diff --git a/rime.lua b/rime.lua index b6b2041..0412771 100644 --- a/rime.lua +++ b/rime.lua @@ -65,7 +65,7 @@ end ------------------------------------------------------------- -- 以词定字 -- https://github.com/BlindingDark/rime-lua-select-character --- 删除了默认按键,需要在 key_binder(default.custom.yaml)下设置 +-- 删除了默认按键,需要在 key_binder 下设置 local function utf8_sub(s, i, j) i = i or 1 j = j or -1 @@ -154,37 +154,32 @@ function long_word_filter(input, env) local count = config:get_int(env.name_space .. "/count") or 2 local idx = config:get_int(env.name_space .. "/idx") or 4 + local code = env.engine.context.input -- 当前编码 + env.target_codes = env.target_codes or Set({"xian", "tian", "tuan", "jie"}) + local l = {} local firstWordLength = 0 -- 记录第一个候选词的长度,提前的候选词至少要比第一个候选词长 - -- local s1 = 0 -- 记录筛选了多少个英语词条(只提升 count 个词的权重,并且对comment长度过长的候选进行过滤) - local s2 = 0 -- 记录筛选了多少个汉语词条(只提升 count 个词的权重) + local s = 0 -- 记录筛选了多少个词条(只提升 count 个词的权重) local i = 1 for cand in input:iter() do - leng = utf8.len(cand.text) - if (firstWordLength < 1 or i < idx) then - i = i + 1 - firstWordLength = leng + if not env.target_codes[code] then yield(cand) - -- 不知道这两行是干嘛用的,似乎注释掉也没有影响。 - -- elseif #table > 30 then - -- table.insert(l, cand) - -- 注释掉了英文的 - -- elseif ((leng > firstWordLength) and (s1 < 2)) and (string.find(cand.text, "^[%w%p%s]+$")) then - -- s1 = s1 + 1 - -- if (string.len(cand.text) / string.len(cand.comment) > 1.5) then - -- yield(cand) - -- end - -- 换了个正则,否则中英混输的也会被提升 - -- elseif ((leng > firstWordLength) and (s2 < count)) and (string.find(cand.text, "^[%w%p%s]+$")==nil) then - elseif ((leng > firstWordLength) and (s2 < count)) and (string.find(cand.text, "[%w%p%s]+") == nil) then - yield(cand) - s2 = s2 + 1 else - table.insert(l, cand) + local leng = utf8.len(cand.text) + if (firstWordLength < 1 or i < idx) then + i = i + 1 + firstWordLength = leng + yield(cand) + elseif ((leng > firstWordLength) and (s < count)) and (string.find(cand.text, "[%w%p%s]+") == nil) then + yield(cand) + s = s + 1 + else + table.insert(l, cand) + end end end - for i, cand in ipairs(l) do + for _, cand in ipairs(l) do yield(cand) end end @@ -237,24 +232,17 @@ end -- 把候选项应改为「ā á ǎ à …… van vain」,让单个字符的排在前面 function v_filter(input, env) local code = env.engine.context.input -- 当前编码 + env.v_spec_arr = env.v_spec_arr or Set( + {"0️⃣", "1️⃣", "2️⃣", "3️⃣", "4️⃣", "5️⃣", "6️⃣", "7️⃣", "8️⃣", "9️⃣", "Vs."}) -- 仅当当前输入以 v 开头,并且编码长度为 2,才进行处理 - if (string.len(code) == 2 and string.find(code, "v") == 1) then + if (string.len(code) == 2 and string.find(code, "^v")) then local l = {} for cand in input:iter() do -- 特殊情况处理 - if (cand.text == "Vs.") then + if (env.v_spec_arr[cand.text]) then yield(cand) - end - -- 特殊情况处理 - local arr = { "1️⃣", "2️⃣", "3️⃣", "4️⃣", "5️⃣", "6️⃣", "7️⃣", "8️⃣", "9️⃣" } - for _, v in ipairs(arr) do - if (v == cand.text) then - yield(cand) - break - end - end - -- 候选项为单个字符的,提到前面来。 - if (utf8.len(cand.text) == 1) then + -- 候选项为单个字符的,提到前面来。 + elseif (utf8.len(cand.text) == 1) then yield(cand) else table.insert(l, cand)