Rev 12248 | Go to most recent revision | Details | Compare with Previous | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
10 | reyssat | 1 | /* Copyright (C) 1998-2003 XIAO, Gang of Universite de Nice - Sophia Antipolis |
2 | * |
||
3 | * This program is free software; you can redistribute it and/or modify |
||
4 | * it under the terms of the GNU General Public License as published by |
||
5 | * the Free Software Foundation; either version 2 of the License, or |
||
6 | * (at your option) any later version. |
||
7 | * |
||
8 | * This program is distributed in the hope that it will be useful, |
||
9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
||
10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
||
11 | * GNU General Public License for more details. |
||
12 | * |
||
13 | * You should have received a copy of the GNU General Public License |
||
14 | * along with this program; if not, write to the Free Software |
||
15 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
||
16 | */ |
||
17 | |||
7676 | bpr | 18 | /* Versatile translation according to a dictionary */ |
10 | reyssat | 19 | |
20 | /*************** Customization: change values hereafter ****************/ |
||
8123 | bpr | 21 | #include "suffix.h" |
22 | #include "../Lib/libwims.h" |
||
23 | #include "translator_.h" |
||
7676 | bpr | 24 | /* limit of dictionary entries */ |
10 | reyssat | 25 | #define entrylim 32768 |
7676 | bpr | 26 | /* limit of dictionary length */ |
27 | #define diclim 1024*1024 |
||
10 | reyssat | 28 | |
29 | /***************** Nothing should need change hereafter *****************/ |
||
30 | |||
3718 | reyssat | 31 | |
10 | reyssat | 32 | char inpbuf[MAX_LINELEN+1], outbuf[2*MAX_LINELEN+2]; |
33 | char *dicbuf; |
||
8123 | bpr | 34 | struct entry entry[entrylim]; |
10 | reyssat | 35 | int entrycount; |
8123 | bpr | 36 | int entry_size=sizeof(entry); |
10 | reyssat | 37 | |
38 | int has_digits=0; |
||
39 | int unknown_type=unk_delete; |
||
40 | int nocase=0,leaveline=0; |
||
41 | char *unknown, unkbuf[1024]; |
||
42 | |||
43 | int compare(int i1, const char *s2) |
||
44 | { |
||
45 | int k; |
||
3247 | bpr | 46 | if(nocase) k=strncasecmp((char*)entry[i1].original,s2,entry[i1].olen); |
47 | else k=strncmp((char*)entry[i1].original,s2,entry[i1].olen); |
||
10 | reyssat | 48 | if(k==0 && (isalnum(*(s2+entry[i1].olen)) || (*(s2+entry[i1].olen)&128)!=0)) return -1; |
49 | else return k; |
||
50 | } |
||
51 | |||
7676 | bpr | 52 | /* searches a list. Returns index if found, -1 if nomatch. |
53 | * Uses binary search, list must be sorted. */ |
||
8100 | bpr | 54 | int search_list2(struct entry *list, int items, size_t item_size, const char *str) |
10 | reyssat | 55 | { |
56 | int i1,i2,j,k,t,t1; |
||
57 | unsigned char c; |
||
58 | |||
59 | if(items<=0) return -1; |
||
60 | j=0; c=str[0]; |
||
61 | k=list[0].original[0]-c; if(k==0) k=compare(0,str); |
||
11124 | georgesk | 62 | if(k==0) goto more; |
63 | if(k>0) return -1; |
||
10 | reyssat | 64 | j=items-1; k=list[j].original[0]-c; if(k==0) k=compare(j,str); |
65 | if(k==0) return j; |
||
66 | if(k>0) for(i1=0,i2=j;i2>i1+1;) { |
||
7676 | bpr | 67 | j=i1+(i2-i1)/2; |
68 | k=list[j].original[0]-c; if(k==0) k=compare(j,str); |
||
69 | if(k==0) goto more; |
||
70 | if(k>0) {i2=j; continue;} |
||
71 | if(k<0) {i1=j; continue;} |
||
10 | reyssat | 72 | } |
73 | if(k>0) {j--;k=compare(j,str);} |
||
74 | more: |
||
75 | if((t=list[j].earlier)<0) { |
||
7676 | bpr | 76 | if(k==0) return j; else return -1; |
10 | reyssat | 77 | } |
78 | if(compare(t,str)!=0) return -1; |
||
8100 | bpr | 79 | for(j=t1=t,k=0;j<items && list[j].earlier==t1 && (k=compare(j,str))<=0; j++) { |
10 | reyssat | 80 | if(k==0) t=j; |
8100 | bpr | 81 | } |
10 | reyssat | 82 | return t; |
83 | } |
||
84 | |||
7676 | bpr | 85 | /* change all spaces into ' ', and collapse multiple occurences */ |
8100 | bpr | 86 | void singlespace2(char *p) |
10 | reyssat | 87 | { |
13634 | bpr | 88 | char *pp, *p2; |
89 | for(pp=p;*pp;pp++) { |
||
90 | if(!isspace(*pp)) continue; |
||
91 | if(leaveline) { |
||
92 | if(*pp==13) ovlstrcpy(pp,pp+1); |
||
93 | if(*pp=='\n') { |
||
94 | pp++; |
||
95 | gopt: for(p2=pp; isspace(*p2) && *p2!='\n'; p2++); |
||
96 | if(p2>pp) ovlstrcpy(pp,p2); |
||
97 | pp--; |
||
98 | } |
||
99 | else { |
||
100 | pp++; if(!isspace(*pp) || *pp=='\n') continue; |
||
101 | goto gopt; |
||
102 | } |
||
10 | reyssat | 103 | } |
13634 | bpr | 104 | else { |
105 | if(*pp!=' ') *pp=' '; |
||
106 | if(!isspace(*(pp+1))) continue; |
||
107 | for(pp++,p2=pp;isspace(*p2);p2++); |
||
108 | ovlstrcpy(pp,p2); pp--; |
||
109 | } |
||
110 | } |
||
10 | reyssat | 111 | } |
112 | |||
7676 | bpr | 113 | /* Prepare dictionary */ |
10 | reyssat | 114 | void prepare_dic(char *fname) |
115 | { |
||
13634 | bpr | 116 | int i,l; |
117 | FILE *dicf; |
||
118 | char *p1, *p2, *pp; |
||
119 | long int flen; |
||
7676 | bpr | 120 | |
13634 | bpr | 121 | entrycount=0; |
122 | dicf=fopen(fname,"r"); if(dicf==NULL) return; |
||
123 | fseek(dicf,0,SEEK_END);flen=ftell(dicf); fseek(dicf,0,SEEK_SET); |
||
124 | if(flen>diclim) return; |
||
125 | dicbuf=xmalloc(flen+16);flen=fread(dicbuf,1,flen,dicf); |
||
126 | fclose(dicf); |
||
127 | if(flen>0 && flen<diclim) dicbuf[flen]=0; |
||
128 | else return; |
||
129 | for(i=0,p1=dicbuf;p1!=NULL && *p1!=0 && i<entrylim;p1=p2) { |
||
130 | p2=strchr(p1+1,'\n'); if(p2>p1) *p2++=0; |
||
131 | pp=strchr(p1,':'); if(pp==NULL) continue; |
||
132 | *pp++=0; |
||
133 | strip_trailing_spaces2(p1); strip_trailing_spaces2(pp); |
||
134 | singlespace2(p1); |
||
135 | p1=find_word_start(p1); pp=find_word_start(pp); |
||
136 | if(*p1==0) continue; |
||
137 | if(has_digits==0) { |
||
138 | char *p; |
||
139 | for(p=p1;*p!=0 && p<pp && !isdigit(*p);p++); |
||
140 | if(isdigit(*p)) has_digits=1; |
||
10 | reyssat | 141 | } |
13634 | bpr | 142 | entry[i].original=(unsigned char*)p1; |
143 | entry[i].replace=(unsigned char*)pp; |
||
144 | entry[i].olen=l=strlen(p1); entry[i].earlier=-1; |
||
145 | if(i>0) { |
||
146 | int l1,l2; |
||
147 | l1=entry[i-1].earlier; if(l1>=0) l2=entry[l1].olen; |
||
148 | else {l2=entry[i-1].olen;l1=i-1;} |
||
149 | if(l>l2 && isspace(p1[l2]) |
||
150 | && strncmp((char*)entry[l1].original,p1,l2)==0) |
||
151 | entry[i].earlier=entry[i-1].earlier=l1; |
||
152 | } |
||
153 | i++; |
||
154 | } |
||
155 | entrycount=i; |
||
10 | reyssat | 156 | } |
157 | |||
7676 | bpr | 158 | /* now make the translation. */ |
10 | reyssat | 159 | void translate(char *p) |
160 | { |
||
12248 | bpr | 161 | char *p1, *p2, *pp; |
162 | int t; |
||
10 | reyssat | 163 | |
12248 | bpr | 164 | if(entrycount<=0 && suffixcnt<=0) return; |
165 | snprintf(outbuf,sizeof(outbuf),"%s",p); |
||
166 | for(p1=find_word_start(outbuf); |
||
167 | p1!=NULL && p1-outbuf<MAX_LINELEN && *p1!=0; |
||
168 | p1=p2) { |
||
169 | p2=find_word_end(p1); |
||
170 | for(pp=p1;pp<p2 && |
||
171 | ((!has_digits && isalpha(*pp)) || |
||
172 | (has_digits && isalnum(*pp)) || (*pp&128)!=0 || |
||
173 | strchr("_",*pp)!=NULL);pp++); |
||
174 | p2=find_word_start(p2); |
||
175 | if(pp==p1 || |
||
176 | (has_digits==0 && isdigit(*pp)) || |
||
177 | (*pp!=0 && !isspace(*pp) && strchr(",.?!/;",*pp)==NULL)) continue; |
||
178 | t=search_list2(entry,entrycount,sizeof(entry[0]),p1); |
||
179 | if(t<0) { |
||
180 | switch(unknown_type) { |
||
181 | case unk_leave: break; |
||
182 | case unk_delete: { |
||
183 | ovlstrcpy(p1,find_word_start(pp)); p2=p1; |
||
184 | break; |
||
185 | } |
||
186 | case unk_replace: { |
||
187 | string_modify3(outbuf,p1,pp,unkbuf); |
||
188 | p2=find_word_start(p1+strlen(unkbuf)); |
||
189 | } |
||
7676 | bpr | 190 | } |
12248 | bpr | 191 | continue; |
192 | } |
||
193 | string_modify3(outbuf,p1,p1+strlen((char*)entry[t].original), |
||
7676 | bpr | 194 | (char*)entry[t].replace); |
12248 | bpr | 195 | p2=find_word_start(p1+strlen((char*)entry[t].replace)); |
196 | } |
||
197 | snprintf(p,MAX_LINELEN,"%s",outbuf); |
||
10 | reyssat | 198 | } |