Currently use intetnally for mednacea project

This class is to semi-automatic segment Thai sentences and create a json file from .tsv (tab-separated values) file. The output is a JSON file.

The first row of .tsv file is considered as the top column.

An input file is needed as a template of a each json object. There are 2 types of structures that this class can handle.

o-NAME-NUMBER

meaning

• o = object
• NAME = Name of Object
• NUMBER = range (e.g., 0-5)or number of column (e.g., 0,1,2,3,4,5) starting from 0

s-NUMBER

meaning

• s = single
• NUMBER = range (e.g., 0-5)or number of column (e.g., 0,1,2,3,4,5) starting from 0

Suppose we have

The command o-product-1-2 or o-product-1,2 will return

{
  "product": {
    "price": "20",
    "name": "ไก่ทอด บอน ชอน"
  }
}{
  "product": {
    "price": "25",
    "name": "ไก่ทอด แสน อร่อย"
  }
}

The command s-1-2 or s-1,2 will return

{
  "price": "20",
  "name": "ไก่ทอด บอน ชอน"
}{
  "price": "25",
  "name": "ไก่ทอด แสน อร่อย"
}

The command

s-0
o-product-1,2

will return

{
  "product": {
    "price": "20",
    "name": "ไก่ทอด บอน ชอน"
  },
  "id": "5"
}{
  "product": {
    "price": "25",
    "name": "ไก่ทอด แสน อร่อย"
  },
  "id": "6"
}

The segmentataion algorithm (namely the class LongLexTo, LongParseTree, Trie and the Lexitron dictionary) belongs to NECTEC under the LGPL License.