Skip to content


anonymous edited this page Oct 9, 2011 · 7 revisions

Jacy uses the light weight version of the lexical database (Single User mode). (2006-09-06: incomplete initial documentation --- details may change)


  1. Database Fields
    1. TDL Fields
    2. All Fields
  2. Lexicon Interaction with CVS
  3. Set Up
  4. ToDo
  5. Priveleges

Database Fields

TDL Fields

Defined in japanese/lex/lexdb.dfn.

id      name            sym
orth    orthography             str-rawlst
unifs   orthography     (orth)  str-dlst
unifs   type    nil     sym
unifs   keyrel  (synsem lkeys keyrel pred)      mixed
unifs   keytag  (synsem lkeys keyrel carg)      mixed
unifs   altkey  (synsem lkeys altkeyrel pred)   mixed
unifs   alt2key (synsem lkeys alt2keyrel pred)  mixed
unifs   altkeytag       (synsem lkeys altkeyrel carg)   mixed
unifs   compkey (synsem local cat head ptype)   mixed
unifs   idiom   (idiom) mixed

All Fields

Defined in japanese/lex/su-init.sql

field SQL type Comment Example
name TEXT NOT NULL TDL lexical-id (unique)
userid TEXT DEFAULT user NOT NULL Maintainer ID
modstamp TIMESTAMP Time of last change
type TEXT TDL lexical type
orthography TEXT TDL Orthography
keyrel TEXT TDL Key (Predicate)
altkey TEXT TDL ALT Key
alt2key TEXT TDL ALT2 Key
compkey TEXT TDL ptype
idiom TEXT TDL Idiom Feature
pronunciation TEXT Pronunciation
comments TEXT Any comments
lang TEXT Ja
country TEXT JP
dialect TEXT --
confidence real DEFAULT 1 Reduce for autoamtic additions
source TEXT, Which lexicon it came from

Lexicon Interaction with CVS

  • Before updating the lexicon file lex/Jacy.rev

  • * do your normal checks

  • ** (lkb::batch-check-lexicon)

  • ** dump to tdl and flop

  • * optionally back up your existing lexicon

  • * dump the lexical database

     psql jacy
     jacy=# \copy lex to japanese/lex/Jacy.rev 
  • Merge using normal CVS update

  • * Check things again

  • ** load the revisions, reindex, run the grammar etc

  • ** dump the tdl

  • Update the Jacy.rev and tdl lexicon

Set Up

Set up the database as described in the Single User page, then load the lexicon from CVS.

  • Connect to the database (normally called jacy)

     psql jacy 
  • Create the Databases

     jacy=# \i japanese/lex/su-init.sql 
    • Load the definitions
     jacy=# \copy dfn from  japanese/lex/lexdb.dfn 
    • Load the lexical entries
     jacy=# \\copy lex from japanese/lex/Jacy.rev 
  • Load Jacy and set up the *lexdb-params* through the LKB Options::Set Options... menu

     ((:DBNAME "jacy") (:TYPE :SINGLE-USER)) 
  • Index for parsing and generation

     (lkb::new-lex-key-table lkb::*lexicon*)
     (lkb::new-semi lkb::*lexicon*)


  • currently we can't delete entries, so we mark anything we want to delete with the string delete in the comment field.


How to allow multiple people to access the same DB:

jacy=#  GRANT ALL  ON  lex_key to bond;
jacy=#  GRANT ALL  ON  semi_extra to bond;
jacy=#  GRANT ALL  ON  semi_frame to bond;
jacy=#  GRANT ALL  ON  semi_mod to bond;
jacy=#  GRANT ALL  ON  semi_obj to bond;
jacy=#  GRANT ALL  ON  semi_var to bond;
jacy=# \dp
                         Access privileges for database "jacy"
 Schema |    Name    | Type  |                    Access privileges                     
 public | dfn        | table | {bond=arwdxt/bond,kuribayashi=arwdxt/bond}
 public | lex        | table | {bond=arwdxt/bond,kuribayashi=arwdxt/bond}
 public | lex_key    | table | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
 public | semi_extra | table | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
 public | semi_frame | table | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
 public | semi_mod   | table | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
 public | semi_obj   | view  | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
 public | semi_pred  | table | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
 public | semi_var   | table | {kuribayashi=arwdxt/kuribayashi,bond=arwdxt/kuribayashi}
(9 rows)