Apply control & enrichment
[To be released Q1 2025] Learn here how to apply our specific enrichment & control fonctions on your extracted data
Last updated
[To be released Q1 2025] Learn here how to apply our specific enrichment & control fonctions on your extracted data
Last updated
Think of functions as small AI-powered robots that execute specific control or enrichment tasks on the extracted fields.
Unlike formats, which modify the value extracted from your document, functions generate additional data points called insights.
You can choose to activate one or more functions directly within the extraction template for each general field or repeated field.
Functions are essential for performing operations that requires an additional tool, such as executing complex calculations through a Python script, or matching extracted values against an extensive database, or even conducting a Google search.
You can activate functions in the extraction template, just below the format field.
The database_match feature allows you to reconcile the extracted data with a corresponding entry in your database.
You can upload your database directly into the extraction template in either XLSX or CSV format.
If you upload an Excel format (XLSX), only the first column of the first tab will be processed. For CSV files, the entire column will be considered.
You can specify the type of matching algorithm:
If exact_match is activated, will search for an identical match between the extracted value and an entry in your database. This is case-sensitive and requires a perfect match for the reconciliation to succeed.
If similar_match is activated, the system will search for a close match between the extracted value and an entry in your database. It ignores case sensitivity (“JOHN SMITH” → “John Smith”), punctuation (“J. Smith” → “J Smith”), and minor variations, while also accounting for common OCR errors (e.g., “O” → “0” or “W1lliams” → “Williams”).
If smart_match is activated, the system will search for a match by considering the meaning of the extracted value.
You can also activate the mandatory_match option, which forces the function to select the closest entry in the database. This ensures that a result is always returned.
The context_match uses several extracted values to find the best match within the database.
You can activate the mandatory_match option, which forces the function to select the closest entry in the database. This ensures that a result is always returned.
The iban_check function performs a series of validations to ensure the integrity and validity of an IBAN:
Checksum Validation: The function applies the MOD97 algorithm to verify the IBAN’s check digits, ensuring mathematical integrity.
Format and Structure Validation: It checks the country-specific IBAN format, structure, and character positions, including bank codes, branch identifiers, and account numbers.
Length Verification: The function ensures the IBAN matches the exact length required for its corresponding country.
BBAN Checksum (if supported): For countries supporting internal account number validation, the function performs an additional local checksum verification.
The output values are valid or invalid.
