func (this *MITIE) Process(body string) *list.List {
tokens := C.mitie_tokenize(C.CString(body))
defer C.mitie_free(unsafe.Pointer(tokens))
dets := C.mitie_extract_entities(this.ner, tokens)
defer C.mitie_free(unsafe.Pointer(dets))
num_dets := C.mitie_ner_get_num_detections(dets)
duplicates := set.New()
entites := list.New()
for i := 0; i < int(num_dets); i++ {
centity := C.get_entity(tokens, dets, C.ulong(i))
model := C.GoString(centity.model)
score := float64(centity.score)
value := C.GoString(centity.value)
key := fmt.Sprintf("%s:%s", value, model)
if duplicates.Has(key) {
continue
}
duplicates.Add(key)
if score > 0.5 {
entity := models.NewEntity(model, score, value)
entites.PushBack(entity)
}
}
return entites
}
// Extract runs the extractor and returns a slice of Entities found in the
// given tokens.
func (ext *Extractor) Extract(tokens []string) ([]Entity, error) {
ctokens := C.ner_arr_make(C.int(len(tokens)) + 1) // NULL termination
defer C.ner_arr_free(ctokens, C.int(len(tokens))+1)
for i, t := range tokens {
cs := C.CString(t) // released by ner_arr_free
C.ner_arr_set(ctokens, cs, C.int(i))
}
dets := C.mitie_extract_entities(ext.ner, ctokens)
defer C.mitie_free(unsafe.Pointer(dets))
if dets == nil {
return nil, ErrMemory
}
n := int(C.mitie_ner_get_num_detections(dets))
entities := make([]Entity, n, n)
for i := 0; i < n; i++ {
pos := int(C.mitie_ner_get_detection_position(dets, C.ulong(i)))
len := int(C.mitie_ner_get_detection_length(dets, C.ulong(i)))
entities[i] = Entity{
Tag: int(C.mitie_ner_get_detection_tag(dets, C.ulong(i))),
Score: float64(C.mitie_ner_get_detection_score(dets, C.ulong(i))),
Name: strings.Join(tokens[pos:pos+len], " "),
Range: Range{pos, pos + len},
}
}
return entities, nil
}
// Tokenize returns a slice that contains a tokenized copy of the input text.
func Tokenize(text string) []string {
cs := C.CString(text)
defer C.free(unsafe.Pointer(cs))
ctokens := C.mitie_tokenize(cs)
defer C.mitie_free(unsafe.Pointer(ctokens))
i := 0
// a hack since mitie arrays are NULL terminated.
p := (*[1 << 30]*C.char)(unsafe.Pointer(ctokens))
tokens := make([]string, 0, 20)
for p[i] != nil {
tokens = append(tokens, C.GoString(p[i]))
i++
}
return tokens
}
// Free frees the underlying used C memory.
func (ext *Extractor) Free() {
C.mitie_free(unsafe.Pointer(ext.ner))
}
func (this *MITIE) Release() {
C.mitie_free(unsafe.Pointer(this.ner))
}
