UDPipe2UD.py

#! /usr/bin/python -i
# "UDPipe2UD.py" by Koichi Yasuoka, July 3, 2019.

class UDPipeEntry:
  def __init__(self,result):
    self.result=result
    if "\n" in result:
      t=[]
      for r in result.split("\n"):
        w=UDPipeEntry(r)
        if w.id>0:
          t.append(w)
      for i,w in enumerate(t):
        w.head=w if w.head==0 else t[i+w.head-w.id]
      self.tokens=t
    else:
      w=result.split("\t")
      try:
        w[0],w[6]=int(w[0]),int(w[6])
      except:
        w=[0]*10
      self.id,self.form,self.lemma,self.upos,self.xpos,self.feats,self.head,self.deprel,self.deps,self.misc=w if len(w)==10 else [0]*10
  def __repr__(self):
    if type(self.result) is str:
      return self.result
    return self.result.encode("utf-8")
  def __getitem__(self,item):
    return self.tokens[item]
  def __len__(self):
    return len(self.tokens)
  def browse(self):
    self.editor(url="http://kanji.zinbun.kyoto-u.ac.jp/~yasuoka/kyodokenkyu/ud-kanbun/conllusvg/viewer.svg")
  def editor(self,url="http://kanji.zinbun.kyoto-u.ac.jp/~yasuoka/kyodokenkyu/ud-kanbun/conllusvg/editor.html"):
    import webbrowser
    try:
      import urllib.parse
      u=urllib.parse.quote(str(self))
    except:
      import urllib
      u=urllib.quote(str(self))
    webbrowser.open(url+"#"+u)

class UDPipe2UD:
  def __init__(self,lang="ja",option="tokenizer&tagger&parser"):
    self.parseURL="http://lindat.mff.cuni.cz/services/udpipe/api/process?model="+lang+"&"+option
  def __call__(self,sentence):
    import json
    s=sentence if type(sentence) is str else sentence.encode("utf-8")
    try:
      import urllib.request,urllib.parse
      with urllib.request.urlopen(self.parseURL+"&data="+urllib.parse.quote(s)) as r:
        q=r.read()
    except:
      import urllib,urllib2
      r=urllib2.urlopen(self.parseURL+"&data="+urllib.quote(s))
      q=r.read().decode("utf-8")
    return UDPipeEntry(json.loads(q)["result"])