IssueID #1487: expanded conllu postprocessing and cleaned a bit

scripts/constants.py

@@ -2,6 +2,7 @@
 TMP_DIRECTORY = '../tmp/structure_assignment'
 
 # scripts
+CONLLU_TWEAK_SCRIPT_NAME = 'tweak_conllu.py'
 CONLLU_TEI_SCRIPT_NAME = 'conllu_to_xml.py'
 MWE_EXTRACTION_SCRIPT_NAME = 'wani.py'
 STRUCTURE_ASSIGNMENT_SCRIPT_NAME = 'assign_structures.py'

scripts/fix_xml_ids.py

@@ -1,21 +0,0 @@
-import argparse
-import codecs
-import re
-
-arg_parser = argparse.ArgumentParser(description='Fix invalid XML ids.')
-arg_parser.add_argument('-infile', type=str, help='Input file')
-arg_parser.add_argument('-outfile', type=str, help='Output file')
-arguments = arg_parser.parse_args()
-input_file_name = arguments.infile
-output_file_name = arguments.outfile
-
-output_file = codecs.open(output_file_name, 'w')
-input_file = codecs.open(input_file_name, 'r')
-
-for line in input_file:
-    line = re.sub('xml:id="(?=\d)','xml:id="s', line)
-    line = line.replace('#', '#s')
-    output_file.write(line)
-    
-input_file.close()
-output_file.close()

scripts/pipeline1.py

@@ -2,7 +2,6 @@ import argparse
 import os
 import shutil
 import codecs
-import re
 
 import classla
 from classla import Document
@@ -31,28 +30,20 @@ def run_pipeline(input_file_name, output_file_name):
     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)
-    fix_xml_ids(OBELIKS_RAW_FILE_NAME, OBELIKS_TWEAKED_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(obeliks_file_name, classla_file_name):
+def run_obeliks4J(list_file_name, conllu_file_name):
     print('Running obeliks ...')
-    obeliks_command = 'java -jar ' + OBELIKS_JAR_FILE_NAME + ' -d -if ' + STRING_LIST_FILE_NAME + ' -o ' + OBELIKS_RAW_FILE_NAME
+    obeliks_command = 'java -jar ' + OBELIKS_JAR_FILE_NAME + ' -d -if ' + list_file_name + ' -o ' + conllu_file_name
     os.system(obeliks_command)
 
-def fix_xml_ids(input_file_name, output_file_name):
-    print('Fixing xml ids ...')
-    output_file = codecs.open(output_file_name, 'w')
-    input_file = codecs.open(input_file_name, 'r')
-    regexp = r'^(# sent_id = )(\d+\.\d+)$'
-    for line in input_file:
-        match = re.search(regexp, line)
-        if (match):
-            line = match.group(1) + XML_ID_PREFIX + match.group(2) + '\n'
-        output_file.write(line)
-    input_file.close()
-    output_file.close()
+def tweak_conllu(input_file_name, output_file_name):
+    print('Tweaking conllu results ...')
+    tweak_command = ' '.join(['python', CONLLU_TWEAK_SCRIPT_NAME, '-infile', input_file_name, '-outfile', output_file_name])
+    os.system(tweak_command)
 
 def run_classla(obeliks_file_name, classla_file_name):
     print('Running classla ...')

scripts/tweak_conllu.py

@@ -0,0 +1,46 @@
+import argparse
+import codecs
+import re
+
+arg_parser = argparse.ArgumentParser(description='Fix invalid XML ids.')
+arg_parser.add_argument('-infile', type=str, help='Input file')
+arg_parser.add_argument('-outfile', type=str, help='Output file')
+arguments = arg_parser.parse_args()
+input_file_name = arguments.infile
+output_file_name = arguments.outfile
+
+output_file = codecs.open(output_file_name, 'w')
+input_file = codecs.open(input_file_name, 'r')
+
+def write(output_file, line):
+    output_file.write(line + '\n')
+
+def write_paragraph(output_file, output_map):
+    if (output_map is not None):
+        write(output_file, output_map['paragraph'])
+        write(output_file, output_map['sentence'])
+        write(output_file, '# text = ' + ' '.join(output_map['texts']))
+        for (index, token_line) in enumerate(output_map['tokens'], start=1):
+            write(output_file, '\t'.join([str(index)] + token_line.split('\t')[1:]))
+        write(output_file, '')
+
+output_map = None
+for line in input_file:
+    if (line[0].isdigit()):
+        output_map['tokens'].append(line.strip())
+    else:
+        match = re.search('^# (.+) = (.+)$', line)
+        if (match):
+            (name, value) = match.groups()
+            if (name == 'newpar id'):
+                write_paragraph(output_file, output_map)
+                output_map = {'paragraph': line.strip(), 'sentence':None, 'texts':[], 'tokens':[]}
+            elif (name == 'sent_id'):
+                if (value.endswith('.1')):
+                    output_map['sentence'] = re.sub('^(# sent_id = )(\d+\.1)$', r'\1s\2', line.strip())
+            elif (name == 'text'):
+                output_map['texts'].append(value)
+write_paragraph(output_file, output_map)
+
+input_file.close()
+output_file.close()