concordia-server/fast-aligner/Makefile
2019-02-04 15:27:56 +01:00

62 lines
2.8 KiB
Makefile

SRC_LANG=pl
TRG_LANG=en
CORPUS_NAME=opensubtitles_sample
SEPARATOR=@\#@
DICTIONARY_WEIGHT=3
all: corpora/$(CORPUS_NAME)/src.lem corpora/$(CORPUS_NAME)/trg.lem
clean-intermediate-files:
rm -f corpora/$(CORPUS_NAME)/*.lem
rm -f corpora/$(CORPUS_NAME)/*.tok
rm -f corpora/$(CORPUS_NAME)/*.dict
rm -f corpora/$(CORPUS_NAME)/*.classes
rm -f corpora/$(CORPUS_NAME)/*.classes.cats
rm -f corpora/$(CORPUS_NAME)/*.vcb
rm -f corpora/$(CORPUS_NAME)/*.snt
rm -f corpora/$(CORPUS_NAME)/*.cooc
rm -f corpora/$(CORPUS_NAME)/aligned*part*
rm -f corpora/$(CORPUS_NAME)/aligned.txt
rm -f corpora/$(CORPUS_NAME)/giza.cfg
rm -f corpora/$(CORPUS_NAME)/aligned.gizacfg
rm -f corpora/$(CORPUS_NAME)/pasted.txt
rm -f corpora/$(CORPUS_NAME)/pasted_deduplicated.txt
rm -f corpora/$(CORPUS_NAME)/src_clean.txt
rm -f corpora/$(CORPUS_NAME)/trg_clean.txt
clean: clean-intermediate-files
rm -f corpora/$(CORPUS_NAME)/src_final.txt
rm -f corpora/$(CORPUS_NAME)/trg_final.txt
rm -f corpora/$(CORPUS_NAME)/aligned_final.txt
corpora/$(CORPUS_NAME)/src.dict:
./collect_dict.py $(SRC_LANG) $(TRG_LANG) $(DICTIONARY_WEIGHT) > $@
corpora/$(CORPUS_NAME)/trg.dict:
./collect_dict.py $(TRG_LANG) $(SRC_LANG) $(DICTIONARY_WEIGHT) > $@
corpora/$(CORPUS_NAME)/src.lem: corpora/$(CORPUS_NAME)/src.txt
/usr/local/bin/concordia-sentence-tokenizer -c ../concordia.cfg < $< | ./sentence_lemmatizer.py $(SRC_LANG) > $@
corpora/$(CORPUS_NAME)/trg.lem: corpora/$(CORPUS_NAME)/trg.txt
/usr/local/bin/concordia-sentence-tokenizer -c ../concordia.cfg < $< | ./sentence_lemmatizer.py $(TRG_LANG) > $@
corpora/$(CORPUS_NAME)/src_clean.txt corpora/$(CORPUS_NAME)/trg_clean.txt corpora/$(CORPUS_NAME)/src_clean.tok corpora/$(CORPUS_NAME)/trg_clean.tok: corpora/$(CORPUS_NAME)/src.txt corpora/$(CORPUS_NAME)/trg.txt corpora/$(CORPUS_NAME)/src.tok corpora/$(CORPUS_NAME)/trg.tok
./clean_corpus.py corpora/$(CORPUS_NAME)/src.txt corpora/$(CORPUS_NAME)/trg.txt corpora/$(CORPUS_NAME)/src.tok corpora/$(CORPUS_NAME)/trg.tok corpora/$(CORPUS_NAME)/src_clean.txt corpora/$(CORPUS_NAME)/trg_clean.txt corpora/$(CORPUS_NAME)/src_clean.tok corpora/$(CORPUS_NAME)/trg_clean.tok $(SEPARATOR)
corpora/$(CORPUS_NAME)/pasted.txt: corpora/$(CORPUS_NAME)/src_clean.txt corpora/$(CORPUS_NAME)/trg_clean.txt
./paste.py corpora/$(CORPUS_NAME)/src_clean.txt corpora/$(CORPUS_NAME)/trg_clean.txt $(SEPARATOR)> $@
corpora/$(CORPUS_NAME)/pasted_deduplicated.txt: corpora/$(CORPUS_NAME)/pasted.txt
sort -k 1.13 $< | uniq -s 12 | sort > $@
corpora/$(CORPUS_NAME)/aligned_final.txt corpora/$(CORPUS_NAME)/src_final.txt corpora/$(CORPUS_NAME)/trg_final.txt: corpora/$(CORPUS_NAME)/pasted_deduplicated.txt corpora/$(CORPUS_NAME)/aligned.txt
./extract.py $< corpora/$(CORPUS_NAME)/aligned.txt corpora/$(CORPUS_NAME)/aligned_final.txt corpora/$(CORPUS_NAME)/src_final.txt corpora/$(CORPUS_NAME)/trg_final.txt $(SEPARATOR)