Upper vs lower case - important note regarding `l' and `I'
----------------------------------------------------------
-We maintain separate dictionaries for upper and lower case. At the
+We maintain separate dictionaries for upper case (Upper), lower case
+(Lower), and (initial portions of) mid-phrase words (Word). At the
beginning of each cell in the table, we expect uppercase; in the
middle of a word we expect lowercase; and, unfortunately, after an
inter-word gap, we are not sure.
So any time we see a word starting with `l' or `I', the program has to
ask about it.
+After an interword gap, we first search for a Word entry in the
+dictionary. If there is a match we use it. Otherwise we search both
+the uppercase and lowercase dictionaries; if one matches and the other
+doesn't, or one matches a wider character than the other, we use it.
+If that fails to resolve the ambiguity we must ask.
+
*Do not* make an entry in the character set dictionary mapping `vertical
stick' to `l' or `I'. Instead, select enough of the whole word in
question that no word would start with the other letter, and enter the
-whole word or part of it as a new glyph.
+whole word or part of it as a new glyph as a new Word.
For example, in the supplied dictionary there is already a glyph for
`Iron'; this is OK because there are no words which start `lron'.
Instead, make a new glyph for the last letter of the previous word
plus the (unusually narrow) inter-word space, and end that entry with
-\x20 (yes, type \ x 20).
+a literal space ` '.
For example, you might find that `y<space>G' is treated as
`y<??lowercase>' and the G doesn't get matched. Select the `y<space>'
-region of the bitmap and type `y\x20' into the string box.
-Sorry for this rather poor UI!
+region of the bitmap and type `y ' into the string box.
Overlapping characters - ligatures
all if the dictionary contains errors, you shouldn't rely on this.
If you think you have made mistakes answering OCR queries (for
-example, the recognised data is wrong), you should download a fresh
-copy of charset-15.txt from
- http://www.chiark.greenend.org.uk/~ijackson/ypp-sc-tools/master/pctb/charset-15.txt
+example, the recognised data is wrong), you should delete the file
+#local-char*#.txt, which contains your local updates. It will then
+only use the centrally provided (and vetted) master file (which is
+automatically updated when you run the PCTB client, by default).
Enabling interactive character set update
DatabaseLink *links;
} DatabaseNode;
+typedef struct {
+ OcrReader *rd;
+ OcrCellType ct;
+ int w;
+ Pixcol *cols;
+ int x;
+ unsigned ctxmap;
+} FindCharArgs;
+typedef struct {
+ DatabaseNode *match;
+ int rx;
+} FindCharResults;
+
+enum { ct_Lower, ct_Upper, ct_Word, ct_Digit };
static const char *context_names[]= {
- "Lower",
- "Upper",
- "Digit"
+ "Lower", /* bit 0, value 001 */
+ "Upper", /* bit 1, value 002 */
+ "Word", /* bit 2, value 004 */
+ "Digit", /* bit 3, value 010 */
};
struct OcrCellTypeInfo {
/* bitmaps of indices into context_names: */
unsigned initial, nextword, midword;
int space_spaces;
const char *name;
+ int (*findchar_select)(const FindCharArgs *fca,
+ const FindCharResults results[]);
};
-const struct OcrCellTypeInfo ocr_celltype_number= {
- 4,4,4,
- .space_spaces= 5,
- .name= "number"
-};
-const struct OcrCellTypeInfo ocr_celltype_text= {
- .initial=2, /* Uppercase */
- .nextword=3, /* Either */
- .midword=1, /* Lower only */
- .space_spaces= 4,
- .name= "text"
-};
-
#define NCONTEXTS (sizeof(context_names)/sizeof(context_names[0]))
}
+static DatabaseNode *findchar_1ctx(const FindCharArgs *fca,
+ DatabaseNode *start, int *matchx_r) {
+ DatabaseNode *current= start;
+ DatabaseNode *bestmatch= 0;
+ int i;
+ int x= fca->x;
+
+ for (;;) {
+ debug_flush();
+ debugf(" | x=%d",x);
+ if (x > fca->w) break;
+ Pixcol cv= fca->cols[x];
+ debugf(" cv=%"PSPIXCOL(PRIx),cv);
+ for (i=0; i<current->nlinks; i++)
+ if (current->links[i].col == cv)
+ goto found;
+ /* not found */
+ debugf(" ?");
+ break;
+
+ found:
+ current= current->links[i].then;
+ if (current->match) {
+ debugf(" \"%s\"%s",current->str,current->endsword?"_":"");
+ bestmatch= current;
+ *matchx_r= x;
+ } else {
+ debugf(" ...");
+ }
+
+ x++;
+ }
+ return bestmatch;
+}
+
+static DatabaseNode *findchar(const FindCharArgs *fca, int *match_rx) {
+ FindCharResults results[NCONTEXTS];
+ int ctxi, match=-1, nmatches=0;
+
+ debugf("OCR lx=%d ct_state=%x ", fca->x, fca->ctxmap);
+ for (ctxi=0; ctxi<NCONTEXTS; ctxi++) {
+ results[ctxi].match= 0;
+ if (!(fca->ctxmap & (1u << ctxi))) continue;
+ debugf(" || %s",context_names[ctxi]);
+
+ results[ctxi].match= findchar_1ctx(fca, &fca->rd->contexts[ctxi],
+ &results[ctxi].rx);
+ if (!results[ctxi].match) continue;
+
+ match= ctxi;
+ nmatches++;
+ }
+ if (nmatches==1) {
+ debugf(" unique");
+ } else {
+ debugf(" ambiguous");
+ match= !fca->ct->findchar_select ? -1 :
+ fca->ct->findchar_select(fca,results);
+ debugf(" resolved %s", match<0 ? "<none>" : context_names[match]);
+ }
+ if (match<0)
+ return 0;
+
+ *match_rx= results[ctxi].rx;
+ return results[ctxi].match;
+}
+
+static int findchar_select_text(const FindCharArgs *fca,
+ const FindCharResults results[]) {
+ if (fca->ctxmap != 017) return -1;
+
+ dbassert(! results[ct_Digit].match );
+ if (results[ct_Word].match) return ct_Word;
+ if (results[ct_Lower].rx > results[ct_Upper].rx) return ct_Lower;
+ if (results[ct_Upper].rx > results[ct_Lower].rx) return ct_Upper;
+ return -1;
+}
+
+const struct OcrCellTypeInfo ocr_celltype_number= {
+ 010,010,010,
+ .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 */
+ .space_spaces= 4,
+ .name= "text",
+ .findchar_select= findchar_select_text
+};
+
+
const char *ocr_celltype_name(OcrCellType ct) { return ct->name; }
OcrResultGlyph *ocr(OcrReader *rd, OcrCellType ct, int w, Pixcol cols[]) {
int nspaces;
- unsigned ctxmap;
- int ctxi, i, x;
+ int x;
+
+ FindCharArgs fca;
+ fca.rd= rd;
+ fca.ct= ct;
+ fca.w= w;
+ fca.cols= cols;
+ fca.x= -1;
restart:
nspaces=- w;
- ctxmap= ct->initial;
+ 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]);
x++;
if (nspaces == ct->space_spaces) {
debugf("OCR x=%x nspaces=%d space\n",x,nspaces);
- ctxmap= ct->nextword;
+ fca.ctxmap= ct->nextword;
}
continue;
}
add_result(rd," ",x-nspaces,x+1,0);
nspaces=0;
- /* find character */
- int lx=x;
+ fca.x= x;
- DatabaseNode *uniquematch= 0;
- int uniquematch_rx=-1;
+ int match_rx=-1;
+ DatabaseNode *match= findchar(&fca, &match_rx);
- debugf("OCR lx=%d ctxmap=%x ",lx,ctxmap);
-
- for (ctxi=0; ctxi<NCONTEXTS; ctxi++) {
- DatabaseNode *current= &rd->contexts[ctxi];;
- DatabaseNode *bestmatch= 0;
- int bestmatch_rx=-1;
-
- x= lx;
- if (!(ctxmap & (1u << ctxi))) continue;
- debugf(" || %s",context_names[ctxi]);
-
- for (;;) {
- debug_flush();
- debugf(" | x=%d",x);
- if (x>w) break;
- Pixcol cv= cols[x];
- debugf(" cv=%"PSPIXCOL(PRIx),cv);
- for (i=0; i<current->nlinks; i++)
- if (current->links[i].col == cv)
- goto found;
- /* not found */
- debugf(" ?");
- break;
-
- found:
- current= current->links[i].then;
- if (current->match) {
- debugf(" \"%s\"%s",current->str,current->endsword?"_":"");
- bestmatch= current;
- bestmatch_rx= x;
- } else {
- debugf(" ...");
- }
-
- x++;
- }
-
- if (bestmatch) {
- if (uniquematch && strcmp(bestmatch->str, uniquematch->str)) {
- debugf( " ambiguous");
- uniquematch= 0;
- break;
- }
- uniquematch= bestmatch;
- uniquematch_rx= bestmatch_rx;
- }
- }
-
- if (uniquematch) {
+ if (match) {
debugf(" || YES");
- add_result(rd, uniquematch->str, lx, uniquematch_rx, ctxmap);
- x= uniquematch_rx+1;
- if (uniquematch->match) ctxmap= ct->midword;
+ add_result(rd, match->str, x, match_rx, fca.ctxmap);
+ x= match_rx+1;
+ if (match->match) fca.ctxmap= ct->midword;
else debugf(" (empty)");
- if (uniquematch->endsword) {
+ if (match->endsword) {
nspaces= ct->space_spaces;
debugf("_");
- ctxmap= ct->nextword;
+ fca.ctxmap= ct->nextword;
}
debugf("\n");
} else {
int rx;
debugf(" || UNKNOWN");
- for (rx=lx; rx<w && cols[rx]; rx++);
- debugf(" x=%d ctxmap=%x %d..%d\n",x, ctxmap, lx,rx);
+ for (rx=x; rx<w && cols[rx]; rx++);
+ debugf(" x=%d ctxmap=%x %d..%d\n",x, fca.ctxmap, x,rx);
debug_flush();
- callout_unknown(rd, w,cols, lx,rx-1, ctxmap);
+ callout_unknown(rd, w,cols, x,rx-1, fca.ctxmap);
goto restart;
}
+
}
add_result(rd, 0,-1,-1,0);
debugf("OCR finished %d glyphs\n",rd->nresults);