int rx;
} FindCharResults;
-enum { ct_Word, ct_Upper, ct_Lower, ct_Digit };
-static const char *context_names[]= {
- "Word", /* bit 0, value 001 */
- "Upper", /* bit 1, value 002 */
- "Lower", /* bit 2, value 004 */
- "Digit", /* bit 3, value 010 */
+#define FOR_EACH_CONTEXT(EACH) \
+ EACH(Word) \
+ EACH(Upper) \
+ EACH(Lower) \
+ EACH(Digit)
+
+#define FEC_ENUM(Context) ct_##Context,
+#define FEC_BIT(Context) ctf_##Context = 1 << ct_##Context,
+enum {
+ FOR_EACH_CONTEXT(FEC_ENUM)
+ FOR_EACH_CONTEXT(FEC_BIT)
};
+
+#define FEC_STRINGS(Context) #Context,
+static const char *context_names[]= { FOR_EACH_CONTEXT(FEC_STRINGS) };
+
struct OcrCellTypeInfo {
/* bitmaps of indices into context_names: */
unsigned initial, nextword, midword;
for (i=0, s=rd->results; i<rd->nresults; i++, s++) {
if (!strcmp(s->s," ")) continue;
fprintf(resolver," %d %d ",s->l,s->r);
- cu_pr_ctxmap(resolver, 1u << s->ctxi);
+ cu_pr_ctxmap(resolver, 1u << s->match);
+ fprintf(resolver," ");
+ cu_pr_ctxmap(resolver, s->ctxmap);
fprintf(resolver," ");
for (p=s->s; (c= *p); p++) {
if (c=='\\') fprintf(resolver,"\\%c",c);
readdb(rd);
}
-static void add_result(OcrReader *rd, const char *s, int l, int r, int ctxi) {
+static void add_result(OcrReader *rd, const char *s, int l, int r,
+ int match, unsigned ctxmap) {
if (rd->nresults >= rd->aresults) {
rd->aresults++; rd->aresults<<=1;
rd->results= mrealloc(rd->results, sizeof(*rd->results)*rd->aresults);
rd->results[rd->nresults].s= s;
rd->results[rd->nresults].l= l;
rd->results[rd->nresults].r= r;
- rd->results[rd->nresults].ctxi= ctxi;
+ rd->results[rd->nresults].match= match;
+ rd->results[rd->nresults].ctxmap= ctxmap;
rd->nresults++;
}
}
const struct OcrCellTypeInfo ocr_celltype_number= {
- 010,010,010,
+ ctf_Digit, ctf_Digit, ctf_Digit,
.space_spaces= 5,
.name= "number",
.findchar_select= 0
};
const struct OcrCellTypeInfo ocr_celltype_text= {
- .initial= 012, /* Digit|Upper */
- .nextword= 017, /* Digit|Upper|Lower|Word */
- .midword= 014, /* Digit|Lower */
+ .initial= ctf_Digit | ctf_Upper,
+ .nextword= ctf_Digit | ctf_Upper | ctf_Lower | ctf_Word,
+ .midword= ctf_Digit | ctf_Lower,
.space_spaces= 4,
.name= "text",
.findchar_select= findchar_select_text
/* something here, so we need to add the spaces */
if (nspaces >= ct->space_spaces)
- add_result(rd," ",x-nspaces,x+1,-1);
+ add_result(rd," ",x-nspaces,x+1,-1,0);
nspaces=0;
fca.x= x;
if (match) {
debugf(" || YES");
- add_result(rd, match->str, x, match_rx, match_ctxi);
+ add_result(rd, match->str, x, match_rx, match_ctxi, fca.ctxmap);
x= match_rx+1;
if (match->match) fca.ctxmap= ct->midword;
else debugf(" (empty)");
}
}
- add_result(rd, 0,-1,-1,-1);
+ add_result(rd, 0,-1,-1,-1,0);
debugf("OCR finished %d glyphs\n",rd->nresults);
debug_flush();
return rd->results;