Page 1 of 2
Transym OCR wrapper functions for Harbour
Posted: Tue Mar 28, 2017 11:12 pm
by reinaldocrespo
Hello everyone;
I finally got Transym OCR to work from API using (x)Harbour. Here is the code in case anyone is interested in the future.
Code: Select all
//Reinaldo Crespo-Bazan 3/28/2017 6:39:11 PM
//reinaldo.crespo@gmail.com
//reinaldo.crespo@structuredsystems.com
//This is sample code and wrapper functions to use Transym OCR engine.
//www.Transym.com
#define TOCRJOBTYPE_TIFFFILE 0 // specifies a tiff file
#define TOCRPROCESS_OPTIONS_MERGE 0x00000001 //Merge options ON
#define TOCRPROCESS_OPTIONS_DESKEW 0x00000010 //deskew ON
FUNCTION MAIN()
LOCAL TestTifFile := "sample.tif"
IF !FILE( TestTifFile )
Alert( "testfile sample.tif not found" )
return NIL
ENDIF
MsgInfo( OCRFromFileUsingTransym( TestTifFile ), "TOCR dll results using default options" )
//I find MergeBreakOff and DeskOff turned off works better for some documents.
MsgInfo( OCRFromFileUsingTransym( TestTifFile, ;
TOCRJOBTYPE_TIFFFILE, ;
TOCRPROCESS_OPTIONS_DESKEW | TOCRPROCESS_OPTIONS_MERGE, "TOCR dll results with Merge and Deskew OFF" ) )
RETURN NIL
//-----------------------------------------------------------
//-------------------------------------------------------------------------------------
/**/
#pragma BEGINDUMP
#include <hbapi.h>
#include <windows.h>
#include <TOCRdll.h>
#include <TOCRuser.h>
#include <TOCRerrs.h>
BOOL OCRWait( long JobNo, TOCRJOBINFO2 JobInfo2 );
BOOL GetResults( long JobNo, TOCRRESULTSEX ** Results );
BOOL FormatResults( TOCRRESULTSEX * Results, char * Msg );
//--------------------------------------------------------
//parameters
// 1. input file with image
// 2. type of file to ocr defaults to TIFF
// 3. options mask defaults to 0
//returns OCRed text
HB_FUNC( OCRFROMFILEUSINGTRANSYM )
{
TOCRJOBINFO2 JobInfo2;
TOCRRESULTSEX * Results = 0;
long Status;
long JobNo;
byte OptionsMask = ISNUM( 3 ) ? hb_parnl( 3 ) : 0x00000000 ;
char Msg[8192];
char * InputFile = ( char * ) hb_parcx( 1 ); //parm 1 is input file
//Sets Transym to print error to show errors on screen dialog
TOCRSetConfig( TOCRCONFIG_DEFAULTJOB, TOCRCONFIG_DLL_ERRORMODE, TOCRERRORMODE_MSGBOX );
memset( &JobInfo2, 0, sizeof( TOCRJOBINFO2 ) );
JobInfo2.JobType = ISNUM( 2 ) ? hb_parni( 2 ) : TOCRJOBTYPE_TIFFFILE ;
JobInfo2.InputFile = InputFile ;
//if MergeBreakOff and DeskewOff parameters are sent then change options accordingly.
JobInfo2.ProcessOptions.MergeBreakOff = OptionsMask & 0x00000001 ; //1 bit MergeBreakOff
JobInfo2.ProcessOptions.DeskewOff = OptionsMask>>1 & 0x00000001 ; //2nd bit DeskewOff
Status = TOCRInitialise( &JobNo );
if ( Status == TOCR_OK ) {
if ( OCRWait( JobNo, JobInfo2 ) ) {
if ( GetResults( JobNo, &Results ) ) {
FormatResults( Results, Msg );
hb_xfree( Results );
}
}
TOCRShutdown( JobNo );
}
hb_retc( Msg ) ;
}
//--------------------------------------------------------
BOOL OCRWait( long JobNo, TOCRJOBINFO2 JobInfo2 )
{
long Status;
long JobStatus;
long ErrorMode;
char Msg[4096];
Status = TOCRDoJob2( JobNo, &JobInfo2 );
if (Status == TOCR_OK) {
Status = TOCRWaitForJob(JobNo, &JobStatus);
}
if (Status == TOCR_OK && JobStatus == TOCRJOBSTATUS_DONE)
{
return TRUE;
} else {
// If something hass gone wrong extract the error message and log or return
// not doing anything with it for the time being.
// Check that the OCR engine hasn't already displayed a message first.
TOCRGetConfig(JobNo, TOCRCONFIG_DLL_ERRORMODE, &ErrorMode);
if ( ErrorMode == TOCRERRORMODE_NONE ) {
TOCRGetJobStatusMsg(JobNo, Msg);
}
return FALSE;
}
} // OCRWait()
//--------------------------------------------------------
// Get the results from TOCR
BOOL getresults(long JobNo, long mode, void **Results)
{
long Status;
long ResultsInf;
char Msg[4096];
Status = TOCRGetJobResultsEx(JobNo, mode, &ResultsInf, 0);
if ( Status != TOCR_OK ) return FALSE;
if ( ResultsInf > 0 ) {
// Allocate memory for results
*Results = ( char * ) hb_xgrab( ResultsInf + 1 );
// Retrieve the results
Status = TOCRGetJobResultsEx(JobNo, mode, &ResultsInf, *Results);
if ( Status != TOCR_OK ) {
hb_xfree( Results );
*Results = 0;
return FALSE;
}
} else return FALSE ;
return TRUE;
} // getresults()
//--------------------------------------------------------
// Get extended results
BOOL GetResults( long JobNo, TOCRRESULTSEX **Results )
{
return getresults( JobNo, TOCRGETRESULTS_EXTENDED, (void **)Results );
} // GetResults()
//--------------------------------------------------------
// Convert extended results to a string
BOOL FormatResults(TOCRRESULTSEX *Results, char *Msg)
{
long ItemNo;
long APos = 0;
BOOL Status = FALSE;
if ( Results->Hdr.NumItems > 0 ) {
for (ItemNo = 0; ItemNo < Results->Hdr.NumItems; ItemNo ++ ) {
if ( Results->Item[ItemNo].OCRCha == '\r' )
Msg[APos] = '\n';
else
Msg[APos] = (char)Results->Item[ItemNo].OCRCha;
APos ++;
}
Msg[APos] = 0;
Status = TRUE;
}
return Status;
} // FormatResults()
#pragma ENDDUMP
Reinaldo.
Re: Transym OCR wrapper functions for Harbour
Posted: Wed Mar 29, 2017 7:56 am
by Enrico Maria Giordano
Great!
EMG
Re: Transym OCR wrapper functions for Harbour
Posted: Wed Mar 29, 2017 5:00 pm
by Silvio.Falconi
run with demo Evaluate TOCR ?
How I must make to create a test sample ?
Re: Transym OCR wrapper functions for Harbour
Posted: Wed Apr 05, 2017 11:25 pm
by reinaldocrespo
Hi Silvio;
Transym has a demo that runs for 30 days (I think). You can download from the webpage. After creating the lib from the dll you only have to add the lib to the link script. That's all you have to do to make the project. It is all straight forward. To create a .lib from .dll use ImpLib.exe which is found on the borland bin directory.
http://www.transym.com
Good luck;
Reinaldo.
Re: Transym OCR wrapper functions for Harbour
Posted: Wed Feb 17, 2021 8:02 am
by damianodec
hi Reinaldo,
I'm interested to OCR.
are you using this product?
are there any others product?
any help?
thanks
Re: Transym OCR wrapper functions for Harbour
Posted: Wed Feb 17, 2021 2:36 pm
by reinaldocrespo
Hello Mr. Damianoodec;
Yes, I'm using Transym. I used the free google OCR libs (Leptonica and Tesseract) for years before moving to Transym. Tresseract works pretty well depending on the document being fed and it can be trained so it gets better with time if documents and fonts are the same type. The main problem with Tess is that you are pretty much on your own when it comes to Harbour and it seems Google has released that product after abandoning it.
In all honesty, I don't remember my code although I'm pretty sure I had to write some c. It has been working without my intervention for years and my mind has been somewhere else for a very long time. If you need help, I can try to find time to go back and review my code. But, in short, I do recommend Transym over Tess.
I hope that helps,
Reinaldo.
Re: Transym OCR wrapper functions for Harbour
Posted: Thu Feb 18, 2021 4:05 pm
by damianodec
hi Reinaldo,
I'd like understand OCR.
Are you using transym and FWH ?
in Transym site tehre are:
TOCR version 5.0, C++ samples code and c# samples code.
inside C++ samples there are a lot of folder and files.
have you any example?
thank you.
Re: Transym OCR wrapper functions for Harbour
Posted: Thu Feb 18, 2021 11:30 pm
by reinaldocrespo
Hello Damlanodec;
I just realized I had already shared sample code to test Transym on this same thread. It is the first post on this thread. At the very top.
Now that I look at it; the beauty of that code is that it is a self-contained minimized sample app-code that produces results. Notice on my sample code the wrapper functions you will need and a few other c functions that can help.
Please look at that code and let me know how it works for you. The lib works great.
Thank you.
Reinaldo.
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 9:57 am
by damianodec
hi Reinaldo,
when I try to compile you example by buildx.bat inside samples folder I get thi:
Code: Select all
Embarcadero C++ 7.00 for Win32 Copyright (c) 1993-2015 Embarcadero Technologies, Inc.
ocrreina.c:
Error E2046 ocrreina.prg 35: Bad file name format in include directive
Error E2046 ocrreina.prg 36: Bad file name format in include directive
Error E2046 ocrreina.prg 37: Bad file name format in include directive
Error E2303 ocrreina.prg 39: Type name expected
Error E2303 ocrreina.prg 40: Type name expected
Error E2147 ocrreina.prg 41: 'TOCRRESULTSEX' cannot start a parameter declaration
Error E2451 ocrreina.prg 52: Undefined symbol 'TOCRJOBINFO2' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2379 ocrreina.prg 52: Statement missing ; in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 53: Undefined symbol 'TOCRRESULTSEX' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 53: Undefined symbol 'Results' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2140 ocrreina.prg 54: Declaration is not allowed here in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2140 ocrreina.prg 55: Declaration is not allowed here in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2140 ocrreina.prg 56: Declaration is not allowed here in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2140 ocrreina.prg 57: Declaration is not allowed here in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2140 ocrreina.prg 58: Declaration is not allowed here in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 61: Undefined symbol 'TOCRCONFIG_DEFAULTJOB' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 61: Undefined symbol 'TOCRCONFIG_DLL_ERRORMODE' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 61: Undefined symbol 'TOCRERRORMODE_MSGBOX' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8065 ocrreina.prg 61: Call to function 'TOCRSetConfig' with no prototype in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 63: Undefined symbol 'JobInfo2' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2109 ocrreina.prg 63: Not an allowed type in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 65: Undefined symbol 'TOCRJOBTYPE_TIFFFILE' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8065 ocrreina.prg 72: Call to function 'TOCRInitialise' with no prototype in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2451 ocrreina.prg 74: Undefined symbol 'TOCR_OK' in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8065 ocrreina.prg 75: Call to function 'OCRWait' with no prototype in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8065 ocrreina.prg 76: Call to function 'GetResults' with no prototype in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8065 ocrreina.prg 83: Call to function 'TOCRShutdown' with no prototype in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8004 ocrreina.prg 86: 'InputFile' is assigned a value that is never used in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8004 ocrreina.prg 86: 'OptionsMask' is assigned a value that is never used in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Warning W8004 ocrreina.prg 86: 'Status' is assigned a value that is never used in function HB_FUN_OCRFROMFILEUSINGTRANSYM
Error E2303 ocrreina.prg 89: Type name expected
*** 23 errors in Compile ***
* Linking errors *
can y ou help me?
thank you
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 10:06 am
by Enrico Maria Giordano
Code: Select all
Error E2046 ocrreina.prg 35: Bad file name format in include directive
Can you show the line 35, please?
EMG
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 10:37 am
by damianodec
hi Enrico,
line 35, 36 and 37
Code: Select all
#include <TOCRdll.h>
#include <TOCRuser.h>
#include <TOCRerrs.h>
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 10:40 am
by Enrico Maria Giordano
Do you have TOCRdll.h in the same directory?
EMG
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 10:49 am
by damianodec
yes that files are inside FWH\include with other .ch fwh files (fivewin.ch and others)
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 10:57 am
by damianodec
hi,
I'm over my error (thank you Enrico), now I get this:
Code: Select all
Compiling...
xHarbour 1.2.3 Intl. (SimpLex) (Build 20161218)
Copyright 1999-2016, http://www.xharbour.org http://www.harbour-project.org/
Compiling 'ocrreina.prg' and generating preprocessed output to 'ocrreina.ppo'...
Generating C source output to 'ocrreina.c'...
Done.
Lines 29, Functions/Procedures 1, pCodes 42
Embarcadero C++ 7.00 for Win32 Copyright (c) 1993-2015 Embarcadero Technologies, Inc.
ocrreina.c:
Turbo Incremental Link 6.70 Copyright (c) 1997-2014 Embarcadero Technologies, Inc.
Error: 'C:\XHARBOUR1709\LIB\TOCRDLL.LIB' contains invalid OMF record, type 0x21 (possibly COFF)
* Linking errors *
Re: Transym OCR wrapper functions for Harbour
Posted: Fri Feb 19, 2021 10:58 am
by Enrico Maria Giordano
Can you send me all the required files to make a test here, please?
EMG