Initial commit

config.ini

+; PLIK KONFIGURACYJNY  WORKERA
+; Plik zawiera konfigurację zarówno Api usługi sieciowej jak i narzędzia.
+;
+;       Autor: Tomasz Walkowiak
+;       email: tomasz.walkowiak@pwr.edu.pl
+
+; --------- CZĘŚĆ DLA Serwisu ---------
+[service]
+#root = /mnt2/requests/
+root = /samba/requests/
+tool = maca
+rabbit_host =10.17.0.85
+rabbit_user =clarin
+rabbit_password =clarin123
+
+; --------- CZĘŚĆ DLA Narzedzia ---------
+[tool]
+workers_number = 1
+
+[logging]
+port = 9995
+local_log_level = INFO

maca_worker.py

+#!/usr/bin/python
+# -*- coding: utf-8 -*-
+
+
+import nlp_ws
+import logging
+
+import maca
+import corpus2
+
+def sentences(reader):
+	"""Yields subsequent sentences from a reader."""
+	while True:
+		sentence = reader.get_next_sentence()
+		if not sentence:
+			break
+		yield sentence
+
+def chunks(reader):
+	"""Yields subsequent sentences from a reader."""
+	while True:
+		chunk = reader.get_next_chunk()
+		if not chunk:
+			break
+		yield chunk
+
+
+_log = logging.getLogger(__name__)
+
+class MacaWorker(nlp_ws.NLPWorker):
+	
+	@classmethod
+	def static_init(cls, config):
+		_log.info( "Worker started loading models %s","AS" )
+		cls.configtool = config['tool'];
+		return
+
+	def init(self):
+		_log.info( "Worker started loading models" )
+			
+	def process(self, inputFile, taskOptions, outputFile):	
+		maca_config='morfeusz2-nkjp'
+		if 'morfeusz2' in taskOptions:
+			if not taskOptions['morfeusz2']:
+				maca_config='morfeusz-nkjp-official'
+			
+		_log.info( "Config %s",maca_config)
+		reader = maca.PlainTextReader.create_file_reader(str(inputFile), maca_config)
+		writer = corpus2.TokenWriter.create_path_writer("ccl",str(outputFile),reader.tagset())
+		for chunk in chunks(reader):
+			writer.write_chunk(chunk)
+			
+
+if __name__ == '__main__':
+	nlp_ws.NLPService.main(MacaWorker)
+	

old/maca_service.py

+#!/usr/bin/python
+# -*- coding: utf-8 -*-
+
+from nlp_service import Service, create_service_option_parser
+import shutil
+from subprocess import call
+
+import maca
+import corpus2
+
+
+def sentences(reader):
+	"""Yields subsequent sentences from a reader."""
+	while True:
+		sentence = reader.get_next_sentence()
+		if not sentence:
+			break
+		yield sentence
+
+def chunks(reader):
+	"""Yields subsequent sentences from a reader."""
+	while True:
+		chunk = reader.get_next_chunk()
+		if not chunk:
+			break
+		yield chunk
+
+
+class MacaService(Service):
+	def __init__(self, *args, **kwargs):
+		self.maca_config='morfeusz2-nkjp'
+		super(MacaService, self).__init__(*args, **kwargs)
+
+	def process(self, inputFile, taskOptions, outputFile):
+		reader = maca.PlainTextReader.create_file_reader(inputFile, self.maca_config)
+		writer = corpus2.TokenWriter.create_path_writer("ccl",outputFile,reader.tagset())
+		for chunk in chunks(reader):
+			writer.write_chunk(chunk)
+		#shutil.move(inputFile, outputFile)
+	
+	
+def test(service):
+	service.process('test.txt',0,'out.ccl')
+	
+if __name__ == '__main__':
+	parser = create_service_option_parser()
+	args = parser.parse_args()
+	config_path = args.config_path
+	logfile_path = args.log_file
+	logging_lvl = args.logging_lvl
+	run_as_daemon = args.daemon
+
+	service = MacaService(config_path, logfile_path, logging_lvl, run_as_daemon)
+	#test(service);
+	service.run()