Skip to main content
Version: 8.1

Technical architecture

Introduction​

The Data Capture architecture is based on a modular and extensible structure, designed to facilitate Customization of formatted file processing

  • Management of processing flows.

  • Integration with third-party systems.

  • Adaptability to regulatory and functional changes.

This architecture includes the components of the processing cycle, the associated workflows, and the technical dependencies necessary for its proper functioning in the Axelor environment.

Overview of the Data Capture module architecture​

The Data Capture module is based on a modular architecture designed to ensure automated, reliable, and scalable processing of formatted files. It is structured around a five-step cycle: Receipt, Extraction, Mapping, Control, and Generation.

This process ensures seamless data integration, from file arrival to the creation of the business object in Axelor Open Suite (AOS). Each step is based on specialized components that can be configured according to business needs and file format.

Functional diagram of the workflow​

The Data Capture workflow illustrates the path of a document—whether it is a JSON, EDI, Factur-X, PDF, or image file—to the automatic generation of a business object (invoice, purchase order, delivery, etc.).

The goal is to minimize human intervention while maximizing data quality and interoperability with the ERP.

Processing cycle​

  1. File receipt. The first step in the process is to retrieve the files to be processed. Several modes of receipt are supported:

  • Manual upload by a user.

  • Automatic retrieval from: a local folder / a remote server via SFTP / an external API.

  • Support for single files / batch processing possible via ZIP archives containing multiple files.

  1. Data extraction. Once the file has been received, its content is analyzed to extract usable data:

  • For structured formats (JSON, EDI, Factur-X): data is extracted directly according to the type of application.

  • For unstructured documents (PDF, image, receipt): an OCR connector such as Mindee or Rossum is used to extract the information.

  1. Data mapping

  • The extracted data is aligned with the target structure defined by the schema.

  • Application of mapping rules and correspondence tables to convert external values into internal formats (e.g., currencies, units, VAT rates, supplier codes).

  1. Data control and validation. Before being integrated into the ERP, the data undergoes a verification process:

  • Application of business control rules defined in the template, via Groovy scripts.

  • Verification of data consistency and validity.

  • Generation of detailed logs, differentiating between blocking errors and simple warnings, for accurate diagnosis and effective monitoring.

  1. Generation of business objects. The validated data is used to automatically create the corresponding objects in Axelor Open Suite:

  • Customer or supplier invoices

  • Orders

  • Delivery notes

  • Expense reports, etc.

Post-processing can also be performed:

  • Automatic reconciliation with existing documents,

  • Triggering of workflows,

  • Sending of notifications,

  • Data enrichment through additional business rules.

This architecture is designed for scalability. Thanks to its modular architecture, the Data Capture module ensures:

  • Native integration with AOS;

  • Great flexibility in adapting to specific formats and business processes;

  • Easy interoperability with third-party systems via API or SFTP.