def __init__(self, parserpath, savedir):
     self.savedir = savedir
     self.parserpath = parserpath
     self.kwargs = {}
     # Parser configuration file
     configfilepath = self.parserpath + '/' + self.savedir + '/config.cfg'
     self.kwargs['config_file'] = configfilepath
     # Pre-trained parser model
     self.kwargs['default'] = {'save_dir': self.parserpath + '/' + self.savedir}
     self.kwargs['is_evaluation'] = True
     self.network = Network(**self.kwargs)
Exemple #2
0
def parse(save_dir, **kwargs):
  """"""
  
  kwargs['config_file'] = os.path.join(save_dir, 'config.cfg')
  files = kwargs.pop('files')
  output_file = kwargs.pop('output_file', None)
  output_dir = kwargs.pop('output_dir', None)
  if len(files) > 1 and output_file is not None:
    raise ValueError('Cannot provide a value for --output_file when parsing multiple files')
  kwargs['is_evaluation'] = True
  network = Network(**kwargs)
  network.parse(files, output_file=output_file, output_dir=output_dir)
  return
Exemple #3
0
def train(save_dir, **kwargs):
  """"""
  
  kwargs['config_file'] = kwargs.pop('config_file', '')
  load = kwargs.pop('load')
  try:
    if not load and os.path.isdir(save_dir):
      raw_input('Save directory already exists. Press <Enter> to continue or <Ctrl-c> to abort.')
      if os.path.isfile(os.path.join(save_dir, 'config.cfg')):
        os.remove(os.path.join(save_dir, 'config.cfg'))
  except KeyboardInterrupt:
    print()
    sys.exit(0)
  network = Network(**kwargs)
  network.train(load=load)
  return
def network_init(save_dir, **kwargs):
    kwargs['config_file'] = os.path.join(save_dir, 'config.cfg')
    short = kwargs.pop('short', False)
    if short:
        kwargs['config_file'] = os.path.join(save_dir, 'config_short.cfg')
    kwargs['is_evaluation'] = True
    network = Network(**kwargs)
    return network
Exemple #5
0
def train(save_dir, **kwargs):
    """"""

    load = kwargs.pop('load')
    try:
        if not load and os.path.isdir(save_dir):
            raw_input(
                'Save directory already exists. Press <Enter> to continue or <Ctrl-c> to abort.'
            )
            if os.path.isfile(os.path.join(save_dir, 'config.cfg')):
                os.remove(os.path.join(save_dir, 'config.cfg'))
    except KeyboardInterrupt:
        sys.exit(0)

    #print (kwargs)
    #print ("train files:",kwargs['train_files'])
    print("initializing")
    network = Network(**kwargs)
    print("initialized")
    network.train(load=load)
    return
class UnstableParser:

    """
    Perform parsing using the UnstableParser (universal dependency parser
    """
    
    def __init__(self, parserpath, savedir):
        self.savedir = savedir
        self.parserpath = parserpath
        self.kwargs = {}
        # Parser configuration file
        configfilepath = self.parserpath + '/' + self.savedir + '/config.cfg'
        self.kwargs['config_file'] = configfilepath
        # Pre-trained parser model
        self.kwargs['default'] = {'save_dir': self.parserpath + '/' + self.savedir}
        self.kwargs['is_evaluation'] = True
        self.network = Network(**self.kwargs)


    def parse(self, outputdir, files):
        """
        Parse the files using a pre-trained UnstableParser model
        """
        self.network.parse(files, output_dir=outputdir)
Exemple #7
0
  @property
  def vocabs(self):
    return self._vocabs
  @property
  def datasets(self):
    return self._datasets
  @property
  def optimizer(self):
    return self._optimizer
  @property
  def save_vars(self):
    return filter(lambda x: u'Pretrained/Embeddings:0' != x.name, tf.global_variables())
  @property
  def non_save_vars(self):
    return filter(lambda x: u'Pretrained/Embeddings:0' == x.name, tf.global_variables())
  @property
  def global_step(self):
    return self._global_step
  @property
  def global_epoch(self):
    return self._global_epoch

#***************************************************************
if __name__ == '__main__':
  """"""
  
  from parser import Network
  configurable = Configurable()
  network = Network.from_configurable(configurable)
  network.train()