Hello. There are bank statements in pdf files. Several types of structure type (up to 5). All types contain a text layer, but not all libraries recognize text correctly. In case of difficulty in extracting text, I suggest connecting ABBY finereader or opencv cloud recognition https://github.com/abbyy/ocrsdk.com/tree/master/Python https://support.abbyy.com/hc/en-us/articles/360017326319-Programming-languages-Quick-start-with-OCR-SDK-for-Python https://support.abbyy.com/hc/en-us/articles/360017269420-api-reference https://www.abbyy.com/cloud-ocr-sdk/documentation/ Statements can be downloaded from the link - https://www.dropbox.com/t/vcLLkeHW0I7P3RlO Primarily interested in 2022. It is necessary to extract information from PDF, structure and save it in CSV. Attached an example. Read all PDFs from a PDF folder (may have sub-directories) Save each statement in a separate CSV file with the same name as PDF, to the CSV folder. It is desirable to save the folder structure, but it is not necessary. Each entry is a separate line. Notes: - column D must contain only the amount without the currency; - the + sign is not needed for amounts greater than 0; - currency must be in column E; - column D must contain the sum in the format 3.15 instead of 3.15, for example; Source code required. Project is urgent - need ASAP.