ccutil/ocrclass.h

Go to the documentation of this file.

#ifndef           OCRCLASS_H
#define           OCRCLASS_H

#include          <time.h>
#ifdef __MSW32__
#include          <windows.h>
#endif
#include          "host.h"

/*Maximum lengths of various strings*/
#define MAX_FONT_NAME   34       /* name of font */
#define MAX_OCR_NAME    32       /* name of engine */
#define MAX_OCR_VERSION   17     /* version code of engine */

/*Image parameters*/
#define MIN_IMAGE_SIZE    64     /* smallest image that will be passed */
#define IMAGE_ROUNDING    32     /* all sizes are multiple of this */

#if defined(__SLOW_TIMES__)
/*Maximum timeouts of various functions (in secs)*/
#define STARTUP_TIMEOUT   100    /* start of OCR engine */
#define SHUTDOWN_TIMEOUT  50     /* end of OCR engine */
#define SENDIM_TIMEOUT    50     /* send of image */
#define RELEASE_TIMEOUT   50     /* release of semaphore */
#define READIM_TIMEOUT    100    /* read of image */
#define READTEXT_TIMEOUT  50     /* read of text */
#define PROGRESS_TIMEOUT  30     /* progress every 3 seconds */
#define BADTIMES_TIMEOUT  7      /* max lack of progress */
#else
/*Maximum timeouts of various functions (in secs)*/
#define STARTUP_TIMEOUT   10     /* start of OCR engine */
#define SHUTDOWN_TIMEOUT  6      /* end of OCR engine */
#define SENDIM_TIMEOUT    5      /* send of image */
#define RELEASE_TIMEOUT   5      /* release of semaphore */
#define READIM_TIMEOUT    10     /* read of image */
#define READTEXT_TIMEOUT  5      /* read of text */
#define PROGRESS_TIMEOUT  3      /* progress every 3 seconds */
#define BADTIMES_TIMEOUT  7      /* max lack of progress */
#endif

/*language definitions are identical to RTF*/
#define LANGE_NONE      0x0400   /*no language */
#define LANGE_ALBANIAN    0x041c /*Albanian */
#define LANGE_BRITISH   0x0809   /*International English */
#define LANGE_BULGARIAN   0x0402 /*Bulgarian */
#define LANGE_CROATIAN    0x041a /*Croatian(latin alphabet) */
#define LANGE_CZECH     0x0405   /*Czech */
#define LANGE_DANISH    0x0406   /*Danish */
#define LANGE_DUTCH     0x0413   /*Dutch */
#define LANGE_FINNISH   0x040b   /*Finnish */
#define LANGE_FRENCH    0x040c   /*French */
#define LANGE_GERMAN    0x0407   /*German */
#define LANGE_GREEK     0x0408   /*Greek */
#define LANGE_HUNGARIAN   0x040e /*Hungarian */
#define LANGE_ITALIAN   0x0410   /*Italian */
#define LANGE_JAPANESE    0x0411 /*Japanese */
#define LANGE_KOREAN    0x0412   /*Korean */
#define LANGE_NORWEGIAN   0x0414 /*Bokmal */
#define LANGE_POLISH    0x0415   /*Polish */
#define LANGE_PORTUGESE   0x0416 /*Brazilian Portugese */
#define LANGE_ROMANIAN    0x0418 /*Romanian */
#define LANGE_RUSSIAN   0x0419   /*Russian */
#define LANGE_SCHINESE    0x0804 /*Simplified Chinese */
#define LANGE_SLOVAK    0x041b   /*Slovak */
#define LANGE_SPANISH   0x040a   /*Castilian */
#define LANGE_SWEDISH   0x041d   /*Swedish */
#define LANGE_TCHINESE    0x0404 /*Traditional Chinese */
#define LANGE_TURKISH   0x041f   /*Turkish */
#define LANGE_USENGLISH   0x0409 /*American */

/*font family definitions are identical to RTF*/
#define FFAM_NONE     0          /* unknown */
#define FFAM_ROMAN      1        /* serifed prop */
#define FFAM_SWISS      2        /* sans-serif prop */
#define FFAM_MODERN     3        /* fixed pitch */

/*character set definitions are identical to RTF*/
#define CHSET_ANSI      0        /* Ansi efigs */
#define CHSET_SHIFT_JIS   128    /* JIS X 0208-1990 */
#define CHSET_KOREAN    129      /* KS C 5601-1992 */
#define CHSET_SCHINESE    134    /* GB 2312-80 */
#define CHSET_BIG5      136      /* Big Five */
#define CHSET_CYRILLIC    204    /* Cyrillic */
#define CHSET_EEUROPE   238      /* Eastern Europe */

/*pitch set definitions are identical to RTF*/
#define PITCH_DEF     0          /* default */
#define PITCH_FIXED     1        /* fixed pitch */
#define PITCH_VAR     2          /* variable pitch */

/*Bitmasks for character enhancements.
OR these together for enhancement in ocr_append_char*/
#define EUC_BOLD      1          /* bold character */
#define EUC_ITALIC      2        /* italic char */
#define EUC_UNDERLINE   4        /* underlined char */
#define EUC_SUBSCRIPT   8        /* subscript char */
#define EUC_SUPERSCRIPT   16     /* superscript char */

enum OCR_CHAR_DIRECTION
{
  OCR_CDIR_RIGHT_LEFT,           /* right to left horizontal */
  OCR_CDIR_LEFT_RIGHT,           /* left to right horizontal */
  OCR_CDIR_TOP_BOTTOM,           /* top to bottom vertical */
  OCR_CDIR_BOTTOM_TOP            /* bottom to top vertical */
};

enum OCR_LINE_DIRECTION
{
  OCR_LDIR_DOWN_RIGHT,           /* horizontal lines go down */
  /*vertical lines go right */
  OCR_LDIR_UP_LEFT               /* horizontal lines go up */
};

enum OCR_NEWLINE_TYPE
{
  OCR_NL_NONE,                   // not a newline
  OCR_NL_NEWLINE,                // this is a newline but not new para
  OCR_NL_NEWPARA                 // this is a newline and a new para
};

/*error codes that can be returned from the API functions other than OKAY and HPERR*/
#define OCR_API_NO_MEM    (-2)   /*filled output buffer */
#define OCR_API_BAD_CHAR  (-3)   /*whitespace sent to ocr_append_char */
#define OCR_API_BAD_STATE (-4)   /*invalid call sequence */

enum OCR_ERR_CODE
{
  OCR_ERR_NONE,                  /* no error */
  OCR_ERR_CLEAN_EXIT,            /* no error */
  OCR_ERR_NO_MEM,                /* out of memory */
  OCR_ERR_FILE_READ,             /* failed to read data file */
  OCR_ERR_TMP_WRITE,             /* failed to write temp file */
  OCR_ERR_TMP_READ,              /* failed to read temp file */
  OCR_ERR_BAD_DLL,               /* missing or invalid dll subcomponent */
  OCR_ERR_BAD_EXE,               /* missing or invalid exe subcomponent */
  OCR_ERR_BAD_LOAD,              /* failed to load subcomponent */
  OCR_ERR_BAD_LANG,              /* unable to recognize requested language */
  OCR_ERR_BAD_STATE,             /* engine did call out of sequence */
  OCR_ERR_INTERNAL1,             /* internal error type 1 */
  OCR_ERR_INTERNAL2,             /* internal error type 1 */
  OCR_ERR_INTERNAL3,             /* internal error type 1 */
  OCR_ERR_INTERNAL4,             /* internal error type 1 */
  OCR_ERR_INTERNAL5,             /* internal error type 1 */
  OCR_ERR_INTERNAL6,             /* internal error type 1 */
  OCR_ERR_INTERNAL7,             /* internal error type 1 */
  OCR_ERR_INTERNAL8,             /* internal error type 1 */
  OCR_ERR_TIMEOUT                /* timed out in comms */
};                               /*for calls to ocr_error */

typedef struct                   /*font description */
{
  UINT16 language;               /* default language */
  UINT8 font_family;             /* serif/not, fixed/not */
  UINT8 char_set;                /* character set standard */
  UINT8 pitch;                   /* fixed or prop */
  INT8 name[MAX_FONT_NAME + 1];  /* plain ascii name */
} EFONT_DESC;                    /* font description */

typedef struct                   /*startup info */
{
  INT32 protocol;                /* interface version */
  UINT32 font_count;             /* number of fonts */
  UINT16 language;               /* default language */
  UINT16 name[MAX_OCR_NAME + 1]; /* name of engine */
                                 /* version of engine */
  UINT16 version[MAX_OCR_VERSION + 1];
  EFONT_DESC fonts[1];           /* array of fonts */
} EOCR_DESC;                     /*startup info */

typedef struct                   /*bitmap strip */
{
  INT16 x_size;                  /* width in pixels */
  INT16 y_size;                  /* of full image */
  INT16 strip_size;              /* of this strip */
  INT16 resolution;              /* pixels per inch */
  UINT8 data[8];                 /* image data */
} ESTRIP_DESC;                   /* bitmap strip */

typedef struct                   /*single character */
{
  UINT16 char_code;              /* character itself */
  INT16 left;                    /* of char (-1) */
  INT16 right;                   /* of char (-1) */
  INT16 top;                     /* of char (-1) */
  INT16 bottom;                  /* of char (-1) */
  INT16 font_index;              /* what font (0) */
  UINT8 confidence;              /* 0=perfect, 100=reject (0/100) */
  UINT8 point_size;              /* of char, 72=i inch, (10) */
  INT8 blanks;                   /* no of spaces before this char (1) */
  UINT8 formatting;              /* char formatting (0) */
} EANYCODE_CHAR;                 /*single character */

typedef bool (*CANCEL_FUNC)(void* cancel_this, int words);
typedef struct                   /*output header */
{
  INT16 count;                   /* chars in this buffer(0) */
  INT16 progress;                /* percent complete increasing (0-100) */
  INT8 more_to_come;             /* true if not last */
  INT8 ocr_alive;                /* ocr sets to 1, HP 0 */
  INT8 err_code;                 /* for errcode use */
  CANCEL_FUNC cancel;            /*returns true to cancel */
  void* cancel_this;             /*this or other data for cancel*/
  clock_t end_time;              /*time to stop if not 0*/
  EANYCODE_CHAR text[1];         /*character data */
} ETEXT_DESC;                    /*output header */

#ifdef __MSW32__

typedef struct                   /*shared mem info */
{
  HANDLE shm_hand;               /*handle to shm */
  HANDLE mutex;                  /*alive check */
  HANDLE ocr_sem;                /*ocr semaphore */
  HANDLE hp_sem;                 /*hp semaphore */
  void *shm_mem;                 /*shared memory */
  INT32 shm_size;                /*size of shm */
} ESHM_INFO;                     /*shared mem info */
#elif defined (__MAC__)

typedef struct                   /*shared mem info */
{
  Boolean mutex;                 /*alive check */
  Boolean ocr_sem;               /*ocr semaphore */
  Boolean hp_sem;                /*hp semaphore */
  void *shm_mem;                 /*shared memory */
  INT32 shm_size;                /*size of shm */
  INT16 language;

  // Process management information follows:
  ProcessSerialNumber IPEProcess;
  ProcessSerialNumber OCRProcess;
} ESHM_INFO;
#elif defined (__UNIX__)

typedef struct
{
  void *shm_mem;                 /* shared memory */
  INT32 shm_size;                /* size of shm */
} ESHM_INFO;
#endif
#endif

---