void test_levenshtein_distance(){
char *string_one, *string_two;
asprintf(&string_one, "impute/recodes");
asprintf(&string_two, "impute/recdes");
assert(levenshtein_distance(string_one, string_two) == 1);
asprintf(&string_one, "input/input table");
asprintf(&string_two, "niput/input table");
assert(levenshtein_distance(string_one, string_two) == 2);
asprintf(&string_one, "database");
asprintf(&string_two, "database");
assert(levenshtein_distance(string_one, string_two) == 0);
asprintf(&string_one, "mthod");
asprintf(&string_two, "method");
assert(levenshtein_distance(string_one, string_two) == 1);
free(string_one);
free(string_two);
printf("Leaving test_levenshtein_distance successfully!\n");
}
static constvalue *find_automaton(const char *name,int *last,char *closestmatch)
{
constvalue *ptr;
int dist,closestdist=INT_MAX;
assert(last!=NULL);
*last=0;
if (closestmatch!=NULL)
*closestmatch='\0';
ptr=sc_automaton_tab.next;
while (ptr!=NULL) {
if (strcmp(name,ptr->name)==0)
return ptr;
if (closestmatch!=NULL && strlen(ptr->name)>0) {
dist=levenshtein_distance(name,ptr->name);
if (dist<closestdist && dist<=MAX_EDIT_DIST) {
strcpy(closestmatch,ptr->name);
closestdist=dist;
} /* if */
} /* if */
if (ptr->index>*last)
*last=ptr->index;
ptr=ptr->next;
} /* while */
return NULL;
}
void RecipeActionsHandler::mergeSimilar()
{
QList<Q3ListViewItem> items = parentListView->selectedItems();
if ( items.count() > 1 )
KMessageBox::sorry( kapp->mainWidget(), i18nc("@info", "Please select only one category."), QString() );
else if ( items.count() == 1 && items.at(0)->rtti() == 1001 ) {
CategoryListItem * cat_it = ( CategoryListItem* ) items.at(0);
QString name = cat_it->categoryName();
const double max_allowed_distance = 0.60;
const int length = name.length();
ElementList categories;
database->loadCategories( &categories );
ElementList matches;
for ( ElementList::const_iterator it = categories.begin(); it != categories.end(); ++it ) {
#if 0
if ( levenshtein_distance(name.toLatin1(),(*it).name.toLatin1())/double(qMax(length,(*it).name.length())) >= max_allowed_distance ) {
#else
if ( compareStrings(name,(*it).name) >= max_allowed_distance ) {
#endif
kDebug()<<(*it).name<<" matches";
if ( cat_it->categoryId() != (*it).id )
matches.append(*it);
}
}
for ( ElementList::const_iterator it = categories.begin(); it != categories.end(); ++it ) {
database->mergeCategories(cat_it->categoryId(),(*it).id);
}
}
else //either nothing was selected or a recipe was selected
static int find_closestsymbol_table(const char *name,const symbol *root,int symboltype,symbol **closestsym)
{
int dist,closestdist=INT_MAX;
char symname[2*sNAMEMAX+16];
symbol *sym=root->next;
int ident,critdist;
assert(closestsym!=NULL);
*closestsym=NULL;
assert(name!=NULL);
critdist=strlen(name)/2; /* for short names, allow only a single edit */
if (critdist>MAX_EDIT_DIST)
critdist=MAX_EDIT_DIST;
while (sym!=NULL) {
funcdisplayname(symname,sym->name);
ident=sym->ident;
if (symboltype==iARRAY && ident==iREFARRAY)
ident=iARRAY; /* reference arrays match arrays */
else if (symboltype==iVARIABLE && (sym->ident==iCONSTEXPR || sym->ident==iREFERENCE || sym->ident==iARRAY || sym->ident==iREFARRAY))
ident=iVARIABLE; /* when requesting variables, constants are also ok */
if (symboltype==ident || (symboltype==iVARIABLE && ident==iFUNCTN)) {
dist=levenshtein_distance(name,symname);
if (dist<closestdist && dist<=critdist) {
*closestsym=sym;
closestdist=dist;
} /* if */
} /* if */
sym=sym->next;
} /* while */
return closestdist;
}
int main()
{
const char *a = "hgfdhfgdhgf";
const char *b = "hgfhgfhgfdh";
printf("%d\n", levenshtein_distance(a, b));
}
int benchmark()
{
int i, j;
volatile unsigned sum = 0;
for(i = 0; i < 5; ++i)
for(j = 0; j < 5; ++j)
sum += levenshtein_distance(strings[i], strings[j]);
return sum;
}
void calculate_results_relevance (alpm_list_t *targets)
{
for (const alpm_list_t *t = targets; t; t = alpm_list_next (t)) {
const char *target = t->data;
for (const alpm_list_t *r = results; r; r = alpm_list_next(r)) {
const char *result = results_name (r->data);
const size_t lev_dst = levenshtein_distance (target, result);
if (lev_dst < ((results_t *) r->data)->rel) {
((results_t *) r->data)->rel = lev_dst;
}
}
}
}
void recurse_levenshtein(int idx,int nwords)
{
int loop,loop2,dist;
char tmpword[MAX_WORD_LENGTH];
static int depth,found;
int locfound = 0;
if (depth > 3){
depth--;
return;
}
if (depth == 0)
found = 0;
depth++;
// printf("depth = %d\n",depth);
for (loop = 0; loop < nwords; loop++){
if (word[loop]==NULL)
continue;
dist = levenshtein_distance(word[idx],word[loop],1);
if (dist == 1){
if (depth == 1)
printf("%s\n",word[idx]);
printf("%s\n",word[loop]);
free(word[idx]);
word[idx] = NULL;
found++;
locfound = loop;
recurse_levenshtein(loop,nwords);
}
}
depth--;
if (locfound){
free(word[locfound]);
word[locfound] = NULL;
}
if (depth == 0){
if (found == 0){
printf("%s\n",word[idx]);
free(word[idx]);
word[idx] = NULL;
}
printf("------------------\n");
}
}
static PyObject* jellyfish_levenshtein_distance(PyObject *self, PyObject *args)
{
const char *s1, *s2;
int result;
if (!PyArg_ParseTuple(args, "ss", &s1, &s2)) {
return NULL;
}
result = levenshtein_distance(s1, s2);
if (result == -1) {
// levenshtein_distance only returns failure code (-1) on
// failed malloc
PyErr_NoMemory();
return NULL;
}
return Py_BuildValue("i", result);
}
int check_levenshtein_distances(int max_lev_distance){
int typo_counter=0;
int min_distance;
char *closest;
if (!apop_table_exists("keys")) return 0;
apop_data *userkeys = apop_query_to_text("select key from keys");
for (int i=0; i < *userkeys->textsize; i++){
min_distance = 100;
for (char **keyptr=ok_keys; strlen(*keyptr); keyptr++){
int ld = levenshtein_distance(*keyptr, *userkeys->text[i]);
if (ld < min_distance){
if(ld == 0) {min_distance=0; break;}
min_distance=ld;
closest = *keyptr;
}
}
Apop_stopif(min_distance > 0 && min_distance <= max_lev_distance, typo_counter++ , 0,
"You wrote %s for one of the keys in your spec file. Did you "
"mean to write %s?", *userkeys->text[i], closest);
}
return typo_counter;
}
SC_FUNC int error_suggest_list(int number,const char *name,constvalue *list)
{
assert(name!=NULL);
assert(list!=NULL);
if (sc_status==statWRITE) {
constvalue *closest=NULL;
if (strlen(name)>0) {
int dist,closestdist=INT_MAX;
while (list->next!=NULL) {
list=list->next;
dist=levenshtein_distance(list->name,name);
if (dist<closestdist && dist<=MAX_EDIT_DIST) {
closest=list;
closestdist=dist;
} /* if */
} /* while */
} /* if */
if (closest!=NULL && strcmp(name,closest->name)!=0)
error(makelong(number,1),name,closest->name);
else
error(number,name);
} /* if */
return 0;
}
int main (int argc, char **argv) {
if (argc < 3) return -1;
printf("%.1f%%\n", levenshtein_distance(argv[1],argv[2]) * 100);
return 0;
}
static VALUE lev_dist(VALUE self, VALUE str1, VALUE str2) {
return LONG2FIX(levenshtein_distance( str1, str2 ));
}
int main(int argc, char *argv[])
{
if (argc == 3)
printf("\n%s %s : %d\n",argv[1],argv[2],levenshtein_distance(argv[1],argv[2]));
}
edit_distance_t
levenshtein_distance (const char *s, const char *t)
{
return levenshtein_distance (s, strlen (s), t, strlen (t));
}
/* "/home/abhinav/projects/fb-puzzles/breathalyzer/levenshtein.pyx":4
* int levenshtein_distance(char *s,char *t)
*
* def levenshtein(char* s1, char* s2): # <<<<<<<<<<<<<<
* return levenshtein_distance(s1, s2)
*/
static PyObject *__pyx_pf_11levenshtein_levenshtein(PyObject *__pyx_self, PyObject *__pyx_args, PyObject *__pyx_kwds); /*proto*/
static PyObject *__pyx_pf_11levenshtein_levenshtein(PyObject *__pyx_self, PyObject *__pyx_args, PyObject *__pyx_kwds) {
char *__pyx_v_s1;
char *__pyx_v_s2;
PyObject *__pyx_r = NULL;
PyObject *__pyx_t_1 = NULL;
static PyObject **__pyx_pyargnames[] = {&__pyx_n_s__s1,&__pyx_n_s__s2,0};
__Pyx_RefNannySetupContext("levenshtein");
__pyx_self = __pyx_self;
if (unlikely(__pyx_kwds)) {
Py_ssize_t kw_args = PyDict_Size(__pyx_kwds);
PyObject* values[2] = {0,0};
switch (PyTuple_GET_SIZE(__pyx_args)) {
case 2: values[1] = PyTuple_GET_ITEM(__pyx_args, 1);
case 1: values[0] = PyTuple_GET_ITEM(__pyx_args, 0);
case 0: break;
default: goto __pyx_L5_argtuple_error;
}
switch (PyTuple_GET_SIZE(__pyx_args)) {
case 0:
values[0] = PyDict_GetItem(__pyx_kwds, __pyx_n_s__s1);
if (likely(values[0])) kw_args--;
else goto __pyx_L5_argtuple_error;
case 1:
values[1] = PyDict_GetItem(__pyx_kwds, __pyx_n_s__s2);
if (likely(values[1])) kw_args--;
else {
__Pyx_RaiseArgtupleInvalid("levenshtein", 1, 2, 2, 1); {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
}
}
if (unlikely(kw_args > 0)) {
if (unlikely(__Pyx_ParseOptionalKeywords(__pyx_kwds, __pyx_pyargnames, 0, values, PyTuple_GET_SIZE(__pyx_args), "levenshtein") < 0)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
}
__pyx_v_s1 = __Pyx_PyBytes_AsString(values[0]); if (unlikely((!__pyx_v_s1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
__pyx_v_s2 = __Pyx_PyBytes_AsString(values[1]); if (unlikely((!__pyx_v_s2) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
} else if (PyTuple_GET_SIZE(__pyx_args) != 2) {
goto __pyx_L5_argtuple_error;
} else {
__pyx_v_s1 = __Pyx_PyBytes_AsString(PyTuple_GET_ITEM(__pyx_args, 0)); if (unlikely((!__pyx_v_s1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
__pyx_v_s2 = __Pyx_PyBytes_AsString(PyTuple_GET_ITEM(__pyx_args, 1)); if (unlikely((!__pyx_v_s2) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
}
goto __pyx_L4_argument_unpacking_done;
__pyx_L5_argtuple_error:;
__Pyx_RaiseArgtupleInvalid("levenshtein", 1, 2, 2, PyTuple_GET_SIZE(__pyx_args)); {__pyx_filename = __pyx_f[0]; __pyx_lineno = 4; __pyx_clineno = __LINE__; goto __pyx_L3_error;}
__pyx_L3_error:;
__Pyx_AddTraceback("levenshtein.levenshtein");
return NULL;
__pyx_L4_argument_unpacking_done:;
/* "/home/abhinav/projects/fb-puzzles/breathalyzer/levenshtein.pyx":5
*
* def levenshtein(char* s1, char* s2):
* return levenshtein_distance(s1, s2) # <<<<<<<<<<<<<<
*/
__Pyx_XDECREF(__pyx_r);
__pyx_t_1 = PyInt_FromLong(levenshtein_distance(__pyx_v_s1, __pyx_v_s2)); if (unlikely(!__pyx_t_1)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 5; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
__Pyx_GOTREF(__pyx_t_1);
__pyx_r = __pyx_t_1;
__pyx_t_1 = 0;
goto __pyx_L0;
__pyx_r = Py_None; __Pyx_INCREF(Py_None);
goto __pyx_L0;
__pyx_L1_error:;
__Pyx_XDECREF(__pyx_t_1);
__Pyx_AddTraceback("levenshtein.levenshtein");
__pyx_r = NULL;
__pyx_L0:;
__Pyx_XGIVEREF(__pyx_r);
__Pyx_RefNannyFinishContext();
return __pyx_r;
}
