From 7651774914d1d4fdbdada5962009500fd6f35aab Mon Sep 17 00:00:00 2001
From: Cyprian Laskowski <cyp@cjvt.si>
Date: Tue, 12 Jan 2021 10:00:40 +0100
Subject: [PATCH] IssueID #1487: switched to python obeliks implementation

---
 .gitignore           | 2 +-
 requirements.txt     | 1 +
 scripts/constants.py | 1 -
 scripts/pipeline1.py | 6 +++---
 scripts/setup.sh     | 9 ---------
 5 files changed, 5 insertions(+), 14 deletions(-)

diff --git a/.gitignore b/.gitignore
index 492e6ae..6f03474 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,5 +1,5 @@
 /venv
-/Obeliks4J
+/obeliks
 /data_admin
 /luscenje_struktur
 /nova_slovnica
diff --git a/requirements.txt b/requirements.txt
index 1c46303..c008586 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -6,6 +6,7 @@ future==0.18.2
 idna==2.10
 lxml==4.5.2
 numpy==1.19.1
+obeliks==1.0.2
 pkg-resources==0.0.0
 protobuf==3.13.0
 psycopg2cffi==2.8.1
diff --git a/scripts/constants.py b/scripts/constants.py
index 2652869..96070ef 100644
--- a/scripts/constants.py
+++ b/scripts/constants.py
@@ -13,7 +13,6 @@ TEI_DICTIONARY_SCRIPT_NAME = 'tei_to_dictionary.py'
 DICTIONARY_MERGE_SCRIPT_NAME = 'merge_dictionaries.py'
 
 # resources
-OBELIKS_JAR_FILE_NAME = '../resources/obeliks.jar'
 TRANSLATION_FILE_NAME = '../resources/dict.xml'
 CLASSLA_MODELS_DIRECTORY = '../resources/classla'
 STRUCTURE_CURRENT_FILE_NAME = '../resources/structures.xml'
diff --git a/scripts/pipeline1.py b/scripts/pipeline1.py
index 0cc791b..8cd2bb3 100644
--- a/scripts/pipeline1.py
+++ b/scripts/pipeline1.py
@@ -29,15 +29,15 @@ def run_pipeline(input_file_name, output_file_name):
     shutil.rmtree(TMP_DIRECTORY, True)
     os.makedirs(TMP_DIRECTORY, exist_ok=True)
     shutil.copyfile(input_file_name, STRING_LIST_FILE_NAME)
-    run_obeliks4J(STRING_LIST_FILE_NAME, OBELIKS_RAW_FILE_NAME)
+    run_obeliks(STRING_LIST_FILE_NAME, OBELIKS_RAW_FILE_NAME)
     tweak_conllu(OBELIKS_RAW_FILE_NAME, OBELIKS_TWEAKED_FILE_NAME)
     run_classla(OBELIKS_TWEAKED_FILE_NAME, CLASSLA_FILE_NAME)
     run_tei_conversion(CLASSLA_FILE_NAME, TEI_INIT_FILE_NAME)
     shutil.copyfile(TEI_INIT_FILE_NAME, output_file_name)
 
-def run_obeliks4J(list_file_name, conllu_file_name):
+def run_obeliks(list_file_name, conllu_file_name):
     print('Running obeliks ...')
-    obeliks_command = 'java -jar ' + OBELIKS_JAR_FILE_NAME + ' -d -if ' + list_file_name + ' -o ' + conllu_file_name
+    obeliks_command = ' '.join(['obeliks', '-c', '-if', list_file_name, '-o', conllu_file_name])
     os.system(obeliks_command)
 
 def tweak_conllu(input_file_name, output_file_name):
diff --git a/scripts/setup.sh b/scripts/setup.sh
index 68d6900..ee59aec 100755
--- a/scripts/setup.sh
+++ b/scripts/setup.sh
@@ -11,19 +11,11 @@ python -c "import classla; classla.download('sl_ssj_jos')" <<< $'Y\nresources/cl
 deactivate
 
 ## get needed repositories
-git clone https://github.com/clarinsi/Obeliks4J
 git clone git@gitea.cjvt.si:redmine_projects/nova_slovnica.git
 git clone git@gitea.cjvt.si:ozbolt/luscenje_struktur.git
 git clone git@gitea.cjvt.si:generic/data_admin.git
 git clone git@gitea.cjvt.si:generic/xml_schemas.git
 
-## set up obeliks
-cd Obeliks4J
-javac -encoding UTF-8 src/main/java/org/obeliks/*.java -d target/classes
-cp src/main/resources/* target/classes/org/obeliks/
-jar -cef org.obeliks.Tokenizer obeliks.jar -C target/classes org
-cd ..
-
 ## put needed scripts in place 
 cd scripts
 ln -s ../nova_slovnica/python/scripts/conllu_to_xml.py .
@@ -35,7 +27,6 @@ cd ..
 
 ## put needed resources in place
 cd resources
-mv ../Obeliks4J/obeliks.jar .
 ln -s ../nova_slovnica/resources/dict.xml .
 ln -s ../data_admin/resources/structures.xml .
 ln -s ../data_admin/resources/structures.xsd .