<yosi___> how do I call in from an internal MIT line?
<timbl> Oh good
<timbl> Dial 9 then 617 761 6200
<timbl> Then the code SWCWM
<timbl> Good, Yosi, You'll be able to describe the integration work yuo did
<DanC> bonus points if you can describe it to the point where somebody else can reproduce your results
<timbl> I'm most annoyed that I didn't set up vlad for CVS access
<timbl> ACTION: Tim to get Vlad cwm CVS access [CONTINUES]
<timbl> ACTION: TimBL to work on --why toward log:includes, log:conclusion for rein [CONTINUES]
<timbl> best way to reproduce is cvs update; make
<DanC> make what? what results should I expect?
<timbl> Well, depends whether yosi has it all integrated in tetsts and mainlined
<timbl> Well I'm paying £15 for this connection I might as well update too :-/
<timbl> Dan K, Vlad, Jordan
<vkolovski> hello
<timbl> waiting fro Yosi and DanC to join
<DanC> <DanC> Zakim, who's on the phone?
<DanC> <Zakim> On the phone I see TimBL, [UMD], DanC, Yosi
<DanC> <DanC> Zakim, [UMD] holds eikon, vlad, yarden
<DanC> <Zakim> +eikon, vlad, yarden; got it
<DanC> who was it that volunteered to scribe?
Me
<scribe> scribenick: eikeon_
<DanC> thanks
Vlad finished functional builtins - math / str
<scribe> Scribe: Yarden
Me is Yarden - I am using DanK's computer
<vkolovski> ?x a :testValue. ?y a :testValue.
<vkolovski> (?x ?y) math:sum ?z.
<DanC> easy enough to type: /nick Yarden
<vkolovski> math:sumOf
<vkolovski> math:quotientOf
Can have variables in lists - builtins that generate values (e.g. sum)
<DanC> +1 remove sumOf and the like
Vlad: can be easily extended to
string builtins
... trivial, in fact
... for every list, we keep both the rdf representation of a
list (e.g. first/rest) and also a python list
representation
... had to change builtin representation to handle more than
relational (i.e. true or false) builtins
Tim asks if we still use the cwm interface
for builtins
Vlad: replies, yes
Tim: no point in testing reverse functions since they are not part of the regression suite
<DanC> ACTION: DanC to kill sumOf
<DanC> "the svn"... hmm...
Vlad: code in svn: svn.mindswap.org/pychinko
<scribe> ACTION: Vlad to stick calls to evalObj, with queue, bindings, etc. set to None, at the right places in pychinko ... at which point the math tests should pass [DONE]
Vlad: pychinko code without builtins will run the same - don't know exactly hwo efficient builtins with lists are yet
svn co http://svn.mindswap.org/pychinko
<DanC> $ svn checkout http://svn.mindswap.org/pychinko
<vkolovski> Yosi, the code is in nodes.py, interpreter.py, helpers.py, builtins.py
<vkolovski> do a diff
<DanC> $ python run_tests.py
<DanC> Traceback (most recent call last):
<DanC> File "run_tests.py", line 5, in ?
<DanC> from pychinko.ut.termtests import *
<DanC> ImportError: No module named pychinko.ut.termtests
<DanC> $ PYTHONPATH=.. python run_tests.py
<DanC> ImportError: No module named rdflib.Literal
<yosi___> svn checkout http://svn.mindswap.org/rdflib
<yosi___> rules/sameVarRules5.n3
<vkolovski> python
<vkolovski> main.py --facts=generatedTests/smallTest.n3 --rules=rules/sameVarRules4.n3
<DanC> $ PYTHONPATH=..:../../rdflib/trunk/ python run_tests.py
<DanC> ImportError: No module named llyn
<DanC> aha!
<DanC> Running all Pychinko testcases
<DanC> Generating random facts into generatedtests/testfacts.200.n3
<timbl> agernda is 1) picking up pychinko install by Danc 2) cwm tests question 3) rdflib fork
<DanC> Ran 20 tests in 19.702s
<DanC> FAILED (failures=3, errors=1)
<DanC> (happiness... there is now an effective pair on pychinko at UMD)
DanK: working on diff proj - has new version of rdflib out, some bugs fixed
<DanC> # zope interfaces -- human.txt Daniel Krech (Wednesday, 14 September)
<DanC> ACTION: DK to send notes on Zope Interfaces to public-cwm-talk [DONE]
Yosi: bug related to parseType='Literal' - what's hte status?
DanK: believe it might have been fixed in this release, not sure
DanC: have been doing working group stuff
<DanC> minutes 31 Aug
<DanC> ACTION: DanC to work with timbl to separate operators from term.py, look at other class organization [CONTINUES]
<DanC> oh, I did get this much done: ACTION: DanC to copy http://inferenceweb.stanford.edu/registry/IE/CWM.owl to w3.org/swap space
<DanC> ACTION: DanC to copy http://inferenceweb.stanford.edu/registry/IE/CWM.owl to w3.org/swap space [DONE]
Tim: I've been looking at PML - complication with rdf:collection
<DanC> ... RDF/XML serialization
<DanC> ACTION: dlm to get cwm generating PML that can be seen in iwbrowser, with timbl [CONTINUES]
<DanC> ACTION: Yosi to see if the queue, bindings, proof, query args are needed in the evalObj interface [DONE]
<DanC> YS: none is used
<DanC> YS: well, proof was used in one built-in, but not in a way that works
Tim: yosi, what about pychinko with CWM formula?
<DanC> TBL: I thin proof is needed, but I'd be happy to see the others go
YS: Found new way to get CWM to
use Rete
... original pycwmko.py got stuff into pychinko, ran pychinko,
retrieved output.
... took inferences into formula and kept going
... after talking to UMD peeps in F2F, the new pycwmko shares
Rete
DanC: challenge, use CWM to process my calendar
YS: is there a public version?
DanC: not right now - perhaps
could be scrambled
... can put it on my TODO list, unlikely to be done soon
YS: using Rete almost exactly the same way as Pychinko does ; so more tightly integrated
<DanC> cf " the slow part of my calendar sync" from http://www.policyawareweb.org/2005/pf-dev/08-31-paw-minutes
YS: checked in to CWM cvs
<DanC> checked in where?
YS: Might work on less things
than older pycmko.py
... due to compound terms
<yosi___> http://www.w3.org/2000/10/swap/pycwmko.py
<DanC> pycwmko.py
<DanC> good that it's checked in, but nary a "this end is up" comment
<yosi___> I'm sorry about that, DanC
DanC: registered CWM with IW - need to check with Deb M.
<DanC> 2000/10/swap$ less util/cwm-iweb.rdf
<DanC> I sent mail to Deb M. (copy paw-team)
<timbl> ___________________________________________
<DanC> Ran 20 tests in 19.702s
<DanC> FAILED (failures=3, errors=1)
Vlad: what is the error
exactly?
... Perhaps it's not finding cwm source and cannot do
comparison
<DanC> FAIL: testkeysToList (pychinko.ut.nodetests.TestJoin)
<DanC> ----------------------------------------------------------------------
<DanC> Traceback (most recent call last):
<DanC> File "/home/connolly/paw-mindswap/pychinko/pychinko/ut/nodetests.py", line 34, in testkeysToList
<DanC> self.assert_(i in result)
<DanC> AssertionError
hm this is new
<DanC> that's the last failure
<DanC> 1st one is
<DanC> FAIL: testJoinOneShared (pychinko.ut.nodetests.TestJoin)
<DanC> and 2nd is
<DanC> FAIL: testJoinTwoShared (pychinko.ut.nodetests.TestJoin)
<DanC> ImportError: No module named lex
<scribe> ACTION: check nodetests.py to Yarden
ply - need yacc.py and lex.py i believe
<vkolovski> (sorry about that)
<DanC> no need to apologize; it's documented
make
make: *** No rule to make target `sparql/detailed.tests', needed by `quick'. Stop.
<DanC> that's the cvs -d problem. sigh.
<DanC> (perhaps worth noting at http://www.w3.org/2000/10/swap/doc/cwm#dev )
<DanC> (even though it's really a cvs usability problem)
<DanC> I installed ply, but "import lex" still fails.
cd test; make is now working for me.
<DanC> I think "ply-1.6$ python setup.py install --home=~" didn't install anything
<DanC> "PLY consists of two files : lex.py and yacc.py. To use the system,
<DanC> simply copy these files to your project and import them like standard
<DanC> Python modules."
<DanC> .
<DanC> .
<DanC> ----------
<DanC> rdflib fork
<DanC> I can't figure out what to use for CWMSOURCE in config.py ...
<DanC> pychinko/builtins.py has from llyn import RDFStore
<vkolovski> DanC, I use the cwm/swap path
<DanC> but cwm.py has from swap import diag
<vkolovski> http://www.w3.org/DesignIssues/N3Logic.html, got it
Tim: check whether it makes sense
to you - haven't been reviewed yet
... if something is missing let me know
... wrap of CWM stuff, asked DanC to keep in touch with
Vlad
<timbl> :)
wrap up of*
Well, he'll see the minutes :)
<timbl> Do you want to stay on and chat with UMD re pychinko install?
<timbl> Meeting is adjouned
<timbl> danc
<timbl> darwinports.org?
<timbl> http://darwinports.opendarwin.org/
<DanC> um...
<DanC> I use fink about once every 6 months, and it always rots w.r.t. apple updates since last time I used it
""" DarwinPorts -- A BSD ports system-like implementation for Mac OS X. Most major Linux/BSD tools are in there, works much better than Fink for me. I use this to get Subversion, wxWidgets, etc etc."""
<DanC> quoting from where?
An email someone sent me with a bunch of app recommendations... could probably forward it on.
130/140 sparql/detailed.tests#t1001 sparql builtins from cwm
Files differ, result= 256
# If this is OK, cp ,temp/sparql_query1_result.rl sparql/query1_result.rl
<timbl> Did the cwm run crash with a python error, or did the files not match, eikeon?
<timbl> Maybe you don't have Yosi's latest sparql stuff ... more update -d ?
Didn't see a stack trace.
<timbl> Then diff ,temp/sparql_query1_result.rl sparql/query1_result.rl
<DanC> ACTION: YK to schedule a meeting with TimBL and Bijan to discuss n3 semantics [CONTINUES]