fix: handle the internal structure better, take care of multiprocessing

src/epy_reader/ebooks/epub.py

@@ -169,7 +169,7 @@ def initialize(self) -> None:
         assert relative_toc is not None
         relative_toc_path = relative_toc.get("href")
         assert relative_toc_path is not None
-        toc_path = self.root_dirpath + relative_toc_path
+        toc_path = os.path.normpath(os.path.join(self.root_dirpath, relative_toc_path))
         toc = ET.parse(self.file.open(toc_path)).getroot()
         self.toc_entries = Epub._get_tocs(toc, version, contents)  # *self.contents (absolute path)
 
@@ -199,4 +199,5 @@ def get_img_bytestr(self, impath: str) -> Tuple[str, bytes]:
         return impath, self.file.read(impath)
 
     def cleanup(self) -> None:
-        pass
+        if isinstance(self.file, zipfile.ZipFile):
+            self.file.close()

src/epy_reader/reader.py

@@ -147,7 +147,8 @@ def run_counting_letters(self):
                 self._process_counting_letter = multiprocessing.Process(
                     name="epy-subprocess-counting-letters",
                     target=count_letters_parallel,
-                    args=(copy.deepcopy(self.ebook), self._proc_child),
+                    # Pass only the filepath, not the entire ebook object
+                    args=(self.ebook.path, self._proc_child),
                 )
                 # forking will raise
                 # zlib.error: Error -3 while decompressing data: invalid distance too far back

src/epy_reader/utils.py

@@ -4,6 +4,7 @@
 import sys
 import textwrap
 from functools import wraps
+from multiprocessing.connection import Connection
 from typing import List, Mapping, Optional, Sequence, Tuple, Union
 
 from epy_reader.ebooks import URL, Azw, Ebook, Epub, FictionBook, Mobi
@@ -344,11 +345,10 @@ def construct_relative_reading_state(
         section=abs_reading_state.section,
     )
 
-
 def count_letters(ebook: Ebook) -> LettersCount:
     per_content_counts: List[int] = []
     cumulative_counts: List[int] = []
-    # assert isinstance(ebook.contents, tuple)
+    assert isinstance(ebook.contents, tuple)
     for i in ebook.contents:
         content = ebook.get_raw_text(i)
         src_lines = parse_html(content)
@@ -358,11 +358,18 @@ def count_letters(ebook: Ebook) -> LettersCount:
 
     return LettersCount(all=sum(per_content_counts), cumulative=tuple(cumulative_counts))
 
-
-def count_letters_parallel(ebook: Ebook, child_conn) -> None:
-    child_conn.send(count_letters(ebook))
-    child_conn.close()
-
+def count_letters_parallel(filepath: str, conn: Connection):
+    """
+    Counts letters in an ebook in a separate process.
+    Receives the ebook's filepath and a connection for sending results.
+    """
+    # The child process opens its own ebook object
+    ebook_obj_in_child_process = get_ebook_obj(filepath)
+    ebook_obj_in_child_process.initialize()
+    letters_count = count_letters(ebook_obj_in_child_process)
+    ebook_obj_in_child_process.cleanup()
+    conn.send(letters_count)
+    conn.close()
 
 def construct_speaker(
     preferred: Optional[str] = None, args: List[str] = []