#!/usr/bin/env python3 import argparse import logging import os.path import re import zipfile from typing import Any, Dict, List, Optional, Tuple import epub_meta logging.basicConfig(format='%(levelname)s:%(funcName)s:%(message)s', level=logging.INFO) log = logging.getLogger('epubgrep') def get_chapter_title(mdata: List[Dict[str, Any]], fname: str) -> Optional[Tuple[str, int]]: found_list = [(x['title'], x['index']) for x in mdata if x['src'] == fname] log.debug('found_list = %s', found_list) if len(found_list) > 0: chap_title = found_list[0][0].strip(' \t.0123456789') return chap_title, found_list[0][1] else: return ('Unknown', 0) def grep_book(filename: str, pattern: str, flags: int): assert os.path.isfile(filename), "{} is not EPub file.".format(filename) sought_RE = re.compile(pattern, flags) mline = flags & re.M == re.M metadata = epub_meta.get_epub_metadata(filename) book = zipfile.ZipFile(filename) printed_booktitle = False for zif in book.infolist(): with book.open(zif) as inf: if mline: decoded_str = inf.read().decode(errors='replace') res = sought_RE.search(decoded_str) if res: if not printed_booktitle: print('{}'.format(filename)) printed_booktitle = True chap_info = get_chapter_title(metadata.toc, zif.filename) print("{}. {}:\n".format(chap_info[1], chap_info[0])) print('{}\n'.format(res.group(0))) else: printed_title = False for line in inf: decoded_line = line.decode(errors='replace').strip() if sought_RE.search(decoded_line): if not printed_booktitle: print('{}'.format(filename)) printed_booktitle = True if not printed_title: chap_info = get_chapter_title(metadata.toc, zif.filename) print("{}. {}:\n".format(chap_info[1], chap_info[0])) printed_title = True print(decoded_line) if __name__ == "__main__": parser = argparse.ArgumentParser(description='Grep through EPub book') parser.add_argument('pattern') parser.add_argument('filename') parser.add_argument('-i', '--ignore-case', action='store_true', help="make search case insensitive") parser.add_argument('-m', '--multi-line', action='store_true', help="make search multi line") args = parser.parse_args() log.debug('args = %s', args) search_flags = 0 if args.ignore_case: search_flags |= re.I if args.multi_line: search_flags |= re.M | re.S book_fname = os.path.realpath(args.filename) grep_book(book_fname, args.pattern, search_flags)