Ligature

Intelligently replaces Unicode ligatures in a PDF text extraction.

Parameters

key	value	description
type (required)	ligature
mappings (required)	object	An object mapping ligature strings (for example, “\u0000”) to an array of possible ligature replacements (for example, [“ff”, “ffi”, “fi”, “fl”]). Sensible uses a dictionary in the target language to choose replacements that lead to known words. Sensible supports American English (en-us). This approach is conservative and may leave some Unicode characters in proper names or other non-word data.
forceReplaceAll	false	If true, specifies to bypass the dictionary lookup and to replace each ligature with the first replacement listed in the mappings array. This is useful in situations where words containing ligatures are not in the supported dictionary. For example, they are in an unsupported language or are proper nouns.

Examples

This config shows using a Ligature preprocessor and outputting the whole document to check the ligature replacement accuracy:

JSON

{
  "preprocessors": [
    {
      "type": "ligature",
      "mappings": {
        "\u0000": [
          "ff",
          "fi",
          "fl"
        ],
        "\u0001": [
          "ti",
          "tl"
        ]
      }
    }
  ],
  "fields": [
    {
      "id": "all_lines_in_doc",
      "method": {
        "id": "documentRange",
        "includeAnchor": true
      },
      "anchor": {
        "match": {
          "type": "first"
        }
      }
    }
  ]
}
`