diff options
author | Patrick Simianer <p@simianer.de> | 2013-11-13 18:12:10 +0100 |
---|---|---|
committer | Patrick Simianer <p@simianer.de> | 2013-11-13 18:12:10 +0100 |
commit | d6e6babf2cfe49fed040b651624b7e34d1a9b507 (patch) | |
tree | 2a00ab18f10a7f93e7e172551c01b48cc9f20b8c /realtime/realtime.py | |
parent | 2d2d5eced93d58bc77894d8c328195cd9950b96d (diff) | |
parent | 8a24bb77bc2e9fd17a6f6529a2942cde96a6af49 (diff) |
merge w/ upstream
Diffstat (limited to 'realtime/realtime.py')
-rwxr-xr-x | realtime/realtime.py | 95 |
1 files changed, 77 insertions, 18 deletions
diff --git a/realtime/realtime.py b/realtime/realtime.py index dff7e90c..ec15b59d 100755 --- a/realtime/realtime.py +++ b/realtime/realtime.py @@ -3,50 +3,109 @@ import argparse import logging import sys +import threading +import time import rt +ABOUT = '''Realtime adaptive translation with cdec (See README.md) + +Code by Michael Denkowski + +Citation: +@misc{denkowski-proposal2013, + author = {Michael Denkowski}, + title = {Machine Translation for Human Translators}, + year = {2013}, + month = {May}, + day = {30}, + howpublished = {{Ph.D.} Thesis Proposal, Carnegie Mellon University} +} + +''' + class Parser(argparse.ArgumentParser): def error(self, message): + sys.stderr.write(ABOUT) self.print_help() sys.stderr.write('\n{}\n'.format(message)) sys.exit(2) +def handle_line(translator, line, output, ctx_name): + res = translator.command_line(line, ctx_name) + if res: + output.write('{}\n'.format(res)) + output.flush() + +def test1(translator, input, output, ctx_name): + inp = open(input) + out = open(output, 'w') + for line in inp: + handle_line(translator, line.strip(), out, ctx_name) + out.close() + +def debug(translator, input): + # Test 1: multiple contexts + threads = [] + for i in range(4): + t = threading.Thread(target=test1, args=(translator, input, '{}.out.{}'.format(input, i), str(i))) + threads.append(t) + t.start() + time.sleep(30) + # Test 2: flood + out = open('{}.out.flood'.format(input), 'w') + inp = open(input) + while True: + line = inp.readline() + if not line: + break + line = line.strip() + t = threading.Thread(target=handle_line, args=(translator, line.strip(), out, None)) + threads.append(t) + t.start() + time.sleep(1) + translator.drop_ctx(None) + # Join test threads + for t in threads: + t.join() + def main(): - parser = Parser(description='Real-time adaptive translation with cdec.') - parser.add_argument('-c', '--config', required=True, help='Config directory (see README.md)') + parser = Parser() + parser.add_argument('-c', '--config', required=True, help='Config directory') + parser.add_argument('-s', '--state', help='Load state file to default context (saved incremental data)') parser.add_argument('-n', '--normalize', help='Normalize text (tokenize, translate, detokenize)', action='store_true') parser.add_argument('-T', '--temp', help='Temp directory (default /tmp)', default='/tmp') parser.add_argument('-a', '--cache', help='Grammar cache size (default 5)', default='5') parser.add_argument('-v', '--verbose', help='Info to stderr', action='store_true') + parser.add_argument('-D', '--debug-test', help='Run debug tests on input file') args = parser.parse_args() if args.verbose: logging.basicConfig(level=logging.INFO) - rtd = rt.RealtimeDecoder(args.config, tmpdir=args.temp, cache_size=int(args.cache), norm=args.normalize) + with rt.RealtimeTranslator(args.config, tmpdir=args.temp, cache_size=int(args.cache), norm=args.normalize) as translator: + + # Debugging + if args.debug_test: + debug(translator, args.debug_test) + return + + # Load state if given + if args.state: + rtd.load_state(state) - try: + # Read lines and commands while True: line = sys.stdin.readline() if not line: break - input = [f.strip() for f in line.split('|||')] - if len(input) == 1: - hyp = rtd.decode(input[0]) - sys.stdout.write('{}\n'.format(hyp)) + line = line.strip() + res = translator.command_line(line) + if res: + sys.stdout.write('{}\n'.format(res)) sys.stdout.flush() - elif len(input) == 2: - rtd.learn(*input) - - # Clean exit on ctrl+c - except KeyboardInterrupt: - logging.info('Caught KeyboardInterrupt, exiting') - - # Cleanup - rtd.close() - + if __name__ == '__main__': main() |