[developers] Selecting start symbols in web demo

Stephan Oepen oe at ifi.uio.no
Tue Aug 18 21:49:34 CEST 2009


hi berthold,

> does anyone know how to make use of the "allow" option in the LOGON web
> demo? 
> 
> I tried, using the ERG, but it seems even with fragments unchecked,
> fragment analyses  still pop up, i.e. I can't see no difference.  

it appears that the selection of root symbols in the ERG on-line demo
is currently broken.  i will need to look at this problem further ...

the general mechanism is /relatively/ simple.  there is a web setting:

  (setf *www-roots*
    '(("sentences" lkb::root_strict "$root_strict" t)
      ("fragments" lkb::root_frag "$root_frag" nil)
      ("infinitives" lkb::root_inffrag "$root_inffrag" nil)
      ("slang" lkb::root_informal "$root_informal" nil)
      ("errors" lkb::root_robust "$root_robust" nil)))

the first() in each case is the string shown on the web page; the next
two fields are the names of the root in the LKB and PET; finally, the
last field determines which root(s) to activate initially.

when parsing an input through the on-line interface, [incr tsdb()] has
support to pass a so-called `custom' string to the processor, which in
PET is used to establish an `overlay' to the grammar-specific settings
(i.e. select parts from the `.set' file(s)).  the web interface makes
use of this facility, to pass a custom string like the following:

  (tsdb::process_item 262147
   ((:p-input . "(42, 0, 1, <0:2>, 1, \"Kim\", 0, \"null\") \
                 (43, 1, 2, <4:8>, 1, \"barks\", 0, \"null\") \
                 (44, 2, 3, <9:9>, 1, \".\", 0, \"null\")")
    (:edges . 20000) (:i-input . "Kim barks.") 
    (:parse-id . 0) (:i-id . 4))
   0 5 nil "start-symbols := $root_frag.")

i see that the custom information is sent alright, so it appears recent
(t)cheap binaries fail to utilize this information appropriately.  this
is maybe an unwanted side-effect from our recent grand SVN merge ...

--- i hope i can still work this out for the imminent Barcelona release
of the LOGON tree; i will post an update to the `logon' mailing list in
a few days then.

                                                          best  -  oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++     CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++       --- oe at ifi.uio.no; oe at csli.stanford.edu; stephan at oepen.net ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++



More information about the developers mailing list