From 7d3fa9e2ae31ae20420ed889b7edec2316cf1769 Mon Sep 17 00:00:00 2001 From: rjawor Date: Wed, 29 Aug 2018 12:56:47 +0200 Subject: [PATCH] installation on poleng computer, clean up --- .gitignore | 3 +++ INSTALL.txt | 2 +- cat/versions_available/europarl_sample.cfg | 14 +++++++------- cat/versions_enabled/europarl_sample.cfg | 1 + mgiza-aligner/Makefile | 6 +++--- tests/addLemmatizedTM.sh | 6 +++--- 6 files changed, 18 insertions(+), 14 deletions(-) create mode 120000 cat/versions_enabled/europarl_sample.cfg diff --git a/.gitignore b/.gitignore index 8824f97..f5b2913 100644 --- a/.gitignore +++ b/.gitignore @@ -1,6 +1,7 @@ build/ logs/concordia-server.log logs/pgbouncer.log +pgbouncer.log logs/phrase-searches.json concordia.cfg concordia-server/config.hpp @@ -40,3 +41,5 @@ import-requests/request_handler.pid mgiza-aligner/corpus-compilator/corpora mgiza-aligner/corpus-compilator/dictionaries/*lem mgiza-aligner/corpus-compilator/dictionary-compilator/*.tsv +mgiza-aligner/mgiza/mgizapp/inst/ +mgiza-aligner/mgiza/mgizapp/install_manifest.txt diff --git a/INSTALL.txt b/INSTALL.txt index 49f6848..66f10fb 100644 --- a/INSTALL.txt +++ b/INSTALL.txt @@ -6,7 +6,7 @@ - grant all privileges on database concordia_server to concordia; - sudo vim /etc/postgresql/9.3/main/pg_hba.conf: change "local all all peer" to "local all all md5" - pgbouncer: - - sudo apt-get install autoconf automake m4 libtool pkg-config libevent-dev autogen + - sudo apt-get install autoconf automake m4 libtool pkg-config libevent-dev autogen python-docutils $ git clone https://github.com/pgbouncer/pgbouncer.git $ cd pgbouncer $ git submodule init diff --git a/cat/versions_available/europarl_sample.cfg b/cat/versions_available/europarl_sample.cfg index 9494336..d6a4004 100644 --- a/cat/versions_available/europarl_sample.cfg +++ b/cat/versions_available/europarl_sample.cfg @@ -1,10 +1,10 @@ dir@#@europarl_sample concordia_host@#@localhost concordia_port@#@8800 -tmid@#@1 -desc@#@Europarl sample (1000 sentences) -enjoy@#@Życzymy udanej pracy z systemem! -prompt@#@Wprowadź zdanie (po polsku): -suggestion@#@Na każde państwo członkowskie Unii Europejskiej przypada jeden komisarz. -suggestion@#@Komisja Europejska przygotowuje raport na najbliższym posiedzeniu. -suggestion@#@Wspólny Komitet przyjmuje swój statut. +tmid@#@2 +desc@#@Welcome to the interactive Concordia demo. The system finds the longest matches of the pattern sentence in its translation memory. This translation memory is over 1.5M sentences taken from English-Polish corpus of European Law (Europarl + JRC-Acquis). Please enter an English sentence in the field below and press Enter (or use the search button). This instance of Concordia works best with law sentences, but is very likely to output some results for any English sentence. You can also use predefined samples, simply use the link "show/hide samples" and apply one of the sample sentences. After the search, click on the highlighted fragments to see their context. +enjoy@#@Enjoy your work with the system! +prompt@#@Enter search pattern (English sentence): +suggestion@#@Every ship in the European Union must have a crew of 50 or more workers. +suggestion@#@It is impossible to abolish the customs duties on fruit and vegetables. +suggestion@#@The convention on human rights was held in Geneva. diff --git a/cat/versions_enabled/europarl_sample.cfg b/cat/versions_enabled/europarl_sample.cfg new file mode 120000 index 0000000..c90ed2e --- /dev/null +++ b/cat/versions_enabled/europarl_sample.cfg @@ -0,0 +1 @@ +../versions_available/europarl_sample.cfg \ No newline at end of file diff --git a/mgiza-aligner/Makefile b/mgiza-aligner/Makefile index 8ad6958..23df30a 100644 --- a/mgiza-aligner/Makefile +++ b/mgiza-aligner/Makefile @@ -1,9 +1,9 @@ -SRC_LANG=pl -TRG_LANG=en +SRC_LANG=en +TRG_LANG=pl CORPUS_NAME=europarl_sample SEPARATOR=@\#@ -DICTIONARY_WEIGHT=0 +DICTIONARY_WEIGHT=3 all: corpora/$(CORPUS_NAME)/aligned_final.txt corpora/$(CORPUS_NAME)/src_final.txt corpora/$(CORPUS_NAME)/trg_final.txt diff --git a/tests/addLemmatizedTM.sh b/tests/addLemmatizedTM.sh index 8a83148..8cb99b0 100755 --- a/tests/addLemmatizedTM.sh +++ b/tests/addLemmatizedTM.sh @@ -1,7 +1,7 @@ #!/bin/sh -CORPUS_NAME="stocznia_plen" -SRC_LANG_ID=1 -TRG_LANG_ID=2 +CORPUS_NAME="europarl_sample" +SRC_LANG_ID=2 +TRG_LANG_ID=1 ./addAlignedLemmatizedTM.py $CORPUS_NAME ../mgiza-aligner/corpora/$CORPUS_NAME/src_final.txt $SRC_LANG_ID ../mgiza-aligner/corpora/$CORPUS_NAME/trg_final.txt $TRG_LANG_ID ../mgiza-aligner/corpora/$CORPUS_NAME/aligned_final.txt