leeyspaul / kengdic

Joe Speigle's Korean/English dictionary database

kengdic is a large Korean/English dictionary database created by Joe Speigle. It was originally hosted at ezkorean.com, and I have posted it here (https://github.com/garfieldnate/kengdic) because it is no longer available anywhere else. It is released under MPL 2.0.

• kengdic.sql is a PostgreSQL database dump of a the Korean/English dictionary as of 2009. It assumes the existence of a modpgwebuser schema, as well as a korean_english_wordid_seq sequence. It contains one table, korean_english with 143,795 rows. There is no primary key, but indexes are created on doe, wordid, and word. This table contains the following information:

  • Korean/English spelling
  • Hanja spellings, comma-separated
  • English definition
  • part of speech
  • source of the entry

• kengdic_2011.tsv is a tab-separated file containing a database dump of the same Korean/English dictionary, but instead of being a SQL file it is simply a tab-separated file containing all of the data. It contains 133,876 rows. As this is probably cleaner, newer, better data than kengdic.sql, it may not be worth cleaning up that file at all.

• create_kengdic.sql creates the kengdic database and sources the data in kengdic_2011.tsv. To run it:

  • Make sure postgres is running on your machine and run this command in the terminal: psql < create_table.sql
    • You may need to provide a user with the -U switch, like so: psql -U username < create_table.sql
  • This should create a database called 'kengdic' with one table called 'korean_english'
  • Run 'psql kengdic' and make sure everything imported correctly: "SELECT COUNT(*) FROM korean_english;" should tell you there are 133,876 rows in the table.

• ezcorean_6000.sql is a database dump of 6000 common Korean words, along with the definitions and hanja.

TODOs for each file:

• kengdic.sql:

  • fix entries where the entries are HTML containing both hangeul and hanja in text body
  • normalize Hanja (have relation table and hanja table separate instead of comma-separating them)
  • find the difference between pos and posn
  • find which numbers indicate which part of speech
  • add comments for all columns and tables
  • replace 'see 6000' in the definitions with the definition from the 6000 list from ezkorean.com
  • add primary key (I'm pretty sure that would be good; there are already indices, though)
  • export to sqlite

• kengdic_2011.tsv:

  • Do something about entries with no definition

Information may be incomplete, as we are still exploring and documenting the contents of this repository. Any contributions to information about this content would be much appreciated.