Skip to content

OCR Extract Pattern

OCR Extract Pattern

Type: ocr-extract-pattern · Package: OCR Activities v1.0.0 · Output: Dictionary

It combines OCR + regular expressions in a single step: it recognizes the text of an image and, on that text, applies a set of patterns (regex) to extract specific fields. Returns a { campo: valor } dictionary. It is the most direct way to extract structured data from a scanned document (invoice number, date, total).

How it works

Runs OCR on the image and, on the resulting text, evaluates each pattern in the patterns dictionary. For each pattern that matches, add nombreCampo → valor to the result. Use a capture group in the regex to extract only the part you are interested in.

Required parameters

Parameter	Editor	Description
`path`	expression	Image file path.
`patterns`	objeto	`nombreCampo → patrón regex` Dictionary. Use a capture group to extract the value.

Optional parameters

Parameter	Editor	Description
`language`	expression	Language code: `eng` (def.) · `spa` · `eng+spa`.
`preprocess`	expression	Limpieza previa: `none` (def.) · `auto` · `deskew` · `binarize`.

Output

Returns a Dictionary with a pair for each pattern that was matched.

Example

Extract number, date and total from a scanned invoice:

OCR Extract Pattern
  path     = = rutaFactura
  language = "spa"
  preprocess = "auto"
  patterns = {
     "numero": "FAC-(\d+)",
     "fecha":  "Fecha:\s*(\d{4}-\d{2}-\d{2})",
     "total":  "Total:\s*\$?([\d.,]+)"
  }
  → output: datos

Log   message = = "Factura " + datos["numero"] + " — Total " + datos["total"]

OCR Recognize File — get the text and extract it yourself with regexFind.
OCR Extract Form Fields — extract by position instead of pattern.