0.14.4

CHANGES

@@ -1,3 +1,43 @@
+0.14.4 August 15, 2026
+- fixed a regression in text boilerplate stripping. Affected only txt -> html (and thus epub, mobi, etc)
+- fixed setup and added parse() invocation in tests_txt
+
+0.14.3 August 12, 2026
+
+This release includes several important updates. There are quite a few issues still waiting to be addressed; they are not forgotten.
+- To balance the long-standing policy of only using "REC" CSS against legal requirements for accessibility, we are significantly relaxing our CSS validation. This means that post-processors should take care to provide "graceful degradation" for CSS that might not be supported in every reading device.
+    - CSS Grid Layout Level 1 is now supported. Fixes #322. Grid layout should be strongly considered as a replacement for the use of table for layout
+    - CSS Flexbox Level 1 is now supported. Together with Grid, the should be no reason to use table for layout; going forward, layout tables will need to be converted to more accessible CSS based layout.
+    - CSS3 Text Decoration properties are now supported. These were the most often used properties removed by Ebookmaker as not yet REC. 
+    - The first-letter property is now supported. While this is still not supported in every browser, it is supported by most ebook readers. Use this property to make drop caps in an accessible way.  Fixes #310 
+    - Ebookmaker uses the Python module 'CSSUtils' to do CSS validation. It does not support the full grammar used in newer CSS. In particular it does not support the "var()" construct; do not expect this to change anytime soon.
+    - The current version of CSSUtils does not support the `rem` measurement, so CSS with rem in it will cause errors to be reported and CSS 2.1 rules containing `rem` will be removed. The current version of CSSUtils, that supports `rem`, requires a Python version newer than we have running in production. When our python version is updated, we will update CSSUtils.
+
+- In both in HTML and EPUB, for improved accessibility, added the CSS rule `a[href] {text-decoration: underline;}` to make sure links are underlined. Previously, we underlined all `a` elements, which ignored prior usage as an link target.
+
+- The `aside` element is now allowed. becomes div[class=aside] in epub2.
+
+- handling of inline SVG has been improved.
+    - the inline SVG element is now handled whether or not it has set the svg namespace.
+    - inline SVG is changed to standalone files in EPUB2
+    - inline html5 SVG now adapted for EPUB3, which cares about namespaces
+
+- an error is no longer emitted when rst output is requested from html input
+
+- uses hN element to set the title element of each chunk in the epub. fixes #331
+
+- refactored boilerplate stripping for txt files so it only needs to be done once per book. boilerplate is now detected in the text Parser, instead of in the text Writer.
+
+- support HTML5 `time` element. fixes #332 
+
+- `a` tags can't contain block tags in xhtml, so div in `a` and `p` in `a` cause EPUB2 validation errors, so Ebookmaker now changes `p` and `div` occurring in `a` to `span` for EPUB2. I'm not sure why anyone would want to use this markup in a book. Fixes #333
+
+- empty `th` elements are replaced by empty `td` elements for better accessibility. Fixes #315.
+
+
+0.14.2 July 4, 2026
+- `text-decoration` for `a` should be reset to `underline`, not initial. Fixes #316.
+
 0.14.1 June 12, 2026
 - fixed bug where text directly under body is deleted.
 - updated idna and beautifulsoup dependencies

setup.cfg

@@ -1,7 +1,7 @@
 [metadata]
 name = ebookmaker
 
-version = 0.14.1
+version = 0.14.4
 
 [options]
 package_dir=

setup.py

@@ -5,7 +5,7 @@
 
 from setuptools import setup
 
-VERSION = '0.14.1'
+VERSION = '0.14.4'
 
 if __name__ == "__main__":
 
@@ -42,7 +42,7 @@
             'roman',
             'requests',
             'six>=1.4.1',
-            'libgutenberg[covers]>=0.10.36',
+            'libgutenberg[covers]>=0.11.2',
             'cchardet==2.2.0a2',
             'beautifulsoup4',
             'html5lib',

src/ebookmaker/HTMLChunker.py

@@ -165,6 +165,9 @@ def shipout_chunk(self, attribs, chunk_id = None, comment = None):
         for e in xpath(self.chunk, '//mathml:math'):
             attribs.rel.add('mathml')
             break
+        for e in xpath(self.chunk, '//xhtml:svg'):
+            attribs.rel.add('svg')
+            break
         for e in xpath(self.chunk, '//svg:svg'):
             attribs.rel.add('svg')
             break
@@ -305,3 +308,23 @@ def rewrite_internal_links_toc(self, toc):
                 error("HTMLChunker: Cannot rewrite toc entry '%s'" % entry[0])
                 error(repr(self.idmap))
                 del entry
+
+    def set_running_headers(self):
+        """ Use hN elements in chunks to set the chunk title
+
+        """
+        chunk_title = ''
+        for chunk in self.chunks:
+            for tag in ['h1', 'h2', 'h3', 'h4', 'h5', 'h6']:
+                hN_title = ''
+                for hN in xpath(chunk[0], f"//xhtml:{tag}"):
+                    hN_title = hN.text_content()
+                    if hN_title:
+                         break
+                if hN_title:
+                    chunk_title = hN_title
+                    break
+            if chunk_title != '':
+                for title in xpath(chunk[0], f"//xhtml:title"):
+                    title.text = chunk_title
+                    break

src/ebookmaker/ParserFactory.py

@@ -86,7 +86,7 @@ def create(cls, url, attribs=None):
         if attribs is None:
             attribs = parsers.ParserAttributes()
 
-        # debug("Need parser for %s" % url)
+        debug("Need parser for %s" % url)
 
         # first check if input url is in output directory (we've already made it!)
         if gg.is_same_path(os.path.abspath(options.outputdir), os.path.dirname(url)):
@@ -101,7 +101,7 @@ def create(cls, url, attribs=None):
 
 
         if url in cls.parsers:
-            # debug("... reusing parser for %s" % url)
+            debug("... reusing parser for %s" % url)
             # reuse same parser, maybe already filled with data
             parser = cls.parsers[url]
             parser.reset()

src/ebookmaker/Version.py

@@ -1,2 +1,2 @@
-VERSION = '0.14.1'
+VERSION = '0.14.4'
 GENERATOR = 'Ebookmaker %s by Project Gutenberg'

src/ebookmaker/parsers/CSSParser.py

@@ -28,10 +28,146 @@
 RE_ELEMENT = re.compile(r'\[[^\]]*\]|((?:^|\s|\+|>|~|,)[a-z0-9]+)', re.I)
 
 mediatypes = (mt.css, )
-PG_CSS_PROFILE = (
-    'Added Properties for Project Gutenberg',
-    {
-        'display': 'flex|initial',
+
+""" Flex and Grid layouts; generated by Claude
+* The grid track-list grammar (repeat(), minmax(), fit-content(),
+  named lines, etc.) is approximated with regular expressions. It
+  covers the vast majority of real-world usage but is not a perfect
+  transcription of the CSS Grid formal grammar (which isn't regular).
+  Tighten or loosen the macros below if you hit edge cases.
+* Known limitation: named grid lines written with square brackets,
+  e.g. `grid-template-columns: [full-start] 1fr [full-end];`, will
+  fail to parse at all -- not just fail validation. This is a
+  limitation of cssutils' own CSS tokenizer/parser (it doesn't know
+  `[` `]` can appear inside a property value), which sits below the
+  profile system and can't be fixed by adding a profile. Everything
+  else in this file (line-names macro, etc.) is kept for forward
+  compatibility / partial matches and in case you patch the tokenizer
+  yourself, but square-bracket named lines won't round-trip through
+  stock cssutils today.
+"""
+
+
+_MACROS = {
+    # basic tokens (defined locally so this profile doesn't depend on
+    # exactly which internal macro names cssutils happens to expose)
+    'integer': r'[+-]?\d+',
+    'num': r'[+-]?\d*\.?\d+(?:[eE][+-]?\d+)?',
+    'length': r'(?:{num}(?:em|ex|px|in|cm|mm|pt|pc|ch|rem|vh|vw|vmin|vmax|q)|0)',
+    'percentage': r'{num}%',
+    'ident': r'-?[a-zA-Z_][a-zA-Z0-9_-]*',
+    'string': r'''(?:"[^"]*"|'[^']*')''',
+    'lop': r'(?:{length}|{percentage})',           # <length-percentage>
+    'lopa': r'(?:{length}|{percentage}|auto)',     # <length-percentage> | auto
+    'positive-integer': r'\d+',
+
+    # --- Grid: track sizing ---
+    'flex-value': r'{num}fr',
+    'inflexible-breadth': r'(?:{length}|{percentage}|min-content|max-content|auto)',
+    'track-breadth': r'(?:{length}|{percentage}|min-content|max-content|auto|{flex-value})',
+    'track-size': r'(?:{track-breadth}|minmax\(\s*{inflexible-breadth}\s*,\s*{track-breadth}\s*\)|fit-content\(\s*{lop}\s*\))',
+    'line-names': r'\[\s*{ident}(?:\s+{ident})*\s*\]',
+    'track-repeat': r'repeat\(\s*(?:{integer}|auto-fill|auto-fit)\s*,\s*(?:\s*{line-names}?\s*{track-size}\s*)+\s*{line-names}?\s*\)',
+    'track-list': r'(?:\s*{line-names}?\s*(?:{track-size}|{track-repeat})\s*)+\s*{line-names}?',
+    'auto-track-list': r'(?:\s*{line-names}?\s*{track-size}\s*)*\s*{line-names}?\s*repeat\(\s*(?:auto-fill|auto-fit)\s*,\s*(?:\s*{line-names}?\s*{track-size}\s*)+\s*{line-names}?\s*\)\s*(?:\s*{line-names}?\s*{track-size}\s*)*\s*{line-names}?',
+    'grid-line': r'(?:auto|span\s+(?:{integer}\s+{ident}|{ident}\s+{integer}|{integer}|{ident})|{integer}\s+{ident}|{ident}\s+{integer}|{integer}|{ident})',
+
+    # --- Shared box-alignment values (used by both grid and flexbox) ---
+    'content-position': r'(?:center|start|end|flex-start|flex-end|left|right)',
+    'self-position': r'(?:center|start|end|flex-start|flex-end|self-start|self-end)',
+    'overflow-position': r'(?:safe|unsafe)',
+    'baseline-position': r'(?:(?:first|last)\s+)?baseline',
+
+   # simplified <color>: named/keyword ident, #hex, rgb()/rgba()/hsl()/hsla()
+    'hexcolor': r'\#(?:[0-9a-fA-F]{3}|[0-9a-fA-F]{4}|[0-9a-fA-F]{6}|[0-9a-fA-F]{8})',
+    'colorfunc': r'(?:rgb|rgba|hsl|hsla)\([^)]*\)',
+    'namedcolor': r'(?:currentColor|transparent|{ident})',
+    'color': r'(?:{hexcolor}|{colorfunc}|{namedcolor})',
+
+    # text-shadow: <color>? && <length>{2,3} (offsets/blur, in either order relative to color)
+    'shadow-item': r'(?:{color}\s+)?{length}\s+{length}(?:\s+{length})?(?:\s+{color})?',
+
+    # from PG_CSS_PROFILE
+    'numeric-figure-values': 'lining-nums|oldstyle-nums',
+    'numeric-spacing-values': 'proportional-nums|tabular-nums',
+    'numeric-fraction-values': 'diagonal-fractions|stacked-fractions',
+    'font-variant-attrs': '{numeric-figure-values}|{numeric-spacing-values}|{numeric-fraction-values}|ordinal|slashed-zero',
+
+}
+
+# ---------------------------------------------------------------------------
+# CSS Grid Layout Module Level 1
+# https://www.w3.org/TR/css-grid-1/
+# ---------------------------------------------------------------------------
+
+GRID_PROPERTIES = {
+    'display': r'(?:grid|inline-grid)',
+
+    'grid-template-columns': r'(?:none|subgrid|{track-list}|{auto-track-list})',
+    'grid-template-rows': r'(?:none|subgrid|{track-list}|{auto-track-list})',
+    'grid-template-areas': r'(?:none|(?:{string}\s*)+)',
+    'grid-template': r'(?:none|.+)',  # shorthand: permissive (complex grammar)
+    'grid': r'.+',                    # shorthand: permissive (complex grammar)
+
+    'grid-auto-columns': r'(?:{track-size}\s*)+',
+    'grid-auto-rows': r'(?:{track-size}\s*)+',
+    'grid-auto-flow': r'(?:(?:row|column)(?:\s+dense)?|dense)',
+
+    'grid-row-start': r'{grid-line}',
+    'grid-row-end': r'{grid-line}',
+    'grid-column-start': r'{grid-line}',
+    'grid-column-end': r'{grid-line}',
+    'grid-row': r'{grid-line}(?:\s*/\s*{grid-line})?',
+    'grid-column': r'{grid-line}(?:\s*/\s*{grid-line})?',
+    'grid-area': r'{grid-line}(?:\s*/\s*{grid-line}){0,3}',
+
+    # gap properties (introduced alongside Grid Level 1, later folded
+    # into Box Alignment Level 3 -- both spellings are included)
+    'grid-row-gap': r'{lop}',
+    'grid-column-gap': r'{lop}',
+    'grid-gap': r'{lop}(?:\s+{lop})?',
+    'row-gap': r'{lop}',
+    'column-gap': r'{lop}',
+    'gap': r'{lop}(?:\s+{lop})?',
+
+    'justify-items': r'(?:normal|stretch|{baseline-position}|{overflow-position}?\s*{self-position}|legacy(?:\s+(?:left|right|center))?)',
+    'justify-self': r'(?:auto|normal|stretch|{baseline-position}|{overflow-position}?\s*{self-position})',
+}
+
+# ---------------------------------------------------------------------------
+# CSS Flexible Box Layout Module Level 1
+# https://www.w3.org/TR/css-flexbox-1/
+# ---------------------------------------------------------------------------
+
+FLEXBOX_PROPERTIES = {
+    'display': r'(?:flex|inline-flex)',
+
+    'flex-direction': r'(?:row|row-reverse|column|column-reverse)',
+    'flex-wrap': r'(?:nowrap|wrap|wrap-reverse)',
+    'flex-flow': r'(?:(?:row|row-reverse|column|column-reverse)(?:\s+(?:nowrap|wrap|wrap-reverse))?|(?:nowrap|wrap|wrap-reverse)(?:\s+(?:row|row-reverse|column|column-reverse))?)',
+
+    'order': r'{integer}',
+    'flex-grow': r'{num}',
+    'flex-shrink': r'{num}',
+    'flex-basis': r'(?:content|{lopa})',
+    'flex': r'(?:none|(?:auto|{num})(?:\s+{num})?(?:\s+{lopa})?)',
+}
+
+# ---------------------------------------------------------------------------
+# Shared CSS Box Alignment values used by flex/grid containers and items
+# (kept as a separate profile so it isn't tied to grid- or flex-only display)
+# ---------------------------------------------------------------------------
+
+BOX_ALIGNMENT_PROPERTIES = {
+    'align-items': r'(?:normal|stretch|{baseline-position}|{overflow-position}?\s*{self-position})',
+    'align-self': r'(?:auto|normal|stretch|{baseline-position}|{overflow-position}?\s*{self-position})',
+    'align-content': r'(?:normal|{baseline-position}|{overflow-position}?\s*{content-position}|space-between|space-around|space-evenly|stretch)',
+    'justify-content': r'(?:normal|{overflow-position}?\s*{content-position}|space-between|space-around|space-evenly|stretch)',
+}
+
+
+PG_CSS_PROFILE = {
+        'display': 'grid|inline-grid|flex|inline-flex|block|inline|none',
         'justify-content': 'center',
         'speak': r'auto|never|always',
         'speak-as': 'normal|spell-out|digits|literal-punctuation|no-punctuation',
@@ -60,26 +196,58 @@
         'padding-left': 'initial',
         'padding': 'initial',
         'text-align': 'initial',
-        'text-decoration': 'initial',
         'text-indent': 'initial',
         'text-transform': 'initial',
 
-
         # updated for  https://www.w3.org/TR/css-writing-modes-3/
         # direction and unicode-bidi  are not supported based on the standard's recommendation
         'writing-mode': 'vertical-lr|vertical-rl|horizontal-tb',
         'text-orientation': 'mixed|upright|sideways',
         'text-combine-upright': 'none | all',      
-    },
-    {
-        'numeric-figure-values': 'lining-nums|oldstyle-nums',
-        'numeric-spacing-values': 'proportional-nums|tabular-nums',
-        'numeric-fraction-values': 'diagonal-fractions|stacked-fractions',
-        'font-variant-attrs': '{numeric-figure-values}|{numeric-spacing-values}|{numeric-fraction-values}|ordinal|slashed-zero',
-    }
-)
-
-cssutils.profile.addProfiles([PG_CSS_PROFILE])
+
+        # add css3 decoration properties
+        'text-decoration-line': r'none|underline|overline|line-through',
+        'text-decoration-style': r'solid|double|dotted|dashed|wavy',
+        'text-decoration-color': r'{color}',
+        'text-decoration': r'(?:none|(?:(?:underline|overline|line-through|blink)(?:\s+(?:underline|overline|line-through|blink)){0,3}|solid|double|dotted|dashed|wavy|{color})(?:\s+(?:(?:underline|overline|line-through|blink)(?:\s+(?:underline|overline|line-through|blink)){0,3}|solid|double|dotted|dashed|wavy|{color})){0,2})',
+
+        # shorthand: <text-emphasis-style> || <text-emphasis-color>
+        'text-emphasis': r'(?:none|(?:{string}|(?:filled|open)(?:\s+(?:dot|circle|double-circle|triangle|sesame))?|(?:dot|circle|double-circle|triangle|sesame)(?:\s+(?:filled|open))?|{color})(?:\s+(?:{string}|(?:filled|open)(?:\s+(?:dot|circle|double-circle|triangle|sesame))?|(?:dot|circle|double-circle|triangle|sesame)(?:\s+(?:filled|open))?|{color}))?)',
+        'text-emphasis-style': r'(?:none|{string}|(?:filled|open)(?:\s+(?:dot|circle|double-circle|triangle|sesame))?|(?:dot|circle|double-circle|triangle|sesame)(?:\s+(?:filled|open))?)',
+        'text-emphasis-color': r'{color}',
+
+        # text-underline-position: auto | [ under || [ left | right ] ]
+        'text-underline-position': r'(?:auto|under(?:\s+(?:left|right))?|(?:left|right)(?:\s+under)?)',
+
+        # text-emphasis-position: [ over | under ] && [ right | left ]
+        'text-emphasis-position': r'(?:(?:over|under)\s+(?:right|left)|(?:right|left)\s+(?:over|under))',
+
+        # text-shadow: none | <shadow>#
+        'text-shadow': r'(?:none|{shadow-item}(?:\s*,\s*{shadow-item})*)',
+
+        # initial letter for accessible drop caps
+        # <number> = the sink's size, in number of lines; the optional
+        # <integer> = the sink, how many lines it descends (defaults to
+        # round(<number>) when omitted)
+        'initial-letter': r'(?:normal|{num}(?:\s+{integer})?)',
+
+        # initial-letter-align: [ auto | alphabetic | hanging | ideographic ]
+        'initial-letter-align': r'(?:auto|alphabetic|hanging|ideographic)',
+
+        # initial-letter-wrap: none | first | all | grid | <length-percentage>
+        'initial-letter-wrap': r'(?:none|first|all|grid|{lop})',
+
+
+}
+
+# ---------------------------------------------------------------------------
+# Register the profiles with cssutils
+# ---------------------------------------------------------------------------
+
+cssutils.profile.addProfile('CSS Grid Layout Module Level 1', GRID_PROPERTIES, _MACROS)
+cssutils.profile.addProfile('CSS Flexible Box Layout Module Level 1', FLEXBOX_PROPERTIES, _MACROS)
+cssutils.profile.addProfile('CSS Box Alignment (Flexbox/Grid)', BOX_ALIGNMENT_PROPERTIES, _MACROS)
+cssutils.profile.addProfile('Added Properties for Project Gutenberg', PG_CSS_PROFILE, _MACROS)
 
 class Parser(ParserBase):
     """ Parse an external CSS file. """