用 jieba 为 emacs 提供中文分词扩展选区的功能

先用一个动图展示效果

先是对光标所在行进行分词，这里用的是请求 flask api, 可以改为命令行直接获取

(defun jieba-cut ()
  (interactive)
  (progn
    (setq r nil)
    (let ((text (thing-at-point 'line t)))
      (request
    "http://127.0.0.1:5000/jieba_cut"
      :type "POST"
      :data (encode-coding-string (json-encode `(("text". ,text))) 'utf-8)
      :headers '(("Content-Type" . "application/json"))
      :parser 'json-read
      :encoding 'utf-8
      :sync t
      :success (cl-function
        (lambda (&key data &allow-other-keys)
          (setq r (assoc-default 'result data)))))
    (mapcar 'identity r))))

flask 的分词 api

import jieba
@app.route('/jieba_cut', methods=['POST'])
def jieba_cut():
    data = json.loads(request.data)
    text = data.get('text')
    result = [i for i in jieba.cut(text)]
    print(result)
    return jsonify({'result': result})

下面是为 emacs 提供中文分词扩展选区的功能，并绑定在 C-- 按键上，和 expand-region 的快捷键 C-= 对应，用了递归

(defun jieba-expand2 (list point acc)
  (if (<= point (+ acc (length (car list))))
      (list acc (length (car list)))
    (jieba-expand2 (cdr list) point (+ acc (length (car list))))))

(defun jieba-expand-at-point2 ()
  (interactive)
  (let* ((p1 (point))
     (je2 (progn
        (beginning-of-line)
        (jieba-expand2 (jieba-cut)
                   (- p1 (point)) -1)))
     (beg (car je2))
     (len1 (cadr je2)))
    (forward-char (+ beg 1))
    (set-mark (point))
    (forward-char len1)))


(defun jieba-expand-at-point3 (beg end)
  (interactive "r")
  (if (region-active-p)
      (let* ((p1 (+ end 1))
         (je2 (progn
            (beginning-of-line)
            (jieba-expand2 (jieba-cut)
                   (- p1 (point)) -1)))
         (len1 (if (equal 1 (cadr je2))
               (+ (car je2) 1)
              (+ (car je2) (cadr je2) 1))))
    (set-mark beg)
    (forward-char len1))
    (let* ((p1 (point))
         (je2 (progn
            (beginning-of-line)
            (jieba-expand2 (jieba-cut)
                   (- p1 (point)) -1)))
         (beg (car je2))
         (len1 (cadr je2)))
    (forward-char (+ beg 1))
    (set-mark (point))
    (forward-char len1))))

(global-set-key (kbd "C--") 'jieba-expand-at-point3)