diff options
| author | Patrick Simianer <p@simianer.de> | 2013-01-21 12:29:43 +0100 |
|---|---|---|
| committer | Patrick Simianer <p@simianer.de> | 2013-01-21 12:29:43 +0100 |
| commit | 0d23f8aecbfaf982cd165ebfc2a1611cefcc7275 (patch) | |
| tree | 8eafa6ea43224ff70635cadd4d6f027d28f4986f /python/tests/extractor/run.sh | |
| parent | dbc66cd3944321961c5e11d5254fd914f05a98ad (diff) | |
| parent | 7cac43b858f3b681555bf0578f54b1f822c43207 (diff) | |
Merge remote-tracking branch 'upstream/master'
Diffstat (limited to 'python/tests/extractor/run.sh')
| -rwxr-xr-x | python/tests/extractor/run.sh | 14 |
1 files changed, 14 insertions, 0 deletions
diff --git a/python/tests/extractor/run.sh b/python/tests/extractor/run.sh new file mode 100755 index 00000000..f44da9f8 --- /dev/null +++ b/python/tests/extractor/run.sh @@ -0,0 +1,14 @@ +#!/usr/bin/env bash + +# Make sure that the sa and online extractors are producing the same (correct) output + +set -x verbose + +python -m cdec.sa.compile -a corpus.al.gz -b corpus.fr-en.gz -o extract >| extract.ini + +cat test.in | python -m cdec.sa.extract -c extract.ini -g gold -o 2>&1 | egrep '\[X\].+\|\|\|.+\|\|\|.+\|\|\|.+\|\|\|'|sed -re 's/INFO.+://g' | ./refmt.py | LC_ALL=C sort >| rules.sort + +cd gold && cat grammar.0|sed -re 's/Egiv.+(IsSingletonF=)/\1/g'|LC_ALL=C sort >| rules.sort && cd .. + +diff gold/rules.sort gold-rules.sort +diff rules.sort gold-rules.sort |