change: Update EPUB parsing method to use lxml instead of BeautifulSoup

Author: mashu3

README.md

@@ -1,12 +1,18 @@
 # manga2pdf
+[![License: MIT](https://img.shields.io/pypi/l/manga2pdf)](https://opensource.org/licenses/MIT)
+[![PyPI - Python Version](https://img.shields.io/pypi/pyversions/manga2pdf)](https://pypi.org/project/manga2pdf)
+[![GitHub Release](https://img.shields.io/github/release/mashu3/manga2pdf?color=orange)](https://github.com/mashu3/manga2pdf/releases)
+[![PyPi Version](https://img.shields.io/pypi/v/manga2pdf?color=yellow)](https://pypi.org/project/manga2pdf/)
+[![Downloads](https://static.pepy.tech/badge/manga2pdf)](https://pepy.tech/project/manga2pdf)
+
 ## Overview
 This Python script is specifically designed to convert manga and comic files, including various formats such as zip, epub, and directories containing image files, to PDF format.
 
 The resulting PDF files are optimized to resemble Japanese manga in terms of page layout and direction. By default, the script uses a "TwoPageRight" page layout that displays two pages side-by-side for a spread view, and a "R2L" (right-to-left) reading direction that is commonly used in Japanese manga.
 
 ## Requirement
 The script uses the Python libraries **[img2pdf](https://pypi.org/project/img2pdf/)** and **[pikepdf](https://pypi.org/project/pikepdf/)** to do the conversion.
-Moreover, it uses **[BeautifulSoup](https://pypi.org/project/beautifulsoup4/)** to read the EPUB files and **[rarfile](https://pypi.org/project/rarfile/)** to read the RAR archive files.
+Moreover, it uses **[lxml](https://pypi.org/project/lxml/)** to read the EPUB files and **[rarfile](https://pypi.org/project/rarfile/)** to read the RAR archive files.
 
 It requires the installation of these packages in order to work properly.
 
@@ -91,4 +97,6 @@ The interface supports English and Japanese, and all settings that can be specif
 The GUI is currently under development and additional features are planned for future updates.
 
 ## Author
-[mashu3](https://github.com/mashu3)
+[mashu3](https://github.com/mashu3)
+
+[![Authors](https://contrib.rocks/image?repo=mashu3/manga2pdf)](https://github.com/mashu3/manga2pdf/graphs/contributors)

setup.py

@@ -1,14 +1,14 @@
 from setuptools import setup, find_packages
 
-VERSION = "0.1.4"
+VERSION = "0.2.0"
 
 INSTALL_REQUIRES = (
+    "lxml",
     "numpy",
     "img2pdf",
     "Pillow",
     "pikepdf",
-    "rarfile",
-    "beautifulsoup4"
+    "rarfile"
 )
 CLASSIFIERS=[
     'License :: OSI Approved :: MIT License',

src/manga2pdf.py

@@ -15,12 +15,10 @@
 import warnings
 import numpy as np
 from PIL import Image
+from lxml import etree
 import concurrent.futures
-from bs4 import BeautifulSoup
-
 warnings.filterwarnings('ignore', category=UserWarning)
-from bs4 import XMLParsedAsHTMLWarning
-warnings.filterwarnings('ignore', category=XMLParsedAsHTMLWarning)
+
 
 class MangaPdfConverter():   
     def __init__(self, input_path: str, output_path: str, pagelayout:str, pagemode:str, direction:str):
@@ -176,44 +174,45 @@ def extract_epub_contents(self, epub):
             opf_name = opf_names[0]
         page_names = []
         with epub.open(opf_name) as opf:
-            content_opf = opf.read().decode()
-            opf_soup = BeautifulSoup(content_opf, 'xml')
-            for item in opf_soup.find_all('item', {'media-type': 'image/jpeg'}):
+            opf_content = opf.read()
+            opf_tree = etree.fromstring(opf_content)
+            namespace = {'dc': 'http://purl.org/dc/elements/1.1/', 'opf': 'http://www.idpf.org/2007/opf'}
+            manifest = opf_tree.find('opf:manifest', namespaces=namespace)
+            for item in manifest.findall('opf:item[@media-type="image/jpeg"]', namespaces=namespace):
                 page_names.append(os.path.join(os.path.dirname(opf_name), item.get('href').replace('/', os.sep)).replace(os.sep, '/'))
-            for item in opf_soup.find_all('item', {'media-type': 'image/png'}):
+            for item in manifest.findall('opf:item[@media-type="image/png"]', namespaces=namespace):
                 page_names.append(os.path.join(os.path.dirname(opf_name), item.get('href').replace('/', os.sep)).replace(os.sep, '/'))
         page_items = []
         for page_name in page_names:
             page_items.append(epub.open(page_name))
         return page_names, page_items, ncx_name, opf_name
-    
+
     # Function to extract the index of an EPUB file
     def extract_epub_index(self, epub, page_names, ncx_name: str):
         page_index = []
         with epub.open(ncx_name) as ncx_file:
             ncx_content = ncx_file.read()
-        ncx_soup = BeautifulSoup(ncx_content, 'lxml')
-        navpoints = ncx_soup.find_all('navpoint')
+        ncx_tree = etree.fromstring(ncx_content)
+        namespace = {'ncx': 'http://www.daisy.org/z3986/2005/ncx/', 'html': 'http://www.w3.org/1999/xhtml', 'svg':'http://www.w3.org/2000/svg'}
+        navmap = ncx_tree.find('ncx:navMap', namespaces=namespace)
+        navpoints = navmap.findall('ncx:navPoint', namespaces=namespace)
         ncx_path = os.path.dirname(ncx_name) + '/'
         for navpoint in navpoints:
-            nav_label = navpoint.navlabel.text.strip()
-            nav_text = navpoint.content['src']
+            nav_label = navpoint.find('ncx:navLabel/ncx:text', namespaces=namespace).text.strip()
+            nav_text = navpoint.find('ncx:content', namespaces=namespace).get('src')
             if nav_text.startswith(ncx_path):
                 nav_text = nav_text[len(ncx_path):]
             else:
                 nav_text = os.path.join(ncx_path, nav_text)
             if nav_text.endswith('.xhtml'):
                 with epub.open(nav_text) as xhtml_file:
                     xhtml_content = xhtml_file.read()
-                xhtml_soup = BeautifulSoup(xhtml_content, 'html.parser')
-                img_tags = xhtml_soup.find_all('image')
+                xhtml_tree = etree.fromstring(xhtml_content)
+                img_tags = xhtml_tree.findall('.//svg:image', namespaces=namespace)
                 if len(img_tags) == 0:
-                    img_tags = xhtml_soup.find_all('img')
+                    img_tags = img_tags = xhtml_tree.findall('.//svg:img', namespaces=namespace)
                 for img_tag in img_tags:
-                    try:
-                        img_link = img_tag['xlink:href']
-                    except KeyError:
-                        img_link = img_tag['src']
+                    img_link = img_tag.get('{http://www.w3.org/1999/xlink}href', img_tag.get('src'))
                     image_href = os.path.abspath(os.path.join(os.path.dirname(nav_text), img_link))
                     image_href = os.path.relpath(image_href, os.getcwd()).replace(os.sep, '/')
                     index_number = page_names.index(image_href)
@@ -227,21 +226,22 @@ def extract_epub_index(self, epub, page_names, ncx_name: str):
     def extract_epub_metadata(self, epub, opf_name: str):
         with epub.open(opf_name) as opf_file:
             opf_content = opf_file.read()
-        opf_soup = BeautifulSoup(opf_content, 'lxml')
-        metadata = opf_soup.find('metadata')    
+        opf_tree = etree.fromstring(opf_content)
+        namespace = {'dc': 'http://purl.org/dc/elements/1.1/', 'opf': 'http://www.idpf.org/2007/opf'}
+        metadata = opf_tree.find('opf:metadata', namespaces=namespace)    
         epub_metadata = {}
         for key in ['title', 'creator', 'publisher', 'date', 'language']:
-            values = metadata.find_all('dc:'+key)
+            values = metadata.xpath('./dc:' + key, namespaces=namespace)
             if key == 'creator':
                 if values:
-                    epub_metadata[key] = [value for value in values]
+                    epub_metadata[key] = [value.text for value in values]
             else:
                 if values:
                     epub_metadata[key] = values[0].text
                 else:
                     epub_metadata[key] = None
         return epub_metadata
-    
+
     # Function to convert input files to a PDF file
     def convert(self):
         if self.is_epub_file(self.input_path):