks, but I don't think that's gonna answer my questions either
11:06 Bmagic Dyrcona: You go with your bad self!
11:06 Dyrcona But, Thread::Queue in Perl seemed like a natural choice for managing data flow.
11:07 Dyrcona Bmagic: By years ago are we talking prior to Perl 5.8?
11:08 Dyrcona 'Cause everything changed in Perl 5.8 when it comes to threads.
11:08 Bmagic no, I believe it was on 14.04
11:08 Bmagic You can see the code here https://github.com/mcoia/mobius_e​vergreen/tree/master/bib_extract
11:12 jihpringle joined #evergreen
11:14 Dyrcona That's not actually threading...That's a lot of overhead to manage fork/exec. :)
11:32 bmills joined #evergreen
11:36 Bmagic Dyrcona: nope's it's not real threading, by design because of the limitation of Encode.pm. This was the solution that I used
11:37 Bmagic But, it does extract bib records from the database super fast :)
11:37 Dyrcona Bmagic: Yep. Perls threads  like old userland threads and fork a new interpreter, hence the multiple free errors and segmentation faults when you use Encode.pm.
11:37 Bmagic 10 threads each gathering different sections of biblio.record_entry
11:38 Dyrcona Yeah. I'm going the other direction and have options similar to yours. But, not sure, I might just do it in Java.
11:40 berick say what you will about Java, the threading support is excellent
11:41 Dyrcona Indeed!
11:41 Bmagic Java has a special place in my heart. I wrote some of my first programs with it
11:41 Bmagic oracle-- # laying ruin to pure sweet coffee
11:41 berick :)
11:42 Dyrcona oracle-- # Just because...Oracle.
11:42 berick java's a fun place to visit
11:42 Dyrcona Yeah, it is, but it can brutal if you stay there too long.
11:42 Bmagic qbasic++
11:42 Dyrcona Fortunately, I use Perl to write the more tedious bits for me. :)
11:44 Dyrcona Perl is now my favorite editor after last night's one-liner that I wrote for something that I'm doing on my own: for (/\W?(\w+)(\W*)/g) { print ucfirst lc $_ }
11:44 Dyrcona That would have been murder in sed.
11:45 abneiman berick: Yeah, I misunderstood what you were asking.  One can infer that Authority field 100 will control Bib fields 100/600/700; Auth 130 will control Bib fields 130/730/830; etc. -- but I'm also not finding something that explicitly lays that out.
11:46 Bmagic anyone remember virtual pimp? http://www.petesqbsite.com/secti​ons/express/issue3/virtualp.bas
11:46 Bmagic qbasic++
11:52 berick abneiman: on that note, any idea why in EG the authority control set for bib field 600 (for example) is controlled by auth fields 100,180,181,182,185 ?
11:52 berick the linking script only looks at auth 100
11:52 berick trying to understand the relationship between the two
11:55 abneiman berick: Auth 180,181,182,185 control subdivisions -- in MARC Bib they would correspond to 6xx subfields x, z, y, and v respectively
11:56 brahmina joined #evergreen
11:57 berick abneiman: i'm confused by the fact that a specific subfield, say 600z, appears to be controlled by 5 different authority fields.  100z, 180z, 181z, ...  i must be misinterpreting something
11:59 berick or it's context-specific and I don't yet understand the context
12:17 jyorio joined #evergreen
12:18 barbara joined #evergreen
12:29 jwoodard joined #evergreen
12:33 abneiman berick: I can see all of them playing a role, but I've never done high-level auth work so I'm not sure what all of the associations there are.  I know that for the 18x auth fields there are restrictions on when the x/z/y/v subfields can be added and that subdiv headings get wonky.
12:33 abneiman Relatively simple example: You can have the established heading -- "651 (151) $a France $x History $y 18th century", which is its own thing as a whole (https://lccn.loc.gov/sh85051305) but also contains pieces "18th century" (Auth field 182, https://lccn.loc.gov/sh2002012474), "History" (Auth field 150, https://lccn.loc.gov/sh85061212), "History -- 18th century" (Auth field 180 with subfield y, https://lccn.loc.gov/sh2002006124), "France -- History
12:33 abneiman (Auth field 151 with subfield x, https://lccn.loc.gov/sh85051256).
12:34 abneiman And to FURTHER complicate matters, that whole bit can be put under another heading i.e. 650 (150) $a Architecture $z France $x History $y 18th century (https://lccn.loc.gov/sh2009115453)
12:34 abneiman Moral of the story is authorities are hard to get a handle on.
12:35 berick abneiman++ thanks, this gives me something to chew on
12:35 abneiman np :) it was a fun little bit of digging
12:41 berick ok, so i take from this the bib 651 is controlled by the auth 151, and the auth 151 in turn is controlled by the various related authority records containing 18X tags
12:42 berick roughly speaking
12:43 berick for the purposes of the bib field control, we don't need to concern ourselves with the 18x tags, because that part of the logic (keeping the authority records themselves under control) is handled elsewhere.
12:54 abneiman berick: that sounds like it should be right, but that's where my knowlege breaks down
12:55 jeff ``As you can see, individual libraries are the direct beneficiaries of all of this MARC 21 standards work. Without it, libraries would not be able to share authority records with ease and accuracy.''
12:56 berick abneiman: I'll take it :)  it's a good start, anyway
12:59 * Dyrcona tries to figure out if a fund transfer can be undone in the database.
13:38 Dyrcona It looks like I can just delete the acq.fund_transfer entries...
13:40 Dyrcona Looks like I need to delete the corresponding acq.fund_allocation entries, too.
13:43 Dyrcona And, I should decrement the amounts in fud_allocation_total?
13:43 Dyrcona :)
13:43 Dyrcona fund...
13:44 Dyrcona Oh, this is getting messy....
13:46 jihpringle Dyrcona was it a transfer from a funding source to a fund or between funds?
13:47 csharp Dyrcona: everything in acq is this messy :-/
13:49 Dyrcona jihpringle: It was between funds I'm told. I found the acq.fund_transfer entries.
13:52 jihpringle_ joined #evergreen
13:53 Dyrcona :)
13:53 Dyrcona jihpringle: It was between funds.
13:53 jihpringle Drycona: we delete transfers between the funding sources and funds for our libraries but tell them they have to do a second transfer to fix issues between funds since it's so messy
13:55 Dyrcona Well, this doesn't look too bad, yet. ;)
13:55 jihpringle good luck :)
13:55 Dyrcona I'll probably have to mess with funding_source_balance, too..
13:55 jihpringle if you figure out all the tables that have to be adjusted we'd be interested to know
13:56 Dyrcona I should just tell them too bad, and be more careful next time. :)
13:56 jihpringle that's what we ended up doing
13:57 jihpringle since they can do another transfer in the staff client
13:58 Dyrcona Ah... The source fund is closed, so they can't transfer back.
13:59 jeff musing on a few things from earlier today (both here and elsewhere), it's sometimes the simplest questions that uncover the complexity of a topic or standard. like how a question of "is there a widely available list of all Dewey Decimal Classification numbers?" quickly leads to subdivisions and the discovery of oddities like the largest (currently known) valid DDC number of 331.890410263362014098133, etc.
14:00 Dyrcona jeff: Technically, 332.0 is larger than that number. :P
14:00 jeff best thing about a large palindrome? you only have to remember half of it.
14:01 jihpringle Drycona: they should be able to switch the fund back to Active, transfer the money, and then set the fund back to inactive (in the funds interface)
14:01 jeff Dyrcona: s/larg/long/ :P
14:02 Dyrcona jihpringle: It was last year's fund and fiscal year has been closed.
14:02 Dyrcona jeff: :P I knew that..... just yanking your chain.
14:03 Dyrcona Anyway, I've got another pointless question to answer: Doe is pay lost fees or fines first? I should say, "Does it really matter?"
14:03 Dyrcona With typing like that, I could ruin the database. :)
14:03 jeff "Doe is pay lost fees or fines first?"?
14:04 jeff remember: late doe season permits are only valid on private property.
14:04 jwoodard 4 more days till Christmas and my Director wants an Inventory by closing tomorrow......
14:04 jwoodard challenge accepted!
14:06 jihpringle Drycona: if year end close out caused the fund transfer issue (rather than staff) you may have fallen victim to https://bugs.launchpad.net/evergreen/+bug/1478712 in which case it's messier and involves multiple libraries
14:06 Dyrcona jeff: Ha!
14:06 pinesol_green Launchpad bug 1478712 in Evergreen "ACQ: Fiscal Year Close-out operation does not check org unit" [Undecided,New]
14:07 Dyrcona jihpringle: I don't think close out caused the transfer. They mistakenly transferred from the closed fund. Apparently, it lets you do that.
14:09 jeff Dyrcona: what were you trying to say with your typo'd line?
14:09 Dyrcona They wanna know if SIP pays lost items or overdues first.
14:10 jeff ah.
14:10 jeff "they"?
14:10 Dyrcona The answer is, "It pays bills in whatever order they come out of the database, so mostly oldest to newest, depending on the state of the tables....basically random order."
14:11 Dyrcona They --- people I work with/for.
14:11 jeff got it.
14:11 jeff wasn't sure if i was missing context that was available elsewhere (looked at list).
14:11 Dyrcona Nah, I'm just playing pinball today, like most Wednesdays.
14:12 jeff what is their reason for asking, and/or their preference?
14:15 Dyrcona Well, it's complicated and has to do with Evergreen SIP not supporting fine item details properly.
14:16 jeff i'll take some blame for that.
14:16 Dyrcona It is coming up because you can't pay a specific bill via Evergreen SIP and any existing self check.
14:16 Dyrcona Heh.
14:16 Dyrcona It's on our RADAR to fix it, soonish.
14:16 Dyrcona I don't think it would be very difficult the required info is there.
14:17 jeff i tried, but once the vendor making the spec stated that they had no way to handle the spec and stated that they would prefer unformatted text, it was a bit of a losing battle / argument to spend more time on it, especially since i had nothing to then test with :P
14:17 Dyrcona Just a matter of do you prefer the 3M format or the Envisionware format.
14:17 * Dyrcona sympathizes. :)
14:21 Dyrcona OK. Back to acq... I think I'll peruse the fund transfer code to see what it looks like I need to do.
14:23 miker berick: aiui, authority 18x does not control other authorities, they control bibs. they're for when a pre-composed heading with the subdivision doesn't already exist. let's you build a bib heading from a 150 and a 18x. or, that's how it's designed to work today.
14:23 Dyrcona See, simple. :)
14:23 miker heh
14:24 berick miker: ok, i was wondering about that..  so it's there in case the 1xx authority record has not yet caught up
14:24 miker now, they can control other authorities, too
14:25 miker auth-auth linking
14:25 berick miker: right, ok.  and we don't link bibs to 18x auths
14:25 berick we just use the 18x data for auth propagation?
14:25 berick (to bibs)
14:25 miker well, you could... by hand
14:26 miker a cataloger could add a $0
14:26 miker but, right. we don't attempt that by machine today
14:27 berick now if a cataloger were to do that, wouldn't it require the bib field in question to link to multiple auth records?  a 1xx link for the main heading, plus then 18x link ?
14:27 miker fwiw, I don't recall ever seeing an in-use 18x in the wild in an actual ILS
14:28 berick ok, yea, that last question was mostly just me being curious
14:28 Dyrcona I know this'll go over like a lead zeppelin, but anyway.... scottmk++ # For comments.
14:28 miker yes, that's what that would mean. or, a "local" uncontrolled main heading and a controlled subdivision
14:29 miker Dyrcona: seconded
14:29 berick miker++ thanks!
14:29 berick @praise scottmk
14:29 * pinesol_green scottmk is the hardest working person in #evergreen.
14:30 berick still!
14:31 * Dyrcona thought he left under a cloud.....
14:31 Dyrcona Anyway, good comments on acq.transfer_fund().
14:52 Dyrcona So, it looks like the easiest thing would be to do what jihpringle suggested and just transfer the money back.
15:11 Dyrcona Awesome sauce!
15:12 Dyrcona I have a test db with a dump from last week loaded and I found a similar transfer from an inactive fund to an active one that I can try this out on.
15:26 bmills joined #evergreen
15:41 rjackson_isl gotta love circs with due dates manually entered for beginning of 2016 and items getting marked Lost with max fines accrued the night of the circ :(
15:44 Dyrcona :)
15:45 rjackson_isl can I get an Amen?!
16:01 JBoyer Dyrcona, Am I correct in remembering that you have a script that can import records from a file and replace existing bre.marc based on the 901c in the modified records?
16:01 JBoyer "modified" being the records in the file.
16:03 Dyrcona JBoyer: You are mostly correct, yes.
16:03 Dyrcona Just a second and I'll post the git URL.
16:04 Dyrcona git://git.mvlcstaff.org/jason/backstage.git
16:04 JBoyer Thanks. I wanted to load these bibs a while ago and it seems I can take something mostly correct and adjust it much faster than I can start from an empty file and build the whole thing. :/
16:05 Dyrcona Yep.
16:06 jeff today's amusing string of text on a screen: CLIFFORD THE BIG RED ARTIST NOT PROVIDED
16:06 Dyrcona If that git server ever shuts down, I'll probably move that to github.
16:06 JBoyer I knew Dali liked dogs.
16:06 Dyrcona heh
16:07 JBoyer Dyrcona, at the risk of bringing it to anyone's attention, if you or tsbere don't turn it off, will anyone else even know it's there?
16:07 jeff also, "THAT IS NOT A GOOD WILLEMS, MO"
16:08 Dyrcona JBoyer: I've advertised it to the community a few times, but no one seems to notice.
16:08 Dyrcona I did move a couple of things from there to github when I left MVLC.
16:09 JBoyer I meant people that could potentially shut it down. :) And I also apparently keep forgetting about it. I knew if you had what I was talking about it would be for Backstage.
16:13 Dyrcona Ah... Yeah, I think they do know.
16:22 Dyrcona JBoyer: I've got a fun, but slow, one for mergin 856s into existing records or creating new ones that I wrote for C/W MARS.
16:23 Dyrcona That's actually what I've been trying to get working with threads and grumbling about with Bmagic. ;)
16:23 Dyrcona It's not shared publicly, yet, though.
16:28 Dyrcona I love documentation lines like this: It is probably not the command you want to use; see "save" above.
16:29 Bmagic agreed, those are sweet
16:34 Dyrcona jihpringle++
16:34 Dyrcona jihpringle: We told them to do what you suggested and it worked. I also know how to reverse it in the database, now. :)
16:35 Dyrcona "There are not the commands you are looking for." *hand wave*
16:48 Dyrcona Another gem: "However, a file generated in this manner will have some items that should not be preseeded, and the example file is a better starting place for most users."
16:49 Dyrcona You just want to say, "So why tell me about it, then?"
16:49 Dyrcona 'Nix guys..... ;)
16:51 Dyrcona Well, I'm heading home. TTYT!
17:00 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
17:08 bmills joined #evergreen
17:13 pinesol_green [evergreen|Galen Charlton] LP#1651808: avoid a class of intermittent search failures - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f67b07b>
17:19 jvwoolf left #evergreen
17:37 bmills joined #evergreen
17:44 jeff i really shouldn't be thinking about the ability to merge copies.
17:57 _adb left #evergreen
18:03 NawJO joined #evergreen
18:54 bmills joined #evergreen
18:57 cbush06 joined #evergreen
19:40 bmills joined #evergreen
20:07 Christineb joined #evergreen

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat