Merge pull request #28 from SUSYUSTC/dev

Dev

mathtranslate/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "2.1.7"
+__version__ = "2.1.8"
 __author__ = "Jiace Sun"
 
 import os

mathtranslate/process_latex.py

@@ -5,7 +5,7 @@
 match_code = r"(" + math_code + r"_\d+(?:_\d+)*)"
 match_code_replace = math_code + r"_(\d+(?:_\d+)*)*"
 
-options = r"\[[a-zA-Z\s,\\\*\.\+\-=_]*?\]"  # ,\*.+-=_
+options = r"\[[a-zA-Z\s,\\\*\.\+\-=_{}]*?\]"  # ,\*.+-=_{}
 
 pattern_env = r"\\begin[ \t]*\{(.*?)\}[ \t]*(options)?(.*?)\\end[ \t]*\{\1\}".replace('options', options)  # \begin{xxx} \end{xxx}, group 1: name, group 2: option, group 3: content
 pattern_command_full = r"\\([a-zA-Z]+\*?)[ \t]*(options)?[ \t]*(\{((?:[^{}]++|(?3))++)\})".replace('options', options)   # \xxx[xxx]{xxx} and \xxx{xxx}, group 1: name, group 2: option, group 4: content
@@ -100,7 +100,7 @@ def replace_latex_objects(text):
         pattern = regex.compile(regex_symbol, regex.DOTALL)
         while pattern.search(text):
             latex_obj = pattern.search(text).group()
-            replaced_objs.append(f'{latex_obj}')
+            replaced_objs.append(f' {latex_obj} ')
             text = pattern.sub(' ' + variable_code(count) + ' ', text, 1)
             count += 1
 
@@ -297,3 +297,16 @@ def replace_function(match):
     text = re.compile(match_code_accent).sub(replace_function, text)
 
     return text
+
+
+def combine_sentences(text):
+    pattern = re.compile(r'\n(\s*([^\s]))')
+
+    def process_function(match):
+        char = match.group(2)
+        if char == '\\':
+            return match.group(0)
+        else:
+            return ' '+match.group(1)
+
+    return pattern.sub(process_function, text)

mathtranslate/translate.py

@@ -92,13 +92,14 @@ def translate_paragraph_latex(self, latex_original_paragraph, num, complete):
         '''
         Translate a latex paragraph, which means that it could contain latex objects
         '''
+        latex_original_paragraph = process_latex.combine_sentences(latex_original_paragraph)
         text_original_paragraph, objs = process_latex.replace_latex_objects(latex_original_paragraph)
         # Since \n is equivalent to space in latex, we change \n back to space
         # otherwise the translators view them as separate sentences
-        text_original_paragraph = text_original_paragraph.replace('\n', '')
         text_original_paragraph = process_text.split_too_long_paragraphs(text_original_paragraph)
         if not complete:
             text_original_paragraph = process_text.split_titles(text_original_paragraph)
+        text_original_paragraph = re.sub(r'  +', ' ', text_original_paragraph)
         if self.debug:
             print(f'\n\nParagraph {num}\n\n', file=self.f_old)
             print(text_original_paragraph, file=self.f_old)