1 files changed, 185 insertions, 0 deletions
diff --git a/converter/extractors.py b/converter/extractors.py
new file mode 100644
index 0000000..e49b48a
--- /dev/null
+++ b/converter/extractors.py
@@ -0,0 +1,185 @@
+"""Text extraction from book files (TXT, PDF, EPUB) and text/HTML cleaning."""
+
+import logging
+import re
+import zipfile
+from html import unescape
+from pathlib import Path
+
+try:
+    from bs4 import BeautifulSoup
+    BS4_AVAILABLE = True
+except ImportError:
+    BS4_AVAILABLE = False
+
+logger = logging.getLogger(__name__)
+
+
+def extract_text(file_path: Path) -> str:
+    """Extract text from a book file based on its extension."""
+    extension = file_path.suffix.lower()
+    if extension == ".txt":
+        return _extract_txt(file_path)
+    if extension == ".pdf":
+        return _extract_pdf(file_path)
+    if extension == ".epub":
+        return extract_epub(file_path)
+    raise ValueError(f"Unsupported file format: {extension}")
+
+
+def clean_text(text: str) -> str:
+    """Normalize whitespace and strip standalone page numbers.
+
+    Page numbers are removed only when they appear as a short number alone on
+    its own line (before whitespace collapsing), so inline numbers like
+    "42 years", "1,000" or "3.5" are preserved.
+    """
+    if not text:
+        return ""
+    # Standalone page numbers (digits alone on a line) must go BEFORE the
+    # newline-collapsing step below.
+    text = re.sub(r"(?m)^\s*\d{1,4}\s*$", " ", text)
+    text = re.sub(r"\s+", " ", text)
+    return text.strip()
+
+
+def clean_html(html_content: str) -> str:
+    """Strip markup, scripts and styles from HTML content."""
+    if not html_content:
+        return ""
+
+    if BS4_AVAILABLE:
+        try:
+            soup = BeautifulSoup(html_content, "html.parser")
+            for tag in soup(["script", "style"]):
+                tag.decompose()
+            text = soup.get_text()
+            lines = (line.strip() for line in text.splitlines())
+            chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
+            return " ".join(chunk for chunk in chunks if chunk)
+        except Exception as exc:
+            logger.debug("BeautifulSoup cleaning failed, falling back to regex: %s", exc)
+
+    # Fallback regex cleaning
+    html_content = re.sub(r"<style[^>]*>.*?</style>", "", html_content, flags=re.DOTALL | re.IGNORECASE)
+    html_content = re.sub(r"<script[^>]*>.*?</script>", "", html_content, flags=re.DOTALL | re.IGNORECASE)
+    html_content = re.sub(r"<[^>]+>", " ", html_content)
+    html_content = unescape(html_content)
+    html_content = re.sub(r"\s+", " ", html_content)
+    return html_content.strip()
+
+
+def extract_epub(file_path: Path) -> str:
+    """Extract text from EPUB, trying several methods in order."""
+    methods = [
+        _extract_epub_ebooklib,
+        _extract_epub_zipfile,
+        _extract_epub_manual,
+    ]
+
+    for method in methods:
+        try:
+            text = method(file_path)
+            if text and text.strip():
+                logger.info("EPUB extraction successful (%s): %d characters", method.__name__, len(text))
+                return text
+        except Exception as exc:
+            logger.warning("EPUB method %s failed: %s", method.__name__, exc)
+
+    raise RuntimeError("All EPUB extraction methods failed")
+
+
+def _extract_epub_ebooklib(file_path: Path) -> str:
+    """Extract using ebooklib, following the spine (reading) order."""
+    import ebooklib
+    from ebooklib import epub
+
+    book = epub.read_epub(str(file_path))
+    text_parts = []
+
+    for entry in book.spine:
+        item_id = entry[0] if isinstance(entry, (tuple, list)) else entry
+        try:
+            item = book.get_item_with_id(item_id)
+            if item and item.get_type() == ebooklib.ITEM_DOCUMENT:
+                content = item.get_body_content()
+                if content:
+                    if isinstance(content, bytes):
+                        content = content.decode("utf-8", errors="ignore")
+                    cleaned = clean_html(str(content))
+                    if cleaned.strip():
+                        text_parts.append(cleaned)
+        except Exception as exc:
+            logger.debug("Skipping EPUB spine item %r: %s", item_id, exc)
+
+    return "\n\n".join(text_parts)
+
+
+def _extract_epub_zipfile(file_path: Path) -> str:
+    """Extract by parsing HTML members of the EPUB zip directly."""
+    text_parts = []
+    with zipfile.ZipFile(file_path, "r") as epub_zip:
+        for file_name in sorted(epub_zip.namelist()):
+            if file_name.lower().endswith((".html", ".xhtml", ".htm")):
+                try:
+                    content = epub_zip.read(file_name).decode("utf-8", errors="ignore")
+                    cleaned = clean_html(content)
+                    if cleaned.strip():
+                        text_parts.append(cleaned)
+                except Exception as exc:
+                    logger.debug("Skipping EPUB member %r: %s", file_name, exc)
+    return "\n\n".join(text_parts)
+
+
+def _extract_epub_manual(file_path: Path) -> str:
+    """Last-resort extraction from any markup-looking EPUB member."""
+    skipped_extensions = (".jpg", ".jpeg", ".png", ".gif", ".css", ".js")
+    text_parts = []
+    with zipfile.ZipFile(file_path, "r") as epub_zip:
+        for file_name in sorted(epub_zip.namelist()):
+            if file_name.lower().endswith(skipped_extensions):
+                continue
+            try:
+                content = epub_zip.read(file_name).decode("utf-8", errors="ignore")
+                if "<" in content and len(content.strip()) > 100:
+                    cleaned = clean_html(content)
+                    if cleaned:
+                        text_parts.append(cleaned)
+            except Exception as exc:
+                logger.debug("Skipping EPUB member %r: %s", file_name, exc)
+    return "\n\n".join(text_parts)
+
+
+def _extract_txt(file_path: Path) -> str:
+    """Extract from TXT, trying common encodings (latin-1 is the catch-all)."""
+    for encoding in ("utf-8", "utf-16", "cp1252", "latin-1"):
+        try:
+            with open(file_path, "r", encoding=encoding) as f:
+                return clean_text(f.read())
+        except UnicodeError:
+            continue
+    raise ValueError(f"Could not decode text file: {file_path}")
+
+
+def _extract_pdf(file_path: Path) -> str:
+    """Extract from PDF."""
+    from pypdf import PdfReader
+
+    text = ""
+    with open(file_path, "rb") as file:
+        pdf_reader = PdfReader(file)
+        total_pages = len(pdf_reader.pages)
+        logger.info("PDF has %d pages", total_pages)
+
+        for page_num, page in enumerate(pdf_reader.pages, 1):
+            try:
+                page_text = page.extract_text() or ""
+                if page_text.strip():
+                    text += f"\n\n{page_text}"
+                if page_num % 10 == 0:
+                    logger.debug("Extracted %d/%d pages", page_num, total_pages)
+            except Exception as exc:
+                logger.warning("Failed to extract page %d: %s", page_num, exc)
+
+        logger.info("Extracted text from %d pages, %d characters total", total_pages, len(text))
+    return clean_text(text)