Translation

book/book.original.tex

@@ -0,0 +1,73 @@
+\documentclass[paper=a4, twoside=false, fontsize=12pt, parskip=half,
+                bibliography=totoc, listof=totoc]{scrbook}
+
+% General packages
+\usepackage{sourdough}
+
+% Basic attributes
+\author{Hendrik Kleinwächter}
+\title{The Sourdough Framework}
+
+\begin{document}
+\input{cover/cover}
+\titlepage
+
+\frontmatter
+{%
+\hypersetup{hidelinks}
+\ifdefined\HCode\else\tableofcontents\fi
+}
+
+\input{intro/foreword}
+\input{intro/preface}
+\input{intro/acknowledgements}
+
+\mainmatter
+
+\input{history/sourdough-history}
+
+\chapter{How sourdough works}
+\input{basics/how-sourdough-works}
+
+\chapter{Making a sourdough starter}
+\input{sourdough-starter/sourdough-starter}
+
+\chapter{Sourdough starter types}
+\input{sourdough-starter/sourdough-starter-types}
+
+\chapter{Flour types}
+\input{flour-types/flour-types}
+
+\chapter{Bread types}
+\input{bread-types/bread-types}
+
+\chapter{Wheat sourdough}%
+\label{chapter:wheat-sourdough}
+\input{wheat-sourdough/wheat-sourdough}
+
+\chapter{Non wheat sourdough}%
+\label{chapter:non-wheat-sourdough}
+\input{non-wheat-sourdough/non-wheat-sourdough}
+
+% \input{mix-ins/mix-ins}
+\chapter{Baking}%
+\label{chapter:baking}
+\input{baking/baking}
+
+\chapter{Storing bread}%
+\label{chapter:storing-bread}
+\input{storing-bread/storing-bread}
+
+\chapter{Troubleshooting}
+\input{troubleshooting/misc}
+
+\backmatter
+\printbibliography
+{%
+\hypersetup{hidelinks}
+\listofflowcharts
+\listoftables
+\listoffigures
+}
+
+\end{document}

book/book.tex

@@ -9,7 +9,7 @@
 \title{The Sourdough Framework}
 
 \begin{document}
-\input{cover/cover.tex}
+\input{cover/cover-translated}
 \titlepage
 
 \frontmatter
@@ -23,7 +23,7 @@
 
 \mainmatter
 
-\input{history/sourdough-history}
+\input{history/sourdough-history-translated}
 
 \input{basics/how-sourdough-works}
 
@@ -36,8 +36,8 @@
 \input{bread-types/bread-types}
 
 \chapter{Wheat sourdough}%
-\label{chapter:wheat-sourdough}
-\input{wheat-sourdough/wheat-sourdough}
+\label{chapter:wheat-sourdough-translated}
+\input{wheat-sourdough/wheat-sourdough-translated}
 
 \input{non-wheat-sourdough/non-wheat-sourdough}
 
@@ -48,7 +48,7 @@ \chapter{Wheat sourdough}%
 \input{storing-bread/storing-bread}
 
 \chapter{Troubleshooting}
-\input{troubleshooting/misc}
+\input{troubleshooting/misc-translated}
 
 \backmatter
 \input{glossary/glossary}

translater.py

@@ -0,0 +1,130 @@
+import openai
+import os
+import re
+import sys
+
+from dotenv import load_dotenv
+
+load_dotenv()
+
+openai.api_key = os.getenv('OPENAI_KEY')
+
+lang = "french"
+model = "gpt-4"
+system = f"""
+You are a translating assistant. 
+You translate in {lang}. 
+You directly translate messages your receive, without saying anything else than the translation of the messages. 
+The content you take as input are LaTeX document. 
+You need to only translate the text that is rendered on the output. Do not translate comments, labels, references, file paths, or anything that is syntax-related or that would break the LaTex compilation.
+"""
+max_chunk_length = 1000*4
+
+def read_file(filepath):
+    file = open(filepath, 'r')
+    file_content = file.read()
+    return file_content
+
+def write_file(filepath, content):
+    file = open(filepath, 'w')
+    file.write(str(content))
+
+def translate(content):
+    chat_completion = openai.ChatCompletion.create(
+        model=model, 
+        messages=[
+            {"role": "system", "content": system},
+            {"role": "user", "content": content},
+        ]
+    )
+
+    translation = chat_completion.choices[0].message.content
+
+    return translation
+
+def translate_big(big_content):
+    content_paragraphs = big_content.split('\n\n') # maybe take lines instead of paragraphs
+
+    chunks = []
+    chunk = ""
+
+    for index, paragraph in enumerate(content_paragraphs):
+        if len(chunk + paragraph) > max_chunk_length:
+            if len(chunk) == 0:
+                raise Exception("Paragraph length exceeded maximum chunk size")
+            chunks.append(chunk)
+            chunk = ""
+        chunk += paragraph
+        if index == len(content_paragraphs)-1:
+            chunks.append(chunk)
+            break;
+        else:
+            chunk += "\n\n"
+
+    translated_chunks = []
+    for index, chunk in enumerate(chunks):
+        print(f'PENDING - Translating chunk {index+1}/{len(chunks)}')
+
+        translated_chunk = translate(chunk)
+        # translated_chunk = chunk
+
+        print(f'OK - Done translating chunk {index+1}/{len(chunks)}')
+        translated_chunks.append(translated_chunk)
+
+    translated = "".join(translated_chunks)
+
+    return translated
+
+def list_files():
+    book_root = './book'
+    tex_string = read_file(os.path.join(book_root, 'book.original.tex'))
+
+    pattern = r'\\input\{(.*?)\}'
+    matches = re.findall(pattern, tex_string)
+
+    files_to_translate = [os.path.join(book_root, match) for match in matches]
+
+    return files_to_translate
+
+
+def translate_one_file(input_path, output_path):
+    if os.path.exists(output_path):
+        print(f'SKIP - File already translated')
+        return
+
+    content = read_file(input_path)
+    translated_content = translate_big(content)
+    write_file(output_path, translated_content)
+
+def translate_all_files():
+    print(f'Translating book to {lang}')
+
+    files_to_translate = list_files()
+
+    #files_to_translate = [files_to_translate[1]]
+    #print(files_to_translate)
+
+    for file_to_translate in files_to_translate:
+        print(f'PENDING - Translating {file_to_translate}')
+        input_file = file_to_translate + '.tex'
+        output_file = file_to_translate + f'.{lang}.tex'
+        translate_one_file(input_file, output_file)
+        print(f'OK - Done translating {file_to_translate}')
+
+    print(f'Done translating book')
+
+def copy_translations():
+    files_to_translate = list_files()
+    for file_to_translate in files_to_translate:
+        filename = file_to_translate + f'.{lang}.tex'
+        if not os.path.exists(filename):
+            filename = file_to_translate + '.tex'
+        content = read_file(filename)
+        write_file(file_to_translate + '-translated.tex', content)
+
+if __name__ == "__main__":
+    if "--copy" in sys.argv:
+        copy_translations()
+    else:
+        translate_all_files()
+        copy_translations()