From dbc3dac871616c8549cc3ddce01cd0e98f83c1f6 Mon Sep 17 00:00:00 2001 From: Michael Williamson Date: Sat, 19 Nov 2016 20:02:19 +0000 Subject: Include embedded style map when explicit style map is set --- NEWS | 7 +++++++ README.md | 7 ++++++- mammoth/__init__.py | 11 ++++++++--- mammoth/options.py | 23 +++++++++++++++++------ tests/mammoth_tests.py | 16 ++++++++++++++++ 5 files changed, 54 insertions(+), 10 deletions(-) diff --git a/NEWS b/NEWS index 9105d08..fb917a0 100644 --- a/NEWS +++ b/NEWS @@ -1,3 +1,10 @@ +# 1.3.0 + +* Add include_embedded_style_map argument to allow embedded style maps to be + disabled. + +* Include embedded styles when explicit style map is passed. + # 1.2.2 * Ignore bold, italic, underline and strikethrough elements that have a value of diff --git a/README.md b/README.md index 428cf8a..b0e83e5 100644 --- a/README.md +++ b/README.md @@ -270,6 +270,11 @@ Converts the source document to HTML. * `style_map`: a string to specify the mapping of Word styles to HTML. See the section "Writing style maps" for a description of the syntax. +* `include_embedded_style_map`: by default, + if the document contains an embedded style, then it is combined with the default style map. + To ignore any embedded style maps, + pass `include_embedded_style_map=False`. + * `include_default_style_map`: by default, the style map passed in `style_map` is combined with the default style map. To stop using the default style map altogether, pass `include_default_style_map=False`. @@ -316,7 +321,7 @@ Each paragraph is followed by two newlines. Embeds the style map `style_map` into `fileobj`. When Mammoth reads a file object, -it will use the embedded style if no explicit style map is provided. +it will use the embedded style map. * `fileobj`: a file-like object containing the source document. Files should be opened for reading and writing in binary mode. diff --git a/mammoth/__init__.py b/mammoth/__init__.py index 7735fcf..fe480ec 100644 --- a/mammoth/__init__.py +++ b/mammoth/__init__.py @@ -5,6 +5,9 @@ from .docx.style_map import write_style_map, read_style_map __all__ = ["convert_to_html", "extract_raw_text", "images", "underline"] +_undefined = object() + + def convert_to_html(*args, **kwargs): return convert(*args, output_format="html", **kwargs) @@ -13,11 +16,13 @@ def convert_to_markdown(*args, **kwargs): return convert(*args, output_format="markdown", **kwargs) -def convert(fileobj, transform_document=None, id_prefix=None, **kwargs): +def convert(fileobj, transform_document=None, id_prefix=None, include_embedded_style_map=_undefined, **kwargs): + if include_embedded_style_map is _undefined: + include_embedded_style_map = True if transform_document is None: transform_document = lambda x: x - if kwargs.get("style_map") is None: - kwargs["style_map"] = read_style_map(fileobj) + if include_embedded_style_map: + kwargs["embedded_style_map"] = read_style_map(fileobj) return options.read_options(kwargs).bind(lambda convert_options: docx.read(fileobj).map(transform_document).bind(lambda document: conversion.convert_document_element_to_html( diff --git a/mammoth/options.py b/mammoth/options.py index ddac3de..d33b7ba 100644 --- a/mammoth/options.py +++ b/mammoth/options.py @@ -2,13 +2,24 @@ from . import style_reader, lists, results def read_options(options): - custom_style_map_result = _read_style_map(options.get("style_map") or "") - include_default_style_map = options.get("include_default_style_map", True) - options["ignore_empty_paragraphs"] = options.get("ignore_empty_paragraphs", True) + custom_style_map_text = options.pop("style_map", "") or "" + embedded_style_map_text = options.pop("embedded_style_map", "") or "" + include_default_style_map = options.pop("include_default_style_map", True) + + read_style_map_result = results.combine([ + _read_style_map(custom_style_map_text), + _read_style_map(embedded_style_map_text), + ]) + + custom_style_map, embedded_style_map = read_style_map_result.value + style_map = custom_style_map + embedded_style_map - options["style_map"] = custom_style_map_result.value + \ - (_default_style_map if include_default_style_map else []) - return custom_style_map_result.map(lambda _: options) + if include_default_style_map: + style_map += _default_style_map + + options["ignore_empty_paragraphs"] = options.get("ignore_empty_paragraphs", True) + options["style_map"] = style_map + return read_style_map_result.map(lambda _: options) def _read_style_map(style_text): diff --git a/tests/mammoth_tests.py b/tests/mammoth_tests.py index 4a524e8..322ebf7 100644 --- a/tests/mammoth_tests.py +++ b/tests/mammoth_tests.py @@ -63,6 +63,22 @@ def explicit_style_map_takes_precedence_over_embedded_style_map(): assert_equal([], result.messages) +@istest +def explicit_style_map_is_combined_with_embedded_style_map(): + with open(test_path("embedded-style-map.docx"), "rb") as fileobj: + result = mammoth.convert_to_html(fileobj=fileobj, style_map="r => strong") + assert_equal("

Walking on imported air

", result.value) + assert_equal([], result.messages) + + +@istest +def embedded_style_maps_can_be_disabled(): + with open(test_path("embedded-style-map.docx"), "rb") as fileobj: + result = mammoth.convert_to_html(fileobj=fileobj, include_embedded_style_map=False) + assert_equal("

Walking on imported air

", result.value) + assert_equal([], result.messages) + + @istest def embedded_style_map_can_be_written_and_then_read(): with _copy_of_test_data("single-paragraph.docx") as fileobj: -- cgit v1.2.3