Skip to content

Commit

Permalink
follow opencc conversion chain (#688)
Browse files Browse the repository at this point in the history
* follow opencc conversion chain

* when a dict doesn't contain a word, pass as-is

* de-duplication
  • Loading branch information
eagleoflqj committed Aug 12, 2023
1 parent 44ae945 commit 75e6b1a
Showing 1 changed file with 39 additions and 10 deletions.
49 changes: 39 additions & 10 deletions src/rime/gear/simplifier.cc
Original file line number Diff line number Diff line change
Expand Up @@ -58,19 +58,48 @@ class Opencc {
}

bool ConvertWord(const string& text, vector<string>* forms) {
if (dict_ == nullptr)
return false;
opencc::Optional<const opencc::DictEntry*> item = dict_->Match(text);
if (item.IsNull()) {
// Match not found
if (converter_ == nullptr) {
return false;
} else {
const opencc::DictEntry* entry = item.Get();
for (auto&& value : entry->Values()) {
forms->push_back(std::move(value));
}
const list<opencc::ConversionPtr> conversions =
converter_->GetConversionChain()->GetConversions();
vector<string> original_words{text};
bool matched = false;
for (auto conversion : conversions) {
opencc::DictPtr dict = conversion->GetDict();
if (dict == nullptr) {
return false;
}
return forms->size() > 0;
set<string> word_set;
vector<string> converted_words;
for (const auto& original_word : original_words) {
opencc::Optional<const opencc::DictEntry*> item =
dict->Match(original_word);
if (item.IsNull()) {
// Current dictionary doesn't convert the word. We need to keep it for
// other dicts in the chain. e.g. s2t.json expands 里 to 里 and 裏,
// then t2tw.json passes 里 as-is and converts 裏 to 裡.
if (word_set.insert(original_word).second) {
converted_words.push_back(original_word);
}
continue;
}
matched = true;
const opencc::DictEntry* entry = item.Get();
for (const auto& converted_word : entry->Values()) {
if (word_set.insert(converted_word).second) {
converted_words.push_back(converted_word);
}
}
}
original_words.swap(converted_words);
}
if (!matched) {
// No dictionary contains the word
return false;
}
*forms = std::move(original_words);
return forms->size() > 0;
}

bool RandomConvertText(const string& text, string* simplified) {
Expand Down

0 comments on commit 75e6b1a

Please sign in to comment.