The Library

4:00 PM

@DanO'Day helloooooooooooo

4:19 PM

@swasheck helloooooooooooooo

@DanO'Day have you ever done anything with CoreNLP? or could you point me to a good tutorial that would help me train an NLTK POS tagger?

i've spend my weekend searching for ways to handle either ... and cant find one

@swasheck I have not worked with CoreNLP before (but didn't @MonicaCellio mention experience with it? Perhaps helped develop it, I could be wrong)

nltk.googlecode.com/svn/trunk/doc/book/ch05.html

see that link, is that what you are looking for?

@DanO'Day i may be missing something, but i wanted to train NLTK on Tauber's tagged NT. i am not sure how to set up the training data set

4:38 PM

@DanO'Day sorry, no experience here either. I helped develop an NLP system for tagging, but it was rule-driven, not a learning system.

nltk-trainer.readthedocs.org/en/latest/train_tagger.html

streamhacker.com/2011/03/21/…

4

Q: custom tagging with nltk

I'm trying to create a small english-like language for specifying tasks. The basic idea is to split a statement into verbs and noun-phrases that those verbs should apply to. I'm working with nltk but not getting the results i'd hoped for, eg: >>> nltk.pos_tag(nltk.word_tokenize("select ...

4:54 PM

@DanO'Day thanks. i'll do some reading. i think i'm just missing a crucial cognitive power at this point.

As a general heads-up, we are in the middle of a community evaluation review. When you have some time, it would be a good idea to participate.

@JonEricson fabulous ... just completed min on dba. wanna post a community message?

@swasheck they have removed that power from us!

@JackDouglas even in chat?

We can still pin in chat

(good idea)

It's in the 'Community Bulletin' on the main site, but that is in my advert blind spot :)

5:03 PM

@JackDouglas sorry. i'm abhorrently illiterate in the ways, means, and terminology of SE

@swasheck little yellow box on the top-right of the front page

@swasheck which probably proves you have a life ;)

@JackDouglas ah. that thing. modifying that's been removed?

@JonEricson it looks like this is currently happening on all sites, by the way -- so after you finish here, check out other sites you care about.

@MonicaCellio read: stay away from SO

;)

@swasheck we used to be able to add a banner on the main site that shifted all the content down: unmissable. Far too much power for mods :)

5:09 PM

@JackDouglas well i clearly don't pay much attention to it because i didnt know what it was called :)

@swasheck I'd like to be able to ask a site what its 10 questions are without having tos tep through them all and vote (for the ones where I'm curious but not enough of an expert to really judge).

@MonicaCellio I just finished the Christianity eval. I looked at Mi Yodeya and I can't make heads or tails of the first question! That almost certainly means it's Excellent, but it would be wrong of me to skew the stats. ;-)

@JonEricson :)

@JackDouglas Same here. It does include more signal than noise, however.

5:25 PM

@swasheck check out this guy's blog too: thelibrarybasement.com/tag/nltk

@DanO'Day yes i've been reading through that quite a bit, too

thelibrarybasement.com/2012/02/10/koine-greek-wordlists

@JonEricson yeah, I can see that. "Skip" is one of the options if you want to get past it and try another. (Or sometimes reloading a review page gets you a different thing to review.)

A: What does the Bible say about life before conception?

5:49 PM

Let's see if a good dose of history goes over well on this question :P

0

The only clear reference is in the apocryphal book Book of Wisdom 8:19-20: As a child I was naturally gifted, and a good soul fell to my lot; or rather, being good, I entered an undefiled body (NRSVCE). The author of this book was almost certainly heavily influenced by Middle-Platonism. S...

@DanO'Day nicely done.

@MonicaCellio something tells me I may have wasted my breath, but meh ;)

6:43 PM

@MonicaCellio Yeah. That's what I'll probably do.

A: How authentic is Codex Sinaiticus?

10:47 PM

Frank bringin' it

3

6

To answer your first question, we should not simply accept Sinaiticus as the source of the truth for the New Testament. It has great weight in debates from its age, but age is not the final arbiter in textual considerations. Codex Sinaiticus was made in the 4th century on parchment using capital...

@swasheck yes it is an excellent response

@DanO'Day i'm such an idiot. i'm seriously at a loss trying to figure out how to format my training data. i can see how to do it once it's formatted, but just trying to get it to a place where nltk will like it. ... le sigh. i'm totally out of my depth.

@swasheck first step, do you understand xml and have you looked at any tagged corpora for nltk? Second step, do you understand data structures in python (list, tuple, dictionary)?

I'd start there

@DanO'Day in my spare moments i'm looking for the tagged corpora.

@swasheck nltk.googlecode.com/svn/trunk/nltk_data/index.xml

@swasheck guide to it: code.google.com/p/nltk/wiki/Corpora

ok can you see if stackoverflow.com loads fine for you?

it is only loading content for me, no design

11:02 PM

Tilghman/np leapt/vbd on/in to/in him/ppo ,/,

@DanO'Day looks good to me

@DanO'Day so this is a tagged file? i set up the NT like this with Tauber and train NLTK on that?

Tilghman/np leapt/vbd on/in to/in him/ppo ,/,

@swasheck usually you tag it programmatically. Use an existing xml corpus of scripture that has morphology and syntax (if there is one) and write a script to tag it in a way nltk can understand

@swasheck someone has to have done this already

My guess is opentext.org

@DanO'Day that's the biggest obstacle to the basement library guy's thing

ah ... yeah

@swasheck yes they are doing it, it is on logos but only in their proprietary format - should be posted soon: opentext.org/resources/articles/a8.html

@swasheck I'm just waiting for them to post it all: opentext.org/model/guidelines/wordgroup/0-1.html

@DanO'Day so then you'd train NLTK on it? or would this be something more usable?

@swasheck here it is: opentext.org/texts.html

dangit no

that is not the xml

why is it available visually but not the xml files?

weird

11:12 PM

@DanO'Day yeah.

doesnt look like a fun HTML scrape

they are supposed to have one (PROIEL): hf.uio.no/ifikk/english/research/projects/proiel

looks promising: greekings.wordpress.com

github.com/janabeck/PPCHiG

@DanO'Day yes ... jana is someone i've been following ... but she's a one-person shop as far as i cantell

so PROIEL may be it for the most part: lists.ibiblio.org/pipermail/b-greek/2010-March/052867.html

hey i gotta run, ttyl @swasheck

@DanO'Day later

thanks.

@swasheck we may have to start our own project if nothing else....

11:17 PM

@DanO'Day i just dont understand the internals, otherwise i would have already

@swasheck but shouldn't have to when there is an abundance of tagged ones out there, but all in propietary bible software... arg

@DanO'Day different story. i assume that was the purpose of morphgnt but again ...

@swasheck i've tagged corpora with affect and behavioral markers, i rely on pos-tagging in nltk to do the syntax tagging or use established pre-tagged corpora. Might be time for me to learn too ;)

@swasheck ttyl