How to extract only TXs details from a bank statement

[herou]

Hi all,
I am building something and I need to extract TX details from the bank statement below and store them in my DB.
Can I do it with stanfordnlp?
The bank statement below is just an example. But can be hundert of them.
For me the most interesting part is: amount, name date,

CUSTOMER NAME										 	BRANCH NAME 		
John Smith										 	<Branch Name>		
ADDRESS										 	CARD NUMBER 		
"2450 Courage St, STE 108
Brownsville, TX 78521"										 	2547-96574-2541-4004		
										 	ACCOUNT NUMBER		
											1234567890 		
OPENING BALANCE		TOTAL CREDIT		TOTAL DEBIT		CLOSING BALANCE			 		PERIOD		PAGE
10,895.92		22,032.72		1,249.95		31,678.69			 		mm/dd/yyyy		1   of   1
													
Date	Reference		Transaction Description							Credit	Debit	Balance	
mm/dd/yyyy	0302432		Account Interest								19.95	10,875.97	
mm/dd/yyyy	0302433		Branch Deposit							20,000.00		30,875.97	
mm/dd/yyyy	0302434		Branch Deposit								600.00	30,275.97	
mm/dd/yyyy	0302435		Branch Deposit								630.00	29,645.97	
mm/dd/yyyy	0302436		Branch Cheque - MC Kean & Park							1,430.00		31,075.97	
mm/dd/yyyy	0302437		Cost of Bank Cheque							100.00		31,175.97	
mm/dd/yyyy	0302438		Branch Deposit							500.00		31,675.97	
mm/dd/yyyy	0302439		FED Bank Account Debits Tax							1.05		31,677.02	
mm/dd/yyyy	0302440		VIC FID Charge							1.67		31,678.69	
													
													
													
													
													
													
													
													
													
													
													
													
										n			
													
													
													
													
			Total							22,032.72	1,249.95	31,678.69	
												

[AngledLuffa]

That seems closer to just doing text processing on the input than anything else

[herou]

Can you give me some hints what should I do? Examples etc…

[herou]

That seems closer to just doing text processing on the input than anything else

[AngledLuffa]

If the bank statements all follow the same format, you just need to write something that divides the text up in to the appropriate columns once and then apply it to all the statements. CoreNLP isn't really the right tool for that. It's meant to find POS tags, named entities, parse trees for the sentences, etc.