summaryrefslogtreecommitdiff
path: root/word-aligner
diff options
context:
space:
mode:
authorMichael Denkowski <mdenkows@cs.cmu.edu>2013-08-19 08:23:42 -0700
committerMichael Denkowski <mdenkows@cs.cmu.edu>2013-08-19 08:23:42 -0700
commitac469cdf4c70154a1c2cedce9edf5cdc3bdb2d61 (patch)
tree3e8bcbc6b00533e7a79e3cf28c2ac6aa4bdadd8d /word-aligner
parentf4a3a2547316ca5d31366e6808858fe94981415c (diff)
Realtime translation (mostly a cdec wrapper for now)
Diffstat (limited to 'word-aligner')
-rwxr-xr-xword-aligner/force_align.py46
1 files changed, 8 insertions, 38 deletions
diff --git a/word-aligner/force_align.py b/word-aligner/force_align.py
index a0c1aad7..ad6d95fa 100755
--- a/word-aligner/force_align.py
+++ b/word-aligner/force_align.py
@@ -2,7 +2,6 @@
import os
import sys
-import subprocess
def main():
@@ -15,45 +14,16 @@ def main():
sys.stderr.write(' {} fwd_params fwd_err rev_params rev_err <in.f-e >out.f-e.gdfa\n'.format(sys.argv[0]))
sys.exit(2)
- (f_p, f_err, r_p, r_err) = sys.argv[1:]
-
- (f_T, f_m) = find_Tm(f_err)
- (r_T, r_m) = find_Tm(r_err)
-
- fast_align = os.path.join(os.path.dirname(__file__), 'fast_align')
- f_cmd = [fast_align, '-i', '-', '-d', '-T', f_T, '-m', f_m, '-f', f_p]
- r_cmd = [fast_align, '-i', '-', '-d', '-T', r_T, '-m', r_m, '-f', r_p, '-r']
-
- atools = os.path.join(os.path.dirname(os.path.dirname(__file__)), 'utils', 'atools')
- tools_cmd = [atools, '-i', '-', '-j', '-', '-c', 'grow-diag-final-and']
-
- sys.stderr.write('running: {}\n'.format(' '.join(f_cmd)))
- sys.stderr.write('running: {}\n'.format(' '.join(r_cmd)))
- sys.stderr.write('running: {}\n'.format(' '.join(tools_cmd)))
-
- f_a = subprocess.Popen(f_cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE)
- r_a = subprocess.Popen(r_cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE)
- tools = subprocess.Popen(tools_cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE)
+ # Hook into realtime
+ sys.path.append(os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), 'realtime'))
+ from rt import ForceAligner
+ aligner = ForceAligner(*sys.argv[1:])
+
for line in sys.stdin:
- f_a.stdin.write(line)
- r_a.stdin.write(line)
- # f words ||| e words ||| links ||| score
- f_line = f_a.stdout.readline().split('|||')[2].strip()
- r_line = r_a.stdout.readline().split('|||')[2].strip()
- tools.stdin.write('{}\n'.format(f_line))
- tools.stdin.write('{}\n'.format(r_line))
- sys.stdout.write(tools.stdout.readline())
-
-def find_Tm(err):
- (T, m) = ('', '')
- for line in open(err):
- # expected target length = source length * N
- if 'expected target length' in line:
- m = line.split()[-1]
- elif 'final tension' in line:
- T = line.split()[-1]
- return (T, m)
+ sys.stdout.write('{}\n'.format(aligner.align(line.strip())))
+ aligner.close()
+
if __name__ == '__main__':
main()