Transformation is  in, or out, transmission. The in-transmission control process takes received information in the form of data, document, image or voice, applies any conversion or mapping and produces flat, formatted text files as output; the out-transmission process takes the flat, formatted text files and applies conversions to produce the required output format.

In-Transmission

In-transmission is a series of processes under an intelligent control process. As messages are received, the control process calculates which process(es) should be applied in order to transform the message in to a data format (e.g. XML, fixed length text or delimited text).

After receiving the message from the communications adaptors, the in-transmission process establishes the format of the content. With some formats, converters are used to flatten the structure and then maps are applied to extract data from geographic structuring. Other formats use either a converter or a map.

Data received in the form of XML, EDI, Text or SMS, requires no additional conversion or mapping and the message passes through the in-transmission process and on to the syntax ontology.

Documents such as Word or Excel are converted to CSV before being passed on to the next process. No mapping is required as the delimited structure of CSV can be read in the syntax ontology. Other documents such as geographic text files (neither fixed length nor delimited) and PDFs with a text layer, are not converted, rather intelligent mapping is used to extract text as data and render it as XML.

Images, for example Fax or PDFs without a text layer, are required to pass through both a converter and a map. Having established the received format is an image, the control process passes the message to an optical character recognition (OCR) engine to convert it in to data. The control process establishes that further mapping is required to transform the OCR data content into meaningful data, and passes it to the visual data processing (VDP) process. VDP has been developed by OmPrompt and is a library of visual ontologies (VOs) which automatically recognise data structures (e.g. addresses, dates, product descriptions, reference numbers, etc.) within image data and equates them to expected data values. Having established the structures, the data is output as XML.

Voice, received as a WAV and XML file requires no conversion or mapping. Both formats pass through the in-transmission process but only the XML file goes into the syntax ontology while the WAV file is held in its original form as a supporting record.

Out-Transmission

The out-transmission process calls various converters according to the required output format.  As input from the outbound syntax ontology is already structured it requires no mapping before being passed to the conversion process and output to the communications adaptors.

The required format(s) for output are pre-configured and the control process calculates which CSV converters are to be used.