Commit Graph

262 Commits

Author SHA1 Message Date
zonble 9faed2153f Uses property wrappers to manage preferences. 2022-01-14 18:06:26 +08:00
Lukhnos Liu c698c61432 Merge pull request #220 from lukhnos/custom-phrase-reader
Use a more tolerant parser for user phrases
2022-01-13 23:43:35 -08:00
Lukhnos Liu d6cc5479f6 Use a more tolerant parser for user phrases
A generic key-value blob reader, KeyValueBlobReader, is implemented to
allow more flexibility in user-editable files. For example, this allows
comments in the file, as well as tolerating leading or trailing spaces,
tabs, or even Windows CR LF line endings.

Unit tests are supplied for KeyValueBlobReader although they are not
part of the Xcode project. A separate CMakeLists.txt is provided.

UserPhrasesLM is refactored to use KeyValueBlobReader. A small stylistic
change is appiled to reduce "using namespace" uses, but otherwise no
major style changes were applied to UserPhrasesLM.

Please note that McBopomofo's user phrase LM uses the value in a
key-value pair as the reading, and the key as the actual "value". We
don't plan to change that order so that we don't have to migrate data.

std::string_view is used to allow efficient reference to char buffers
and interop with std::string (and so no c_str() is needed). C++17 is now
enabled for the project to enable the use of std::string_view.

Copyright headers are added to McBopomofoLM and UserPhrasesLM.
2022-01-13 23:27:31 -08:00
zonble fcdd59dd6b Wraps OpenCCBridge into a SPM package. 2022-01-14 00:57:41 +08:00
zonble e01eb46c9f Wraps InputSourceHelper to a SPM package. 2022-01-14 00:43:21 +08:00
zonble d4772ffa99 Adds notifier UI to notify user Chinese conversion on/off. 2022-01-13 23:38:56 +08:00
zonble a7e38b5b2d Fine-tunes tooltip UI. 2022-01-13 22:00:29 +08:00
zonble 366453820d Adds a tiny tooltop for shift-left/right selections. 2022-01-13 21:47:52 +08:00
zonble 4c1781d970 Renames half-size to half-width. 2022-01-13 19:59:49 +08:00
zonble 9cd2306313 Adds emacs-style hotkeys. 2022-01-13 19:57:08 +08:00
zonble 232a944953 Implements half-size punctuations. 2022-01-13 17:07:22 +08:00
zonble d590d748f8 Adds UserPhrasesLM for user phrases.
Since there is no probability information for users' custom phrases,
they should be stored in a format differs from data.txt. Using the same
format and FastLM to parse user phrases just because of laziness but it
is not the right way.

The pull request adds a new language model class to parse user phrases.
It also update the input method controller to adopt the new user phrase
format.
2022-01-12 16:53:51 +08:00
zonble f1e56a7e01 Lets McBopomofoLM to accept NULL as the parameter in loadUserPhrases. 2022-01-12 13:17:41 +08:00
zonble 84fc2f068b Removes unused code and fixes a typo. 2022-01-12 13:16:10 +08:00
zonble abdf97f652 Adds McBopomofoLM as the facade of three language models.
- main language model
- user phrases
- user excluded phrases
2022-01-12 12:26:24 +08:00
zonble 56896625e3 Removes unused comments. 2022-01-12 01:17:39 +08:00
zonble cbd21cbe1d Updates localization. 2022-01-12 01:10:39 +08:00
zonble ea36061a41 Implements excluding punctuations. 2022-01-12 00:36:55 +08:00
zonble 9b485b799c Implements excluding phrases. 2022-01-12 00:16:55 +08:00
zonble 144d133463 Adds Language Model Manager.
The reference of the global language models were stored in the class
InputMethodController, however, the global models are global but not a
part of the input method controller, and the input method controller
only use one of the models (McBopomofo/Plain Bopomofo). I guess it
somehow violates SRP and there should be a better place for the global
models.
2022-01-11 17:12:58 +08:00
zonble f339948219 Fixes duplicated code and typos. 2022-01-11 13:46:29 +08:00
zonble df3914eeed Fixes a minor bug in the new Swift app delegate. 2022-01-11 02:35:31 +08:00
zonble a7b2edcf26 Converts AppDelegate to Swift. 2022-01-11 02:22:13 +08:00
zonble 61e2751702 Converts candidate UI to a Swift package. 2022-01-11 01:07:17 +08:00
zonble 867a828722 Fixes minor layout issues. 2022-01-11 00:47:48 +08:00
zonble f7e927d67d Starts to use Swift candidate UI.
There are bugs still.
2022-01-11 00:30:02 +08:00
zonble a97cc5ca6c Converts VerticalCandidateController to Swift. 2022-01-11 00:03:32 +08:00
zonble 5aafe64751 Starts to convert candidate UI to Swift. 2022-01-10 22:01:40 +08:00
zonble ba6889fa63 Converts OVInputSourceHelper to Swift. 2022-01-10 22:01:40 +08:00
zonble 84849bdb3d Converts the preference and non modal view controller to Swift. 2022-01-10 22:01:40 +08:00
zonble 75a0f68a9c Fixes the bug that I forgot to create the user phrases folder.
There was a legacy user override model which creates a folder and a
plist file. If a user uses McBopomofo for years, the folder would
exist. However, when the old override model was removed, I forgot
to create the folder for the new user phrase file.

The bug would let the users with new installation of McBopomofo unable
to add user phrases.
2022-01-10 21:59:18 +08:00
zonble 6bdd2aab44 Fixes a bug on building the unigrams. 2022-01-09 13:00:19 -08:00
zonble b4276f0488 Fixes a bug on building the vector for unigrams from both global language model and user phrases. 2022-01-09 13:00:19 -08:00
zonble 1e5bad20c2 Removes unused references. 2022-01-09 08:38:32 -08:00
zonble 5b72e48a4e Minor fine-tunes on the preference window. 2022-01-09 08:38:32 -08:00
zonble 0af238ef79 Cleans-up unused logs. 2022-01-09 08:38:32 -08:00
zonble 3763688275 Fixes a typo. 2022-01-09 08:38:32 -08:00
zonble a5247d958c Makes it able to reload user phrases. 2022-01-09 08:38:32 -08:00
zonble e909dc20b5 Uses user phrases in the block builder. 2022-01-09 08:38:32 -08:00
zonble 6f761ecbcd Implements adding phrase from shift and arrow keys. 2022-01-09 08:38:32 -08:00
zonble 358462dff1 [WIP] Starts to work on the user phrases. 2022-01-09 08:38:32 -08:00
ovadmin 789d2a5687 計算選字事件時,若遇到常用標點,將標點視為句尾
如此一來標點後的單字詞,在計算時,等同於句首第一詞。
2022-01-06 18:28:37 -08:00
ovadmin aeb774a8ed 小幅重構重複的程式碼 2022-01-06 18:28:37 -08:00
ovadmin 2e8e78971c 傳統注音不要記住用戶選字 2022-01-06 18:28:37 -08:00
ovadmin eef6f8c0ce 加大用戶選字詞模型的容量跟半衰期 2022-01-06 18:28:37 -08:00
ovadmin 3e0e859feb 將用戶選字記憶機制整合入 InputMethodController 2022-01-06 18:28:37 -08:00
ovadmin d672136843 實作簡單的用戶選字記憶模型
這個模型基本上只是根據游標前的兩個 unigram 記憶當前的用戶選字。當有超過
一個以上的用戶選字時,則要給每個選字評分。評分標準是選字頻率乘上一個透過
半衰期遞減的最近選字經歷時間。如此一來我們在「少用但最近選過」及「常用但
最近少選」之間取得一個平衡。半衰期透過經驗法則決定。

目前這個簡易模型並不存入磁碟,因此下一次重開機後就會洗掉重來。目前這樣選
擇純粹是因為模型有半衰期,因此長時間存放後還是會遺忘。

這個模型的好處是對既有詞庫提供詞的影響很小,對於連續單字詞的 override 有
還不錯的幫助。如此對於人名、地名、公司名等專有名詞,應該可以減少選字的頻
率。這個模型應用起來的缺點是,如果用戶修改的字詞原來是個雙字詞,例如先前
的兩個 unigram 分別是 A, BB ,而用戶想改的是 BB 的第二個字,使選完後的三
個字分別是 A, B', C,這個 C 往往是記不起來的,但如果一開始用戶逐字選取,
亦即在 BB 只出現 B 時就選取 B' 然後再打 C ,則 A, B', C 這個組合往往能被
正確記憶。實際發生原因在此不討論,但跟底層所用的組字網架的架構有關。確實
要改進的話得要從底層重新架構來下手,但至少目前這個模型給的建議偏保守,不
至干擾原有的預設選字。衡諸得失,這個模型提供一些邊際上的改善,應該還是值
得採用的。
2022-01-06 18:28:37 -08:00
ovadmin a17438b67a 修正一些選字機制 C++ 檔案 #include 不完整的問題 2022-01-06 18:28:37 -08:00
ovadmin 3760d24350 移除早期的候選歷史記憶機制
這個機制從未正式發布,設計本身也有很多缺陷,因此決定移除。
2022-01-06 18:28:29 -08:00
zonble 23100153cc Adds an option to clear entire input buffer by ESC key.
This fixes #146.
2022-01-02 22:09:23 +08:00