chiark
/
gitweb
/
~yarrgweb
/
ypp-sc-tools.db-test.git
/ blobdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
raw
|
inline
| side by side
antialiasing text conversion: seems to handle it in the C code so far ....
[ypp-sc-tools.db-test.git]
/
pctb
/
ocr.c
diff --git
a/pctb/ocr.c
b/pctb/ocr.c
index 6482b81430292af3a7c4da942088ce0e0ce00ace..c747b205cfc8f7eaca3c31bd40a130181341e8de 100644
(file)
--- a/
pctb/ocr.c
+++ b/
pctb/ocr.c
@@
-44,7
+44,7
@@
typedef struct {
OcrReader *rd;
OcrCellType ct;
int w;
OcrReader *rd;
OcrCellType ct;
int w;
- Pixcol *cols;
+
const
Pixcol *cols;
int x;
unsigned ctxmap;
} FindCharArgs;
int x;
unsigned ctxmap;
} FindCharArgs;
@@
-127,7
+127,7
@@
static void readdb1(OcrReader *rd, const char *which, int local) {
goto x;
FGETSLINE;
goto x;
FGETSLINE;
- dbassert(!strcmp(lbuf,"# ypp-sc-tools pctb font v
1
"));
+ dbassert(!strcmp(lbuf,"# ypp-sc-tools pctb font v
2
"));
dbassert( dbfile_scanf("%d", &h) == 1);
dbassert(h==rd->h);
dbassert( dbfile_scanf("%d", &h) == 1);
dbassert(h==rd->h);
@@
-161,12
+161,20
@@
static void readdb1(OcrReader *rd, const char *which, int local) {
for (;;) {
FGETSLINE;
if (!lbuf[0]) { dbassert(current != &rd->contexts[ctxi]); break; }
for (;;) {
FGETSLINE;
if (!lbuf[0]) { dbassert(current != &rd->contexts[ctxi]); break; }
- char *ep;
- cv= strtoul(lbuf,&ep,16); dbassert(!*ep);
- dbassert(!(cv & ~((1UL << rd->h)-1)));
+ dbassert( strlen(lbuf) == rd->h );
+ FILLZERO(cv);
+ int y;
+ for (y=0; y<h; y++) {
+ char pb[2];
+ pb[0]= lbuf[y];
+ pb[1]= 0;
+ char *ep;
+ unsigned pv= strtoul(pb,&ep,16); dbassert(!*ep);
+ pixcol_p_add(&cv, y, pv);
+ }
for (j=0; j<current->nlinks; j++)
for (j=0; j<current->nlinks; j++)
- if (
current->links[j].col == cv
) {
+ if (
!pixcol_cmp(&cv, ¤t->links[j].col)
) {
current= current->links[j].then;
goto found_link;
}
current= current->links[j].then;
goto found_link;
}
@@
-225,12
+233,11
@@
static void cu_pr_ctxmap(FILE *resolver, unsigned ctxmap) {
fprintf(resolver,"}");
}
fprintf(resolver,"}");
}
-static void callout_unknown(OcrReader *rd, int w, Pixcol cols[],
+static void callout_unknown(OcrReader *rd, int w,
const
Pixcol cols[],
int unk_l, int unk_r, unsigned unk_ctxmap) {
int c,i, x,y;
const OcrResultGlyph *s;
const char *p;
int unk_l, int unk_r, unsigned unk_ctxmap) {
int c,i, x,y;
const OcrResultGlyph *s;
const char *p;
- Pixcol pv;
FILE *resolver= resolve_start();
if (!resolver || !(o_flags & ff_editcharset))
FILE *resolver= resolve_start();
if (!resolver || !(o_flags & ff_editcharset))
@@
-258,20
+265,12
@@
static void callout_unknown(OcrReader *rd, int w, Pixcol cols[],
fputc('\n',resolver);
fprintf(resolver,
fputc('\n',resolver);
fprintf(resolver,
- "/* XPM */\n"
- "static char *t[] = {\n"
- "/* columns rows colors chars-per-pixel */\n"
- "\"%d %d 2 1\",\n"
- "\" c black\",\n"
- "\"o c white\",\n",
- w,rd->h);
- for (y=0, pv=1; y<rd->h; y++, pv<<=1) {
- fputc('"',resolver);
+ "P2\n%d %d\n%d\n", w, rd->h, AAMAXVAL);
+ for (y=0; y<rd->h; y++) {
for (x=0; x<w; x++)
for (x=0; x<w; x++)
- fp
utc(cols[x] & pv ? 'o' : ' ', resolver
);
- fput
s("\",\n"
,resolver);
+ fp
rintf(resolver, " %x", pixcol_p_get(&cols[x], y)
);
+ fput
c('\n'
,resolver);
}
}
- fputs("};\n",resolver);
resolve_finish();
readdb(rd);
resolve_finish();
readdb(rd);
@@
-304,9
+303,9
@@
static DatabaseNode *findchar_1ctx(const FindCharArgs *fca,
debugf(" | x=%d",x);
if (x > fca->w) break;
Pixcol cv= fca->cols[x];
debugf(" | x=%d",x);
if (x > fca->w) break;
Pixcol cv= fca->cols[x];
- debugf(" cv=
%"PSPIXCOL(PRIx),cv
);
+ debugf(" cv=
"PIXCOL_PRFMT, PIXCOL_PRVAL(cv)
);
for (i=0; i<current->nlinks; i++)
for (i=0; i<current->nlinks; i++)
- if (
current->links[i].col == cv
)
+ if (
!pixcol_cmp(&cv, ¤t->links[i].col)
)
goto found;
/* not found */
debugf(" ?");
goto found;
/* not found */
debugf(" ?");
@@
-408,7
+407,8
@@
const struct OcrCellTypeInfo ocr_celltype_text= {
const char *ocr_celltype_name(OcrCellType ct) { return ct->name; }
const char *ocr_celltype_name(OcrCellType ct) { return ct->name; }
-OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w, Pixcol cols[]) {
+OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w,
+ const Pixcol cols[]) {
int nspaces;
int x;
int nspaces;
int x;
@@
-425,7
+425,7
@@
OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w, Pixcol cols[]) {
fca.ctxmap= ct->initial;
rd->nresults=0;
debugf("OCR h=%d w=%d",rd->h,w);
fca.ctxmap= ct->initial;
rd->nresults=0;
debugf("OCR h=%d w=%d",rd->h,w);
- for (x=0; x<w; x++) debugf("
%"PSPIXCOL(PRIx),cols[x]
);
+ for (x=0; x<w; x++) debugf("
"PIXCOL_PRFMT, PIXCOL_PRVAL(cols[x])
);
debugf("\n");
debug_flush();
debugf("\n");
debug_flush();
@@
-436,7
+436,7
@@
OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w, Pixcol cols[]) {
if (x>=w)
break;
if (x>=w)
break;
- if (!
cols[x]
) {
+ if (!
pixcol_nonzero(&cols[x])
) {
nspaces++;
x++;
if (nspaces == ct->space_spaces) {
nspaces++;
x++;
if (nspaces == ct->space_spaces) {
@@
-472,7
+472,7
@@
OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w, Pixcol cols[]) {
} else {
int rx;
debugf(" || UNKNOWN");
} else {
int rx;
debugf(" || UNKNOWN");
- for (rx=x; rx<w &&
cols[rx]
; rx++);
+ for (rx=x; rx<w &&
pixcol_nonzero(&cols[rx])
; rx++);
debugf(" x=%d ctxmap=%x %d..%d\n",x, fca.ctxmap, x,rx);
debug_flush();
callout_unknown(rd, w,cols, x,rx-1, fca.ctxmap);
debugf(" x=%d ctxmap=%x %d..%d\n",x, fca.ctxmap, x,rx);
debug_flush();
callout_unknown(rd, w,cols, x,rx-1, fca.ctxmap);
@@
-489,6
+489,8
@@
OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w, Pixcol cols[]) {
OcrReader *ocr_init(int h) {
OcrReader *rd;
OcrReader *ocr_init(int h) {
OcrReader *rd;
+ assert(h <= OCR_MAX_H);
+
if (o_flags & ff_dict_fetch) {
char *fetchfile= masprintf("char%d",h);
progress("Updating %s...",fetchfile);
if (o_flags & ff_dict_fetch) {
char *fetchfile= masprintf("char%d",h);
progress("Updating %s...",fetchfile);
@@
-529,7
+531,7
@@
static int show_char_compar(const void *av, const void *bv) {
void ocr_showcharsets(void) {
DIR *d;
struct dirent *de;
void ocr_showcharsets(void) {
DIR *d;
struct dirent *de;
- char found[
32
];
+ char found[
OCR_MAX_H
];
pcre *fnpat;
int matchvec[10];
char hbuf[10];
pcre *fnpat;
int matchvec[10];
char hbuf[10];