Bengali analyzer

The built-in bengali analyzer can be applied to a text field using the following command:

  1. PUT /bengali-index
  2. {
  3.   "mappings": {
  4.     "properties": {
  5.       "content": {
  6.         "type": "text",
  7.         "analyzer": "bengali"
  8.       }
  9.     }
  10.   }
  11. }

copy

Stem exclusion

You can use stem_exclusion with this language analyzer using the following command:

  1. PUT index_with_stem_exclusion_bengali_analyzer
  2. {
  3.   "settings": {
  4.     "analysis": {
  5.       "analyzer": {
  6.         "stem_exclusion_bengali_analyzer": {
  7.           "type": "bengali",
  8.           "stem_exclusion": ["কর্তৃপক্ষ", "অনুমোদন"]
  9.         }
  10.       }
  11.     }
  12.   }
  13. }

copy

Bengali analyzer internals

The bengali analyzer is built using the following components:

  • Tokenizer: standard

  • Token filters:

    • lowercase
    • decimal_digit
    • indic_normalization
    • normalization (Bengali)
    • stop (Bengali)
    • keyword
    • stemmer (Bengali)

Custom Bengali analyzer

You can create a custom Bengali analyzer using the following command:

  1. PUT /bengali-index
  2. {
  3.   "settings": {
  4.     "analysis": {
  5.       "filter": {
  6.         "bengali_stop": {
  7.           "type": "stop",
  8.           "stopwords": "_bengali_"
  9.         },
  10.         "bengali_stemmer": {
  11.           "type": "stemmer",
  12.           "language": "bengali"
  13.         },
  14.         "bengali_keywords": {
  15.           "type":       "keyword_marker",
  16.           "keywords":   [] 
  17.         }
  18.       },
  19.       "analyzer": {
  20.         "bengali_analyzer": {
  21.           "type": "custom",
  22.           "tokenizer": "standard",
  23.           "filter": [
  24.             "lowercase",
  25.             "decimal_digit",
  26.             "indic_normalization",
  27.             "bengali_normalization",
  28.             "bengali_stop",
  29.             "bengali_keywords",
  30.             "bengali_stemmer"
  31.           ]
  32.         }
  33.       }
  34.     }
  35.   },
  36.   "mappings": {
  37.     "properties": {
  38.       "content": {
  39.         "type": "text",
  40.         "analyzer": "bengali_analyzer"
  41.       }
  42.     }
  43.   }
  44. }

copy

Generated tokens

Use the following request to examine the tokens generated using the analyzer:

  1. POST /bengali-index/_analyze
  2. {
  3.   "field": "content",
  4.   "text": "ছাত্ররা বিশ্ববিদ্যালয়ে পড়াশোনা করে। তাদের নম্বরগুলি ১২৩৪৫৬।"
  5. }

copy

The response contains the generated tokens:

  1. {
  2.   "tokens": [
  3.     {"token": "ছাত্র","start_offset": 0,"end_offset": 7,"type": "<ALPHANUM>","position": 0},
  4.     {"token": "বিসসবিদালয়","start_offset": 8,"end_offset": 23,"type": "<ALPHANUM>","position": 1},
  5.     {"token": "পরাসোন","start_offset": 24,"end_offset": 32,"type": "<ALPHANUM>","position": 2},
  6.     {"token": "তা","start_offset": 38,"end_offset": 43,"type": "<ALPHANUM>","position": 4},
  7.     {"token": "নমমর","start_offset": 44,"end_offset": 53,"type": "<ALPHANUM>","position": 5},
  8.     {"token": "123456","start_offset": 54,"end_offset": 60,"type": "<NUM>","position": 6}
  9.   ]
  10. }