From ba62c21945798ed351961c47f60e09a4143d19f1 Mon Sep 17 00:00:00 2001
From: "vladimir.eidelman"
 <vladimir.eidelman@ec762483-ff6d-05da-a07a-a48fb63a330f>
Date: Sun, 12 Sep 2010 19:54:11 +0000
Subject: tagger example

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@644 ec762483-ff6d-05da-a07a-a48fb63a330f
---
 tests/system_tests/tagger/TRAIN       | 1 +
 tests/system_tests/tagger/cdec.ini    | 5 +++++
 tests/system_tests/tagger/input.txt   | 2 ++
 tests/system_tests/tagger/pos.grammar | 4 ++++
 tests/system_tests/tagger/tagset      | 1 +
 tests/system_tests/tagger/weights     | 2 ++
 6 files changed, 15 insertions(+)
 create mode 100644 tests/system_tests/tagger/TRAIN
 create mode 100644 tests/system_tests/tagger/cdec.ini
 create mode 100644 tests/system_tests/tagger/input.txt
 create mode 100644 tests/system_tests/tagger/pos.grammar
 create mode 100644 tests/system_tests/tagger/tagset
 create mode 100644 tests/system_tests/tagger/weights

(limited to 'tests/system_tests')

diff --git a/tests/system_tests/tagger/TRAIN b/tests/system_tests/tagger/TRAIN
new file mode 100644
index 00000000..c4bb433a
--- /dev/null
+++ b/tests/system_tests/tagger/TRAIN
@@ -0,0 +1 @@
+../../../training/cluster-ptrain.pl cdec.ini pos.grammar weights --gaussian_prior --sigma_squared 1 &> training.log &
\ No newline at end of file
diff --git a/tests/system_tests/tagger/cdec.ini b/tests/system_tests/tagger/cdec.ini
new file mode 100644
index 00000000..15cc930d
--- /dev/null
+++ b/tests/system_tests/tagger/cdec.ini
@@ -0,0 +1,5 @@
+formalism=tagger
+feature_function=Tagger_BigramIdentity
+feature_function=LexicalPairIdentity
+intersection_strategy=full
+tagger_tagset=tagset
\ No newline at end of file
diff --git a/tests/system_tests/tagger/input.txt b/tests/system_tests/tagger/input.txt
new file mode 100644
index 00000000..fe5cb58d
--- /dev/null
+++ b/tests/system_tests/tagger/input.txt
@@ -0,0 +1,2 @@
+A B C
+A
\ No newline at end of file
diff --git a/tests/system_tests/tagger/pos.grammar b/tests/system_tests/tagger/pos.grammar
new file mode 100644
index 00000000..3415854b
--- /dev/null
+++ b/tests/system_tests/tagger/pos.grammar
@@ -0,0 +1,4 @@
+[X] ||| A B C ||| a b c
+[X] ||| A ||| a
+[X] ||| A ||| b
+[X] ||| B C ||| b c
\ No newline at end of file
diff --git a/tests/system_tests/tagger/tagset b/tests/system_tests/tagger/tagset
new file mode 100644
index 00000000..3774da60
--- /dev/null
+++ b/tests/system_tests/tagger/tagset
@@ -0,0 +1 @@
+a b c
diff --git a/tests/system_tests/tagger/weights b/tests/system_tests/tagger/weights
new file mode 100644
index 00000000..ec8358f3
--- /dev/null
+++ b/tests/system_tests/tagger/weights
@@ -0,0 +1,2 @@
+Tagger_BigramIdentity 1
+LexicalPairIdentity 1
\ No newline at end of file
-- 
cgit v1.2.3