Input is gathered from its email address and a network shared folder at regular intervals. All documents are converted to .pdf if they did not come in that way. TesseractOCR is then used to fix the rotation on these documents and find key phrases and number combinations that indicate document type, document id, and client or account ids. These pdfs are saved to the searchable indexed PostgreSQL database along with any recognised text and ids.
A simple app to scan bar and QR codes in live view, send the contents to the connected server and retrieve the results. Later repurposed as EthScan for a french association that organises LAN parties for charity, to scan player tickets and display name, any warnings (such as the player being underage and needing an extra permission form, the ticket not having been paid yet, or the ticket having already been scanned), and any notes added by the organisers.