Support narrowing with Chinese pinyin via `pattern-transformer` #2601

zbelial · 2023-05-21T12:07:54Z

zbelial
May 21, 2023

Hi, I'd like to use Chinese pinyin to narrow candidates, so if I enter wenjian, helm will not only match files (or buffers etc) like wenjian.org, but also files like 文件.org。

The following is how I archive that:

I add a new variable helm-pattern-transformer-alist.

(defvar helm-pattern-transformer-alist
  '()
  "An alist of regex building functions for each source.

Each key is the source name:

Each value is a function that should take a string and return a
valid regex.")

I change function helm-process-pattern-transformer to use helm-pattern-transformer-alist in the following way

(defun helm-process-pattern-transformer (pattern source)
  "Execute pattern-transformer attribute function(s) on PATTERN in SOURCE."
  (let* ((name (assoc-default 'name source))
         (transformer (assoc-default name helm-pattern-transformer-alist)))
    (helm-aif (or transformer
                  (assoc-default 'pattern-transformer source))
        (helm-apply-functions-from-source source it pattern)
      pattern)))

If you want a particular source to support narrowing with Chinese pinyin, do this

(add-to-list 'helm-pattern-transformer-alist '("Buffers" . helm-pyim-to-utf8))

My questions are:

Is using pattern-transformer in this way the correct way?
If the answer to the first question is no, how should I do to archive my goal?
I have also tried to use candidate-transformer to transform Chinese character to Chinese pinyin, but haven't make it work as expected so far.

Thanks and welcome any feedback.

Answered by thierryvolpiatto

May 22, 2023

太阳 ***@***.***> writes:

Thanks very much. I tried your method with some sources, but it didn't work in some of them. Following is what I did. It worked in helm-source-buffers and helm-moccur-class, but not for the other three. (cl-defmethod helm-setup-user-source ((source helm-source-basic-bookmarks)) (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

For this one it can't work because the source is already defined in helm-bookmarks, use this instead: (helm-set-attr 'pattern-transformer #'helm-pyim-to-utf8 helm-source-bookmarks)

(cl-defmethod helm-setup-user-source ((source helm-recentf-source)) (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

Same fo…

View full answer

thierryvolpiatto · 2023-05-21T17:49:33Z

thierryvolpiatto
May 21, 2023
Maintainer

太阳 ***@***.***> writes:

1. ( ) multipart/mixed (*) text/html Hi, I'd like to use Chinese pinyin to narrow candidates, so if I enter wenjian, helm will not only match files (or buffers etc) like wenjian.org, but also files like 文件.org。 The following is how I archive that: 1. I add a new variable helm-pattern-transformer-alist. (defvar helm-pattern-transformer-alist '() "An alist of regex building functions for each source. Each key is the source name: Each value is a function that should take a string and return a valid regex.") 2. I change function helm-process-pattern-transformer to use helm-pattern-transformer-alist in the following way (defun helm-process-pattern-transformer (pattern source) "Execute pattern-transformer attribute function(s) on PATTERN in SOURCE." (let* ((name (assoc-default 'name source)) (transformer (assoc-default name helm-pattern-transformer-alist))) (helm-aif (or transformer (assoc-default 'pattern-transformer source)) (helm-apply-functions-from-source source it pattern) pattern))) 3. If you want a particular source to support narrowing with Chinese pinyin, do this (add-to-list 'helm-pattern-transformer-alist '("Buffers" . helm-pyim-to-utf8)) My questions are: 1. Is using pattern-transformer in this way the correct way?

Yes it is correct but IMHO there is no need to modify helm source code, see below.

2. If the answer to the first question is no, how should I do to archive my goal?

What about this in your init file? (cl-defmethod helm-setup-user-source ((source helm-source-buffers)) (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8)) This will affect helm-buffers-list and helm-mini.

…

-- Thierry

0 replies

zbelial · 2023-05-22T02:06:11Z

zbelial
May 22, 2023
Author

Thanks very much.

I tried your method with some sources, but it didn't work in some of them. Following is what I did. It worked in helm-source-buffers and helm-moccur-class, but not for the other three.

            (cl-defmethod helm-setup-user-source ((source helm-source-buffers))
              (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

            (cl-defmethod helm-setup-user-source ((source helm-moccur-class))
              (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))
            
            (cl-defmethod helm-setup-user-source ((source helm-source-basic-bookmarks))
              (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

            (cl-defmethod helm-setup-user-source ((source helm-recentf-source))
              (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

            (cl-defmethod helm-setup-user-source ((source helm-source-ffiles))
              (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

Also I found two problems in helm-find-files, here is a screenshot showing it

As you can see, the first is the file-error (it says the filename is too long) , and the second is when the file does not exist, helm will use the transformed pattern as a new filename.

The file-error happened in function helm-ff-filter-candidate-one-by-one at line 4097 , it's expected but a little annoying. It seems that a file-exists-p check can prevent this error report.

The second problem is because of this line, where it uses path (based on helm-pattern, so it's transformed). I'm not sure, but maybe making path base on helm-input makes sense?

0 replies

thierryvolpiatto · 2023-05-22T05:59:30Z

thierryvolpiatto
May 22, 2023
Maintainer

太阳 ***@***.***> writes:

Thanks very much. I tried your method with some sources, but it didn't work in some of them. Following is what I did. It worked in helm-source-buffers and helm-moccur-class, but not for the other three. (cl-defmethod helm-setup-user-source ((source helm-source-basic-bookmarks)) (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

For this one it can't work because the source is already defined in helm-bookmarks, use this instead: (helm-set-attr 'pattern-transformer #'helm-pyim-to-utf8 helm-source-bookmarks)

(cl-defmethod helm-setup-user-source ((source helm-recentf-source)) (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

Same for this one, use this instead: (helm-set-attr 'pattern-transformer #'helm-pyim-to-utf8 helm-source-recentf)

(cl-defmethod helm-setup-user-source ((source helm-source-ffiles)) (setf (slot-value source 'pattern-transformer) 'helm-pyim-to-utf8))

For helm-find-files, it is more complex (fuzzy matching), as the transformation of pattern occurs in the candidates function, so I think pattern-transformer is not a solution here, hence the errors you have. Could you provide the function helm-pyim-to-utf8? Thanks.

…

-- Thierry

0 replies

zbelial · 2023-05-22T10:24:06Z

zbelial
May 22, 2023
Author

Thanks. You're so kind and patient!

Here is helm-pyim-to-utf8 (it depnds on another package pyim)

(defun helm-pyim-to-utf8 (str)
              (cond ((equal 0 (length str))
	             str)
	            (t
	             (pyim-cregexp-build str 2))))

0 replies

thierryvolpiatto · 2023-05-22T18:41:00Z

thierryvolpiatto
May 22, 2023
Maintainer

太阳 ***@***.***> writes:

1. ( ) multipart/mixed (*) text/html Thanks. You're so kind and patient! Here is helm-pyim-to-utf8 (it depnds on another package pyim) (defun helm-pyim-to-utf8 (str) (cond ((equal 0 (length str)) str) (t (pyim-cregexp-build str 2))))

Thanks, the problem is this doesn't handle fuzzy, so helm-find-files is trying to transform the regexp returned by helm-pyim-to-utf8 to a fuzzy regexp! (pyim-cregexp-build "文件.org" 2) =>"\\(?:文件.\\(?:org\\|[偶区吽呕哦喔噢嚄怄欧殴沤瓯耦藕讴鸥][乳人亻仁仍仞任偌儒入冉冗刃喏嚅嚷壤壬如妊娆孺容嵘弱忍恁惹戎扔扰揉攘日朊枘染柔桡棁榕汝洳润溶溽濡热然熔燃狨瑞瓤睿禳稔穰箬糅纫纴绒绕缛肉肜芮苒若茸茹荏荛荣葚蓉蓐蕊蕤薷蚋蚺蛲蝾融蠕衽褥襦认让讱赁蹂轫软辱铷锐闰阮鞣韧颥饪饶髯鸮][丐丨个乖乾亘亢仡估伽佝供倌傀光公共关冈冠冮刚刮刽刿剐割功勾匦匮卦古句各合告呙呱呷咕咖咣咯哌哏哥哽哿嗝嘎嘏噶囗固国圪圭坩垓垢埂埚塥够夬夹妫姑媾孤宄官宫寡尕尜尬尴岗岣崞崮工巩帼干广庋庚廾弓彀归怪恭惯感戆戈戤扛扞拐括拱挂掴掼搁搞搿擀改攻故敢旮旰晷更杆杠杲构果枸柑柜栝根格桂桄桧梏梗棍棺椁概榖槁槔橄歌毂毌氿汞汩沟沩沽泔涡涫淦港溉滚澉灌炅炔牁牯牿犷狗猓玕珙琯瑰瓘瓜甘疙疳痼癸皈皋盖盥睾瞽矜矸砝硅硌磙秆稿竿笱筻箍管篙篝簋粿糕红纥纲纶绀给绠绲缑缟缸罐罟罡羔羹耕耿聒肛肝股肱胍胱-胳脘脶膈膏臌舡舸艮芥苟苷茛莞菇菏菰葛藁虢虼蚣蛄蛊蛤蜗蜾蝈衮袼裹褂观规觏觚觥诂诖诟诡该诰谷贡购贯贵贾赅赓赣赶跟跪躬轧轨轱辊辜过逛遘郜郭酐酤钆钙钢钩钴铬锅锆锢镉镐闺阁阖陔隔雇革鞲顾颃颈颌馃馆馘骨骼髂高鬲鬼鲑鲠鲧鲴鳏鳜鸪鸹鸽鹄鹘鹳鼓龚龟]\\)\\)" (helm--mapconcat-pattern (pyim-cregexp-build "文件.org" 2)) => "[^\\]*?\\\\[^(]*?([^?]*?\\?[^:]*?:[^文]*?文[^件]*?件 [^.]*?\\.[^\\]*?\\\\[^(]*?([^?]*?\\?[^:]*?:[^o]*?o[^r]*?r[^g]*?g[^\\]*?\\\\[^|]*?|[^[]*?\\[[^ 偶]*?偶[^区]*?区[^吽]*?吽[^呕]*?呕[^哦]*?哦[^喔]*?喔[^噢]*?噢[^嚄]*?嚄[^ 怄]*?怄[^欧]*?欧[^殴]*?殴[^沤]*?沤[^瓯]*?瓯[^耦]*?耦[^藕]*?藕[^讴]*?讴[^ 鸥]*?鸥[^]]*?][^[]*?\\[[^乳]*?乳[^人]*?人[^亻]*?亻[^仁]*?仁[^仍]*?仍 [^仞]*?仞[^任]*?任[^偌]*?偌[^儒]*?儒[^入]*?入[^冉]*?冉[^冗]*?冗[^刃]*?刃 [^喏]*?喏[^嚅]*?嚅[^嚷]*?嚷[^壤]*?壤[^壬]*?壬[^如]*?如[^妊]*?妊[^娆]*?娆 [^孺]*?孺[^容]*?容[^嵘]*?嵘[^弱]*?弱[^忍]*?忍[^恁]*?恁[^惹]*?惹[^戎]*?戎 [^扔]*?扔[^扰]*?扰[^揉]*?揉[^攘]*?攘[^日]*?日[^朊]*?朊[^枘]*?枘[^染]*?染 [^柔]*?柔[^桡]*?桡[^棁]*?棁[^榕]*?榕[^汝]*?汝[^洳]*?洳[^润]*?润[^溶]*?溶 [^溽]*?溽[^濡]*?濡[^热]*?热[^然]*?然[^熔]*?熔[^燃]*?燃[^狨]*?狨[^瑞]*?瑞 [^瓤]*?瓤[^睿]*?睿[^禳]*?禳[^稔]*?稔[^穰]*?穰[^箬]*?箬[^糅]*?糅[^纫]*?纫 [^纴]*?纴[^绒]*?绒[^绕]*?绕[^缛]*?缛[^肉]*?肉[^肜]*?肜[^芮]*?芮[^苒]*?苒 [^若]*?若[^茸]*?茸[^茹]*?茹[^荏]*?荏[^荛]*?荛[^荣]*?荣[^葚]*?葚[^蓉]*?蓉 [^蓐]*?蓐[^蕊]*?蕊[^蕤]*?蕤[^薷]*?薷[^蚋]*?蚋[^蚺]*?蚺[^蛲]*?蛲[^蝾]*?蝾 [^融]*?融[^蠕]*?蠕[^衽]*?衽[^褥]*?褥[^襦]*?襦[^认]*?认[^让]*?让[^讱]*?讱 [^赁]*?赁[^蹂]*?蹂[^轫]*?轫[^软]*?软[^辱]*?辱[^铷]*?铷[^锐]*?锐[^闰]*?闰 [^阮]*?阮[^鞣]*?鞣[^韧]*?韧[^颥]*?颥[^饪]*?饪[^饶]*?饶[^髯]*?髯[^鸮]*?鸮 [^]]*?][^[]*?\\[[^丐]*?丐[^丨]*?丨[^个]*?个[^乖]*?乖[^乾]*?乾[^亘]*?亘[^ 亢]*?亢[^仡]*?仡[^估]*?估[^伽]*?伽[^佝]*?佝[^供]*?供[^倌]*?倌[^傀]*?傀[^ 光]*?光[^公]*?公[^共]*?共[^关]*?关[^冈]*?冈[^冠]*?冠[^冮]*?冮[^刚]*?刚[^ 刮]*?刮[^刽]*?刽[^刿]*?刿[^剐]*?剐[^割]*?割[^功]*?功[^勾]*?勾[^匦]*?匦[^ 匮]*?匮[^卦]*?卦[^古]*?古[^句]*?句[^各]*?各[^合]*?合[^告]*?告[^呙]*?呙[^ 呱]*?呱[^呷]*?呷[^咕]*?咕[^咖]*?咖[^咣]*?咣[^咯]*?咯[^哌]*?哌[^哏]*?哏[^ 哥]*?哥[^哽]*?哽[^哿]*?哿[^嗝]*?嗝[^嘎]*?嘎[^嘏]*?嘏[^噶]*?噶[^囗]*?囗[^ 固]*?固[^国]*?国[^圪]*?圪[^圭]*?圭[^坩]*?坩[^垓]*?垓[^垢]*?垢[^埂]*?埂[^ 埚]*?埚[^塥]*?塥[^够]*?够[^夬]*?夬[^夹]*?夹[^妫]*?妫[^姑]*?姑[^媾]*?媾[^ 孤]*?孤[^宄]*?宄[^官]*?官[^宫]*?宫[^寡]*?寡[^尕]*?尕[^尜]*?尜[^尬]*?尬[^ 尴]*?尴[^岗]*?岗[^岣]*?岣[^崞]*?崞[^崮]*?崮[^工]*?工[^巩]*?巩[^帼]*?帼[^ 干]*?干[^广]*?广[^庋]*?庋[^庚]*?庚[^廾]*?廾[^弓]*?弓[^彀]*?彀[^归]*?归[^ 怪]*?怪[^恭]*?恭[^惯]*?惯[^感]*?感[^戆]*?戆[^戈]*?戈[^戤]*?戤[^扛]*?扛[^ 扞]*?扞[^拐]*?拐[^括]*?括[^拱]*?拱[^挂]*?挂[^掴]*?掴[^掼]*?掼[^搁]*?搁[^ 搞]*?搞[^搿]*?搿[^擀]*?擀[^改]*?改[^攻]*?攻[^故]*?故[^敢]*?敢[^旮]*?旮[^ 旰]*?旰[^晷]*?晷[^更]*?更[^杆]*?杆[^杠]*?杠[^杲]*?杲[^构]*?构[^果]*?果[^ 枸]*?枸[^柑]*?柑[^柜]*?柜[^栝]*?栝[^根]*?根[^格]*?格[^桂]*?桂[^桄]*?桄[^ 桧]*?桧[^梏]*?梏[^梗]*?梗[^棍]*?棍[^棺]*?棺[^椁]*?椁[^概]*?概[^榖]*?榖[^ 槁]*?槁[^槔]*?槔[^橄]*?橄[^歌]*?歌[^毂]*?毂[^毌]*?毌[^氿]*?氿[^汞]*?汞[^ 汩]*?汩[^沟]*?沟[^沩]*?沩[^沽]*?沽[^泔]*?泔[^涡]*?涡[^涫]*?涫[^淦]*?淦[^ 港]*?港[^溉]*?溉[^滚]*?滚[^澉]*?澉[^灌]*?灌[^炅]*?炅[^炔]*?炔[^牁]*?牁[^ 牯]*?牯[^牿]*?牿[^犷]*?犷[^狗]*?狗[^猓]*?猓[^玕]*?玕[^珙]*?珙[^琯]*?琯[^ 瑰]*?瑰[^瓘]*?瓘[^瓜]*?瓜[^甘]*?甘[^疙]*?疙[^疳]*?疳[^痼]*?痼[^癸]*?癸[^ 皈]*?皈[^皋]*?皋[^盖]*?盖[^盥]*?盥[^睾]*?睾[^瞽]*?瞽[^矜]*?矜[^矸]*?矸[^ 砝]*?砝[^硅]*?硅[^硌]*?硌[^磙]*?磙[^秆]*?秆[^稿]*?稿[^竿]*?竿[^笱]*?笱[^ 筻]*?筻[^箍]*?箍[^管]*?管[^篙]*?篙[^篝]*?篝[^簋]*?簋[^粿]*?粿[^糕]*?糕[^ 红]*?红[^纥]*?纥[^纲]*?纲[^纶]*?纶[^绀]*?绀[^给]*?给[^绠]*?绠[^绲]*?绲[^ 缑]*?缑[^缟]*?缟[^缸]*?缸[^罐]*?罐[^罟]*?罟[^罡]*?罡[^羔]*?羔[^羹]*?羹[^ 耕]*?耕[^耿]*?耿[^聒]*?聒[^肛]*?肛[^肝]*?肝[^股]*?股[^肱]*?肱[^胍]*?胍[^ 胱]*?胱[^-]*?-[^胳]*?胳[^脘]*?脘[^脶]*?脶[^膈]*?膈[^膏]*?膏[^臌]*?臌 [^舡]*?舡[^舸]*?舸[^艮]*?艮[^芥]*?芥[^苟]*?苟[^苷]*?苷[^茛]*?茛[^莞]*?莞 [^菇]*?菇[^菏]*?菏[^菰]*?菰[^葛]*?葛[^藁]*?藁[^虢]*?虢[^虼]*?虼[^蚣]*?蚣 [^蛄]*?蛄[^蛊]*?蛊[^蛤]*?蛤[^蜗]*?蜗[^蜾]*?蜾[^蝈]*?蝈[^衮]*?衮[^袼]*?袼 [^裹]*?裹[^褂]*?褂[^观]*?观[^规]*?规[^觏]*?觏[^觚]*?觚[^觥]*?觥[^诂]*?诂 [^诖]*?诖[^诟]*?诟[^诡]*?诡[^该]*?该[^诰]*?诰[^谷]*?谷[^贡]*?贡[^购]*?购 [^贯]*?贯[^贵]*?贵[^贾]*?贾[^赅]*?赅[^赓]*?赓[^赣]*?赣[^赶]*?赶[^跟]*?跟 [^跪]*?跪[^躬]*?躬[^轧]*?轧[^轨]*?轨[^轱]*?轱[^辊]*?辊[^辜]*?辜[^过]*?过 [^逛]*?逛[^遘]*?遘[^郜]*?郜[^郭]*?郭[^酐]*?酐[^酤]*?酤[^钆]*?钆[^钙]*?钙 [^钢]*?钢[^钩]*?钩[^钴]*?钴[^铬]*?铬[^锅]*?锅[^锆]*?锆[^锢]*?锢[^镉]*?镉 [^镐]*?镐[^闺]*?闺[^阁]*?阁[^阖]*?阖[^陔]*?陔[^隔]*?隔[^雇]*?雇[^革]*?革 [^鞲]*?鞲[^顾]*?顾[^颃]*?颃[^颈]*?颈[^颌]*?颌[^馃]*?馃[^馆]*?馆[^馘]*?馘 [^骨]*?骨[^骼]*?骼[^髂]*?髂[^高]*?高[^鬲]*?鬲[^鬼]*?鬼[^鲑]*?鲑[^鲠]*?鲠 [^鲧]*?鲧[^鲴]*?鲴[^鳏]*?鳏[^鳜]*?鳜[^鸪]*?鸪[^鸹]*?鸹[^鸽]*?鸽[^鹄]*?鹄 [^鹘]*?鹘[^鹳]*?鹳[^鼓]*?鼓[^龚]*?龚[^龟]*?龟 [^]]*?][^\\]*?\\\\[^)]*?)[^\\]*?\\\\[^)]*?)" So this can't work, try to see if prepending a space before the basename in the Helm pattern matches your file, i.e. "/home/you/ 文件.org" instead of "/home/you/文件.org". This will have for effect to abort fuzzy matching and switches to regexp matching.

…

-- Thierry

0 replies

zbelial · 2023-05-23T01:28:05Z

zbelial
May 23, 2023
Author

Thanks. I'll try to understand the cause and try what you recommended.

0 replies

zbelial · 2023-05-23T01:59:39Z

zbelial
May 23, 2023
Author

Also, I had another idea to use candidate-transformer (not touching pattern-transformer) and I made a try, but didn't succeed.

The idea is for candidate that contains Chinese character, transform it to Chinese pinyin. For example, 文件.org will be transformed to wen文jian件.org, so I can use wenjian to match this file.

Here is the transformer (depends on pyim too. The complete modification is here. At that time I didn't know that I can use cl-defmethod helm-setup-user-source to modify a source's slot, so I added a helm-candidate-transformer-alist, and some modification was rough):

(defun helm-replace-chinese (arg)
  (let ((cand (or (car-safe arg)
                  arg)))
    (mapconcat (lambda (c)
                 (cond
                  ((eq (char-charset c) 'ascii)
                   (char-to-string c))
                  (t
                   (let ((pinyin (pyim-pymap-cchar2py-get c)))
                     (if pinyin
                         (concat (nth 0 pinyin) (char-to-string c))
                       (char-to-string c))))))
               cand)))

;; it does not handle candidates with form (disp . real)
(defun helm-chinese-candidate-transformer (args)
  (cl-loop for arg in args
           collect (cons arg (helm-replace-chinese arg))   
           ))

I thought this way may be better (fuzzy match will not be affected(right?) and no Regular expression too big problem raised sometimes in the pattern-transformer method), but it just couldn't match Chinese characters.
Can you give me some hints about this method? Thanks.

0 replies

thierryvolpiatto · 2023-05-23T05:00:17Z

thierryvolpiatto
May 23, 2023
Maintainer

太阳 ***@***.***> writes:

I thought this way may be better (fuzzy match will not be affected(right?) and no Regular expression too big problem raised sometimes in the pattern-transformer method), but it just couldn't match Chinese characters.

Not sure to understand, what did you type at prompt at this time? Chinese or normal string?

Can you give me some hints about this method? Thanks.

I have a better idea that may work (and would be much faster). Here is what I am using in my helm config file (in the helm-setup-user-source defmethod) to match accentued characters e.g. typing "telechargement" match "Téléchargement": (helm-aif (slot-value source 'match) (setf (slot-value source 'match) (append it '((lambda (candidate) (string-match (concat (helm-basedir helm-input) (char-fold-to-regexp (helm-basename helm-input))) candidate)))))) Perhaps you could do the same, something like: (helm-aif (slot-value source 'match) (setf (slot-value source 'match) (append it '((lambda (candidate) (string-match (concat (helm-basedir helm-input) (helm-replace-chinese (helm-basename helm-input))) candidate)))))) Is this working for you?

…

-- Thierry

0 replies

thierryvolpiatto · 2023-05-23T05:05:50Z

thierryvolpiatto
May 23, 2023
Maintainer

太阳 ***@***.***> writes: Ah! Forget to mention that this function is correct but you forget the separator ("") in mapconcat, so it may fail silently which would explain why your method doesn't work.

…

(defun helm-replace-chinese (arg) (let ((cand (or (car-safe arg) arg))) (mapconcat (lambda (c) (cond ((eq (char-charset c) 'ascii) (char-to-string c)) (t (let ((pinyin (pyim-pymap-cchar2py-get c))) (if pinyin (concat (nth 0 pinyin) (char-to-string c)) (char-to-string c)))))) cand)))

-- Thierry

0 replies

thierryvolpiatto · 2023-05-23T05:15:34Z

thierryvolpiatto
May 23, 2023
Maintainer

Thierry Volpiatto ***@***.***> writes:

太阳 ***@***.***> writes: > I thought this way may be better (fuzzy match will not be > affected(right?) and no Regular expression too big problem raised > sometimes in the pattern-transformer method), but it just couldn't > match Chinese characters. Not sure to understand, what did you type at prompt at this time? Chinese or normal string? > Can you give me some hints about this method? Thanks. I have a better idea that may work (and would be much faster). Here is what I am using in my helm config file (in the helm-setup-user-source defmethod) to match accentued characters e.g. typing "telechargement" match "Téléchargement": (helm-aif (slot-value source 'match) (setf (slot-value source 'match) (append it '((lambda (candidate) (string-match (concat (helm-basedir helm-input) (char-fold-to-regexp (helm-basename helm-input))) candidate)))))) Perhaps you could do the same, something like: (helm-aif (slot-value source 'match) (setf (slot-value source 'match) (append it '((lambda (candidate) (string-match (concat (helm-basedir helm-input) (helm-replace-chinese (helm-basename helm-input))) candidate))))))

Sorry, instead of helm-replace-chinese you should use IIUC a function that make a chinese regexp.

…

Is this working for you?

-- Thierry

0 replies

thierryvolpiatto · 2023-05-23T05:31:39Z

thierryvolpiatto
May 23, 2023
Maintainer

Thierry Volpiatto ***@***.***> writes:

Thierry Volpiatto ***@***.***> writes: > 太阳 ***@***.***> writes: > >> I thought this way may be better (fuzzy match will not be >> affected(right?) and no Regular expression too big problem raised >> sometimes in the pattern-transformer method), but it just couldn't >> match Chinese characters. > > Not sure to understand, what did you type at prompt at this time? > Chinese or normal string? > >> Can you give me some hints about this method? Thanks. > > I have a better idea that may work (and would be much faster). Here is > what I am using in my helm config file (in the helm-setup-user-source > defmethod) to match accentued characters e.g. typing "telechargement" > match "Téléchargement": > > (helm-aif (slot-value source 'match) > (setf (slot-value source 'match) > (append it > '((lambda (candidate) > (string-match (concat (helm-basedir helm-input) > (char-fold-to-regexp > (helm-basename helm-input))) > candidate)))))) > > Perhaps you could do the same, something like: > > (helm-aif (slot-value source 'match) > (setf (slot-value source 'match) > (append it > '((lambda (candidate) > (string-match (concat (helm-basedir helm-input) > (helm-replace-chinese > (helm-basename helm-input))) > candidate)))))) Sorry, instead of helm-replace-chinese you should use IIUC a function that make a chinese regexp.

Sorry, this is wrong as well, finally I could match your file "文件.org" with this: (helm-aif (slot-value source 'match) (setf (slot-value source 'match) (append it '((lambda (candidate) (string-match helm-input (concat (helm-basedir candidate) (helm-replace-chinese (helm-basename candidate))))))))) Here what I did (correct me if wrong): 1) create file "/home/me/tmp/文件.org". 2) Launch helm-find-files. 3) type /home/me/tmp/we and helm matched you file "文件.org". Is it what you want to achieve?

…

-- Thierry

0 replies

zbelial · 2023-05-23T07:09:29Z

zbelial
May 23, 2023
Author

Not sure to understand, what did you type at prompt at this time?
Chinese or normal string?

Normal string.

Here what I did (correct me if wrong):

create file "/home/me/tmp/文件.org".

Launch helm-find-files.

type /home/me/tmp/we
and helm matched you file "文件.org".

Is it what you want to achieve?

Great! Thanks! It's almost what I want to achieve, I just need to find a better helm-replace-chinese.

Thank you very much.

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support narrowing with Chinese pinyin via `pattern-transformer` #2601

{{title}}

Replies: 12 comments

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Select a reply

Support narrowing with Chinese pinyin via pattern-transformer #2601

zbelial May 21, 2023

Replies: 12 comments

thierryvolpiatto May 21, 2023 Maintainer

zbelial May 22, 2023 Author

thierryvolpiatto May 22, 2023 Maintainer

zbelial May 22, 2023 Author

thierryvolpiatto May 22, 2023 Maintainer

zbelial May 23, 2023 Author

zbelial May 23, 2023 Author

thierryvolpiatto May 23, 2023 Maintainer

thierryvolpiatto May 23, 2023 Maintainer

thierryvolpiatto May 23, 2023 Maintainer

thierryvolpiatto May 23, 2023 Maintainer

zbelial May 23, 2023 Author

Support narrowing with Chinese pinyin via `pattern-transformer` #2601

zbelial
May 21, 2023

thierryvolpiatto
May 21, 2023
Maintainer

zbelial
May 22, 2023
Author

thierryvolpiatto
May 22, 2023
Maintainer

zbelial
May 22, 2023
Author

thierryvolpiatto
May 22, 2023
Maintainer

zbelial
May 23, 2023
Author

zbelial
May 23, 2023
Author

thierryvolpiatto
May 23, 2023
Maintainer

thierryvolpiatto
May 23, 2023
Maintainer

thierryvolpiatto
May 23, 2023
Maintainer

thierryvolpiatto
May 23, 2023
Maintainer

zbelial
May 23, 2023
Author