From 0172721855098ca02b207231a654dffa5e4eb1c9 Mon Sep 17 00:00:00 2001
From: redpony <redpony@ec762483-ff6d-05da-a07a-a48fb63a330f>
Date: Tue, 22 Jun 2010 05:12:27 +0000
Subject: initial checkin

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@2 ec762483-ff6d-05da-a07a-a48fb63a330f
---
 word-aligner/makefiles/makefile.grammars | 70 ++++++++++++++++++++++++++++++++
 1 file changed, 70 insertions(+)
 create mode 100644 word-aligner/makefiles/makefile.grammars

(limited to 'word-aligner/makefiles/makefile.grammars')

diff --git a/word-aligner/makefiles/makefile.grammars b/word-aligner/makefiles/makefile.grammars
new file mode 100644
index 00000000..b89937c1
--- /dev/null
+++ b/word-aligner/makefiles/makefile.grammars
@@ -0,0 +1,70 @@
+all: corpus.f-e.lex-grammar.gz corpus.e-f.lex-grammar.gz corpus.class.e corpus.class.f weights.init.gz
+
+clean:
+	$(RM) orthonorm-dict.* voc2class* corpus.class.* corpus.e-f corpus.f-e weights* corpus.f-e.lex-grammar* *.model1 *voc corpus.e-f.lex-grammar*
+
+SUPPORT_DIR = $(SCRIPT_DIR)/support
+GZIP = /usr/bin/gzip
+ZCAT = zcat
+EXTRACT_WEIGHTS = $(SUPPORT_DIR)/extract_weights.pl
+EXTRACT_GRAMMAR = $(SUPPORT_DIR)/extract_grammar.pl
+SUPPLEMENT_WEIGHTS = $(SUPPORT_DIR)/supplement_weights_file.pl
+EXTRACT_VOCAB = $(SUPPORT_DIR)/extract_vocab.pl
+ORTHONORM_E = $(SCRIPT_DIR)/ortho-norm/$(E_LANG).pl
+ORTHONORM_F = $(SCRIPT_DIR)/ortho-norm/$(F_LANG).pl
+CLASSIFY = $(SUPPORT_DIR)/classify.pl
+MAKE_LEX_GRAMMAR = $(SUPPORT_DIR)/make_lex_grammar.pl
+MODEL1 = $(TRAINING_DIR)/model1
+MERGE_CORPUS = $(SUPPORT_DIR)/merge_corpus.pl
+
+orthonorm-dict.e: corpus.e
+	$(EXTRACT_VOCAB) corpus.e > e.voc
+	$(ORTHONORM_E) < e.voc > e.ortho-voc
+	$(MERGE_CORPUS) e.voc e.ortho-voc > $@
+
+orthonorm-dict.f: corpus.f
+	$(EXTRACT_VOCAB) corpus.f > f.voc
+	$(ORTHONORM_F) < f.voc > f.ortho-voc
+	$(MERGE_CORPUS) f.voc f.ortho-voc > $@
+
+voc2class.e: corpus.e $(MKCLS)
+	$(MKCLS) -c$(NCLASSES) -n10 -pcorpus.e -Vvoc2class.e opt
+
+voc2class.f: corpus.f $(MKCLS)
+	$(MKCLS) -c$(NCLASSES) -n10 -pcorpus.f -Vvoc2class.f opt
+
+corpus.class.e: corpus.e voc2class.e $(CLASSIFY)
+	$(CLASSIFY) voc2class.e corpus.e > $@
+
+corpus.class.f: corpus.f voc2class.f $(CLASSIFY)
+	$(CLASSIFY) voc2class.f corpus.f > $@
+
+corpus.f-e: corpus.f corpus.e $(MERGE_CORPUS)
+	$(MERGE_CORPUS) corpus.f corpus.e > $@
+
+corpus.e-f: corpus.f corpus.e $(MERGE_CORPUS)
+	$(MERGE_CORPUS) corpus.e corpus.f > $@
+
+corpus.f-e.model1: corpus.f-e $(MODEL1)
+	$(MODEL1) corpus.f-e > corpus.f-e.model1
+
+corpus.e-f.model1: corpus.e-f $(MODEL1)
+	$(MODEL1) corpus.e-f > corpus.e-f.model1
+
+bidir.grammars: corpus.f-e corpus.f-e.model1 corpus.e-f.model1 orthonorm-dict.f orthonorm-dict.e voc2class.e voc2class.f
+	$(RM) $@
+	$(MAKE_LEX_GRAMMAR) corpus.f-e corpus.f-e.model1 corpus.e-f.model1 orthonorm-dict.f orthonorm-dict.e voc2class.e voc2class.f > bidir.grammars
+
+corpus.f-e.lex-grammar.gz: bidir.grammars
+	$(EXTRACT_GRAMMAR) 1 bidir.grammars | $(GZIP) -9 > corpus.f-e.lex-grammar.gz
+
+corpus.e-f.lex-grammar.gz: bidir.grammars
+	$(EXTRACT_GRAMMAR) 2 bidir.grammars | $(GZIP) -9 > corpus.e-f.lex-grammar.gz
+
+weights.init.gz: bidir.grammars voc2class.f voc2class.e
+	$(EXTRACT_WEIGHTS) bidir.grammars > weights.init
+	$(SUPPLEMENT_WEIGHTS) voc2class.f > weights.dup
+	$(SUPPLEMENT_WEIGHTS) voc2class.e >> weights.dup
+	sort -u weights.dup >> weights.init
+	$(GZIP) -9 weights.init
+
-- 
cgit v1.2.3