txt2ebook 0.1.112__py3-none-any.whl → 0.1.114__py3-none-any.whl

txt2ebook/__init__.py

@@ -20,9 +20,11 @@ import logging
 import platform
 import sys
 
+import langdetect
+
 logger = logging.getLogger(__name__)
 
-__version__ = "0.1.112"
+__version__ = "0.1.114"
 
 
 def setup_logger(config: argparse.Namespace) -> None:
@@ -69,3 +71,19 @@ def print_env() -> None:
         f"platform: {platform.platform()}",
         sep="\n",
     )
+
+
+def detect_and_expect_language(content: str, config_language: str) -> str:
+    """Detect and expect the language of the txt content."""
+    detect_language = langdetect.detect(content)
+    config_language = config_language or detect_language
+    logger.info("Config language: %s", config_language)
+    logger.info("Detect language: %s", detect_language)
+
+    if config_language and config_language != detect_language:
+        logger.warning(
+            "Config (%s) and detect (%s) language mismatch",
+            config_language,
+            detect_language,
+        )
+    return config_language

txt2ebook/cli.py

@@ -27,10 +27,13 @@ from typing import Optional, Sequence
 
 import txt2ebook.subcommands.env
 import txt2ebook.subcommands.epub
+import txt2ebook.subcommands.gmi
 import txt2ebook.subcommands.massage
 import txt2ebook.subcommands.md
 import txt2ebook.subcommands.parse
+import txt2ebook.subcommands.pdf
 import txt2ebook.subcommands.tex
+import txt2ebook.subcommands.typ
 from txt2ebook import __version__, setup_logger
 
 logger = logging.getLogger(__name__)
@@ -287,7 +290,10 @@ def build_parser(
     txt2ebook.subcommands.massage.build_subparser(subparsers)
     txt2ebook.subcommands.epub.build_subparser(subparsers)
     txt2ebook.subcommands.tex.build_subparser(subparsers)
+    txt2ebook.subcommands.pdf.build_subparser(subparsers)
+    txt2ebook.subcommands.typ.build_subparser(subparsers)
     txt2ebook.subcommands.md.build_subparser(subparsers)
+    txt2ebook.subcommands.gmi.build_subparser(subparsers)
     txt2ebook.subcommands.env.build_subparser(subparsers)
 
     return parser

txt2ebook/subcommands/gmi.py

@@ -0,0 +1,87 @@
+# Copyright (C) 2021,2022,2023,2024 Kian-Meng Ang
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+"""gmi subcommand."""
+
+import argparse
+import logging
+import sys
+
+from txt2ebook.subcommands.parse import run as parse_txt
+from txt2ebook.formats.gmi import GmiWriter
+
+
+logger = logging.getLogger(__name__)
+
+
+def build_subparser(subparsers) -> None:
+    """Build the subparser."""
+    gmi_parser = subparsers.add_parser(
+        "gmi", help="generate ebook in Markdown format"
+    )
+
+    gmi_parser.set_defaults(func=run)
+
+    gmi_parser.add_argument(
+        "input_file",
+        nargs=None if sys.stdin.isatty() else "?",  # type: ignore
+        type=argparse.FileType("rb"),
+        default=None if sys.stdin.isatty() else sys.stdin,
+        help="source text filename",
+        metavar="TXT_FILENAME",
+    )
+
+    gmi_parser.add_argument(
+        "output_file",
+        nargs="?",
+        default=None,
+        help="converted ebook filename (default: 'TXT_FILENAME.md')",
+        metavar="EBOOK_FILENAME",
+    )
+
+    gmi_parser.add_argument(
+        "-sp",
+        "--split-volume-and-chapter",
+        default=False,
+        action="store_true",
+        dest="split_volume_and_chapter",
+        help=(
+            "split volume or chapter into separate file and "
+            "ignore the --overwrite option"
+        ),
+    )
+
+    gmi_parser.add_argument(
+        "-toc",
+        "--table-of-content",
+        default=False,
+        action="store_true",
+        dest="with_toc",
+        help="add table of content",
+    )
+
+
+def run(args: argparse.Namespace) -> None:
+    """Run md subcommand.
+
+    Args:
+        config (argparse.Namespace): Config from command line arguments
+
+    Returns:
+        None
+    """
+    book = parse_txt(args)
+    writer = GmiWriter(book, args)
+    writer.write()

txt2ebook/subcommands/massage.py

@@ -23,7 +23,10 @@ import cjkwrap
 import regex as re
 from bs4 import UnicodeDammit
 
+from txt2ebook import detect_and_expect_language
 from txt2ebook.exceptions import EmptyFileError
+from txt2ebook.formats.txt import TxtWriter
+from txt2ebook.parser import Parser
 from txt2ebook.zh_utils import zh_halfwidth_to_fullwidth
 
 logger = logging.getLogger(__name__)
@@ -44,6 +47,35 @@ def build_subparser(subparsers) -> None:
         metavar="TXT_FILENAME",
     )
 
+    massage_parser.add_argument(
+        "output_file",
+        nargs="?",
+        default=None,
+        help="converted ebook filename (default: 'TXT_FILENAME.txt')",
+        metavar="EBOOK_FILENAME",
+    )
+
+    massage_parser.add_argument(
+        "-sp",
+        "--split-volume-and-chapter",
+        default=False,
+        action="store_true",
+        dest="split_volume_and_chapter",
+        help=(
+            "split volume or chapter into separate file and "
+            "ignore the --overwrite option"
+        ),
+    )
+
+    massage_parser.add_argument(
+        "-ow",
+        "--overwrite",
+        default=False,
+        action="store_true",
+        dest="overwrite",
+        help="overwrite massaged TXT_FILENAME",
+    )
+
     massage_parser.add_argument(
         "-rd",
         "--regex-delete",
@@ -97,6 +129,21 @@ def run(args: argparse.Namespace) -> None:
     Returns:
         None
     """
+    massaged_txt = massage_txt(args)
+    args.language = detect_and_expect_language(massaged_txt, args.language)
+    args.with_toc = False
+    parser = Parser(massaged_txt, args)
+    book = parser.parse()
+
+    if args.debug:
+        book.debug(args.verbose)
+
+    writer = TxtWriter(book, args)
+    writer.write()
+
+
+def massage_txt(args: argparse.Namespace) -> str:
+    """Massage the text file."""
     logger.info("Parsing txt file: %s", args.input_file.name)
 
     unicode = UnicodeDammit(args.input_file.read())

txt2ebook/subcommands/parse.py

@@ -19,6 +19,7 @@ import argparse
 import logging
 import sys
 
+import jieba.analyse
 from bs4 import UnicodeDammit
 from langdetect import detect
 
@@ -80,6 +81,9 @@ def run(args: argparse.Namespace) -> Book:
             detect_language,
         )
 
+    tags = jieba.analyse.extract_tags(content, topK=100)
+    logger.info("tags: %s", " ".join(tags))
+
     parser = Parser(content, args)
     book = parser.parse()
 

txt2ebook/subcommands/pdf.py

@@ -0,0 +1,77 @@
+# Copyright (C) 2021,2022,2023,2024 Kian-Meng Ang
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+"""pdf subcommand."""
+
+import argparse
+import logging
+import sys
+
+from txt2ebook.subcommands.parse import run as parse_txt
+from txt2ebook.formats.pdf import PdfWriter
+from txt2ebook.formats import PAGE_SIZES
+
+
+logger = logging.getLogger(__name__)
+
+
+def build_subparser(subparsers) -> None:
+    """Build the subparser."""
+    pdf_parser = subparsers.add_parser(
+        "pdf", help="generate ebook in Markdown format"
+    )
+
+    pdf_parser.set_defaults(func=run)
+
+    pdf_parser.add_argument(
+        "input_file",
+        nargs=None if sys.stdin.isatty() else "?",  # type: ignore
+        type=argparse.FileType("rb"),
+        default=None if sys.stdin.isatty() else sys.stdin,
+        help="source text filename",
+        metavar="TXT_FILENAME",
+    )
+
+    pdf_parser.add_argument(
+        "output_file",
+        nargs="?",
+        default=None,
+        help="converted ebook filename (default: 'TXT_FILENAME.md')",
+        metavar="EBOOK_FILENAME",
+    )
+
+    pdf_parser.add_argument(
+        "-pz",
+        "--page-size",
+        dest="page_size",
+        default="a5",
+        choices=PAGE_SIZES,
+        help="page size of the ebook (default: '%(default)s')",
+        metavar="PAGE_SIZE",
+    )
+
+
+def run(args: argparse.Namespace) -> None:
+    """Run md subcommand.
+
+    Args:
+        config (argparse.Namespace): Config from command line arguments
+
+    Returns:
+        None
+    """
+    book = parse_txt(args)
+    writer = PdfWriter(book, args)
+    writer.write()

txt2ebook/subcommands/typ.py

@@ -0,0 +1,86 @@
+# Copyright (C) 2021,2022,2023,2024 Kian-Meng Ang
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+"""md subcommand."""
+
+import argparse
+import logging
+import sys
+
+from txt2ebook.subcommands.parse import run as parse_txt
+from txt2ebook.formats.typ import TypWriter
+from txt2ebook.formats import PAGE_SIZES
+
+
+logger = logging.getLogger(__name__)
+
+
+def build_subparser(subparsers) -> None:
+    """Build the subparser."""
+    typ_parser = subparsers.add_parser(
+        "typ", help="generate ebook in Typst format"
+    )
+
+    typ_parser.set_defaults(func=run)
+
+    typ_parser.add_argument(
+        "input_file",
+        nargs=None if sys.stdin.isatty() else "?",  # type: ignore
+        type=argparse.FileType("rb"),
+        default=None if sys.stdin.isatty() else sys.stdin,
+        help="source text filename",
+        metavar="TXT_FILENAME",
+    )
+
+    typ_parser.add_argument(
+        "output_file",
+        nargs="?",
+        default=None,
+        help="converted ebook filename (default: 'TXT_FILENAME.typ')",
+        metavar="EBOOK_FILENAME",
+    )
+
+    typ_parser.add_argument(
+        "-pz",
+        "--page-size",
+        dest="page_size",
+        default="a5",
+        choices=PAGE_SIZES,
+        help="page size of the ebook (default: '%(default)s')",
+        metavar="PAGE_SIZE",
+    )
+
+    typ_parser.add_argument(
+        "-toc",
+        "--table-of-content",
+        default=False,
+        action="store_true",
+        dest="with_toc",
+        help="add table of content",
+    )
+
+
+def run(args: argparse.Namespace) -> None:
+    """Run typ subcommand.
+
+    Args:
+        config (argparse.Namespace): Config from command line arguments
+
+    Returns:
+        None
+    """
+    book = parse_txt(args)
+    writer = TypWriter(book, args)
+    writer.write()

txt2ebook/tokenizer.py

@@ -62,11 +62,7 @@ class Tokenizer:
         """Set the constructor for the Tokenizer."""
         self.raw_content = raw_content
         self.config = config
-
-        if self.config.fullwidth:
-            self.metadata_marker = "－－－"
-        else:
-            self.metadata_marker = "---"
+        self.metadata_marker = "---"
 
         config_lang = config.language.replace("-", "_")
         self.langconf = import_module(f"txt2ebook.languages.{config_lang}")

txt2ebook/txt2ebook.py

@@ -29,9 +29,8 @@ import time
 from typing import Optional, Sequence
 
 from bs4 import UnicodeDammit
-from langdetect import detect
 
-from txt2ebook import __version__, print_env, setup_logger
+from txt2ebook import __version__, print_env, setup_logger, detect_and_expect_language
 from txt2ebook.exceptions import EmptyFileError
 from txt2ebook.formats import (
     EBOOK_FORMATS,
@@ -62,19 +61,7 @@ def run(config: argparse.Namespace) -> None:
                 f"Empty file content in {config.input_file.name}"
             )
 
-        config_language = config.language
-        detect_language = detect(content)
-        config.language = config_language or detect_language
-        logger.info("Config language: %s", config_language)
-        logger.info("Detect language: %s", detect_language)
-
-        if config_language and config_language != detect_language:
-            logger.warning(
-                "Config (%s) and detect (%s) language mismatch",
-                config_language,
-                detect_language,
-            )
-
+        config.language = detect_and_expect_language(content, config.language)
         parser = Parser(content, config)
         book = parser.parse()
 

{txt2ebook-0.1.112.dist-info → txt2ebook-0.1.114.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: txt2ebook
-Version: 0.1.112
+Version: 0.1.114
 Summary: CLI tool to convert txt file to ebook format
 Home-page: https://github.com/kianmeng/txt2ebook
 License: AGPL-3.0-or-later
@@ -29,6 +29,7 @@ Requires-Dist: CJKwrap (>=2.2,<3.0)
 Requires-Dist: EbookLib (>=0.17.1,<0.18.0)
 Requires-Dist: bs4 (>=0.0.1,<0.0.2)
 Requires-Dist: importlib-resources (>=6.1.1,<7.0.0)
+Requires-Dist: jieba (>=0.42.1,<0.43.0)
 Requires-Dist: langdetect (>=1.0.9,<2.0.0)
 Requires-Dist: lxml (>=5.2.2,<6.0.0)
 Requires-Dist: pylatex (>=1.4.2,<2.0.0)

{txt2ebook-0.1.112.dist-info → txt2ebook-0.1.114.dist-info}/RECORD

@@ -1,6 +1,6 @@
-txt2ebook/__init__.py,sha256=aPLmaeNFAoTluwaX47RB9YvnRDaSSl-fPG4ezGS2ckY,2062
+txt2ebook/__init__.py,sha256=oOYsrTNvJqaCLR6K9pfOTy12ZuXFXMtFB6QLyAdaMkI,2681
 txt2ebook/__main__.py,sha256=gMLvgpqc_BL4cBqNe0vqErRF5dlJPAbvqu1zndcAHYI,850
-txt2ebook/cli.py,sha256=LLwnhIqirbpyQVfeHbmJYysSU9NRMHusFVe5bYQ2_Ws,8362
+txt2ebook/cli.py,sha256=qdQQ9r134Lbs58NwzJ1wKDF1EO6blWLBzqjqA_tpZIc,8635
 txt2ebook/exceptions.py,sha256=b2HDsXdqweLJbvSJEGt48nxvGkZq20SfYezSjwp77JU,842
 txt2ebook/formats/__init__.py,sha256=WhiRWGvbUjc8QZfhAIkKCg6GL8vNNlEF73meZSzYhDA,2463
 txt2ebook/formats/base.py,sha256=SMt6Op88-HoIxRA-tgPBNSlrt7-KZ-m5-BytAEJT4m0,5814
@@ -35,15 +35,18 @@ txt2ebook/models/volume.py,sha256=HyT4XO9yZ8d0PgZVfMMyAYUDFv58RrUmsSFNNmU-sHY,15
 txt2ebook/parser.py,sha256=2Dk1n51Czb3Tn_a1tgSSx5A-XNETVyuX4gJ9nQ-fPlQ,8707
 txt2ebook/subcommands/env.py,sha256=26wcAhEbjKfWYeNiIsjm9TTaBUyNDYx1MYUtV2ZFkmg,1481
 txt2ebook/subcommands/epub.py,sha256=evZgvi1cHe3uCHgAwteJ2FlXFk3YOUekYNghVk7ySgE,2578
-txt2ebook/subcommands/massage.py,sha256=on4Oqpv3WwOf6tZVGFzwq972oGa6TaJx7Is_0YSnmBI,5766
+txt2ebook/subcommands/gmi.py,sha256=k-pDtP2OevgMACy6VeUSfiKLV0E4FyPA9t4wGLjJXu8,2405
+txt2ebook/subcommands/massage.py,sha256=wsSFVdheYgCn9NsISsgHVyoRj0D73GfP3NIrTYI13iA,7024
 txt2ebook/subcommands/md.py,sha256=v-xSWEpe6p9hIDRh-Nk1uH43P9gOvy80g-qt11dG-aw,2418
-txt2ebook/subcommands/parse.py,sha256=1QEVRzgDZYiwOs-36WYPNFIIc-GgKCp8eBxZKuzqQa8,2607
+txt2ebook/subcommands/parse.py,sha256=BpEHs-gFYbJK1-50FPUX_3INfuxkgL5Glgpd3IBdJfs,2730
+txt2ebook/subcommands/pdf.py,sha256=cnmjavFg6aXSYWzXpQEH4QfiwkppzsbAsLdO3DNvcRo,2179
 txt2ebook/subcommands/tex.py,sha256=O2Sl-v8tTU_bMFDDukFYEyXBTecGzp3Ll8Gqom63Clk,2330
-txt2ebook/tokenizer.py,sha256=HJEub1NYTaH3FMfzzdaNpmimzgzGN942m9O7oDHqkuQ,9456
-txt2ebook/txt2ebook.py,sha256=GwWjSk2HbnGrOM59qFsPv2CMqAgv1GBxy_r1zkgBEvQ,13969
+txt2ebook/subcommands/typ.py,sha256=bS6AVcZk4mxcHnvlaCPWaIIUWVRmOsCXNHecWDH4nJg,2373
+txt2ebook/tokenizer.py,sha256=zdZwiDreL1QVHdWm4YGPqG8-HzcjPI6wS9vggBlNEkg,9356
+txt2ebook/txt2ebook.py,sha256=ZP-5RxZcKasoowjuJjUR1yEBzHqekOBPruGvyj47sKg,13539
 txt2ebook/zh_utils.py,sha256=EgKVbwqYGaTGswQUGcOCeSfRelzwkAb9WWY9TrsX1x4,4882
-txt2ebook-0.1.112.dist-info/LICENSE.md,sha256=tGtFDwxWTjuR9syrJoSv1Hiffd2u8Tu8cYClfrXS_YU,31956
-txt2ebook-0.1.112.dist-info/METADATA,sha256=amyhdLNk3kzqJy5wis5MAoOYUo4u1CxvHtoauC2-J60,7754
-txt2ebook-0.1.112.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-txt2ebook-0.1.112.dist-info/entry_points.txt,sha256=JLW3Iv7eUyABlQeUFiUWQhLKfRdnB9o5SIcNlneGR0Q,77
-txt2ebook-0.1.112.dist-info/RECORD,,
+txt2ebook-0.1.114.dist-info/LICENSE.md,sha256=tGtFDwxWTjuR9syrJoSv1Hiffd2u8Tu8cYClfrXS_YU,31956
+txt2ebook-0.1.114.dist-info/METADATA,sha256=6CM7gs_KSb6hRpSj9EnK4w1tdYloa050zZnS_If9ycI,7794
+txt2ebook-0.1.114.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+txt2ebook-0.1.114.dist-info/entry_points.txt,sha256=JLW3Iv7eUyABlQeUFiUWQhLKfRdnB9o5SIcNlneGR0Q,77
+txt2ebook-0.1.114.dist-info/RECORD,,