fix: preserve nested Word structures and synchronize valid font families

This commit is contained in:
2026-11-05 21:54:52 +08:00
parent 0e072eed3b
commit 7f81c2015e
12 changed files with 292 additions and 28 deletions
+5 -3
View File
@@ -113,9 +113,11 @@ def main(argv: list[str] | None = None) -> int:
app.setOrganizationName("MDHelper") app.setOrganizationName("MDHelper")
settings = QSettings("MDHelper", "MDHelper") settings = QSettings("MDHelper", "MDHelper")
if project_root is None: if project_root is None:
last_project_root = Path(str(settings.value("project/last_root", ""))).expanduser() last_root = settings.value("project/last_root", "")
if str(last_project_root) and last_project_root.exists(): if isinstance(last_root, str) and last_root.strip():
project_root = last_project_root.resolve() last_project_root = Path(last_root).expanduser()
if last_project_root.is_dir():
project_root = last_project_root.resolve()
logging.getLogger(__name__).info( logging.getLogger(__name__).info(
"Starting MDHelper %s with project_root=%s content_root=%s log_root=%s", "Starting MDHelper %s with project_root=%s content_root=%s log_root=%s",
+17 -4
View File
@@ -248,7 +248,10 @@ def _pandoc_list_tables(source: str) -> str:
if match: if match:
prefix, marker, spacing = match.groups() prefix, marker, spacing = match.groups()
indent = prefix + ' ' * (len(marker) + len(spacing)) indent = prefix + ' ' * (len(marker) + len(spacing))
lines[row] = match.group() + '\n' + prefix + '\n' + indent + lines[row][match.end():] # A bare '-' after parent text is a Setext heading underline.
# '+' is the same bullet without changing the parent's tightness.
marker = '+' if marker == '-' else marker
lines[row] = prefix + marker + spacing + '\n' + prefix + '\n' + indent + lines[row][match.end():]
return ''.join(lines) return ''.join(lines)
@@ -324,6 +327,7 @@ def _add_python_docx_table(document: object, block: TableBlock) -> None:
if block.list_target: if block.list_target:
table._tbl.set(LIST_TARGET, block.list_target) table._tbl.set(LIST_TARGET, block.list_target)
table._tbl.set(LIST_LEVEL, str(block.list_level)) table._tbl.set(LIST_LEVEL, str(block.list_level))
if block.quote_depth: table._tbl.set(QUOTE_DEPTH, str(block.quote_depth))
table.style = "Table Grid" table.style = "Table Grid"
for row_index, row in enumerate(block.rows): for row_index, row in enumerate(block.rows):
for col_index in range(column_count): for col_index in range(column_count):
@@ -371,6 +375,7 @@ class TableBlock:
alignments: list[str] = field(default_factory=list) alignments: list[str] = field(default_factory=list)
list_target: str = '' list_target: str = ''
list_level: int = 0 list_level: int = 0
quote_depth: int = 0
class _DocxBuilder: class _DocxBuilder:
@@ -415,6 +420,8 @@ class _DocxBuilder:
elif token.type == "blockquote_close": elif token.type == "blockquote_close":
self.quote_depth = max(0, self.quote_depth - 1) self.quote_depth = max(0, self.quote_depth - 1)
elif token.type in ('bullet_list_open', 'ordered_list_open'): elif token.type in ('bullet_list_open', 'ordered_list_open'):
if self.item_pending and self.item_pending[-1]:
self._append_list_marker()
self.list_stack.append('bullet' if token.type == 'bullet_list_open' else 'number') self.list_stack.append('bullet' if token.type == 'bullet_list_open' else 'number')
self.list_sequence += 1 self.list_sequence += 1
self.list_ids.append(self.list_sequence) self.list_ids.append(self.list_sequence)
@@ -427,7 +434,7 @@ class _DocxBuilder:
self.list_quote_depths.pop() self.list_quote_depths.pop()
elif token.type == "table_open": elif token.type == "table_open":
if self.item_pending and self.item_pending[-1]: if self.item_pending and self.item_pending[-1]:
self._append_paragraph([], 'Normal', self.list_stack[-1]) self._append_list_marker()
self.blocks.append(self._collect_table()) self.blocks.append(self._collect_table())
continue continue
elif token.type == "hr": elif token.type == "hr":
@@ -435,6 +442,12 @@ class _DocxBuilder:
self.index += 1 self.index += 1
return self.blocks return self.blocks
def _append_list_marker(self) -> None:
self._append_paragraph([], 'Normal', self.list_stack[-1])
# The marker belongs to the list item, outside its leading quote/table.
self.blocks[-1].quote_depth = self.list_quote_depths[-1]
self.blocks[-1].inner_quote = False
def _add_next_inline(self, style: str, list_kind: str = "") -> None: def _add_next_inline(self, style: str, list_kind: str = "") -> None:
inline = self.tokens[self.index + 1] if self.index + 1 < len(self.tokens) else None inline = self.tokens[self.index + 1] if self.index + 1 < len(self.tokens) else None
runs = _inline_runs(inline.children if inline is not None else None) runs = _inline_runs(inline.children if inline is not None else None)
@@ -474,7 +487,7 @@ class _DocxBuilder:
self.index += 1 self.index += 1
break break
self.index += 1 self.index += 1
return TableBlock(rows, alignments, self.item_targets[-1] if self.item_targets else '', max(0, len(self.list_stack)-1)) return TableBlock(rows, alignments, self.item_targets[-1] if self.item_targets else '', max(0, len(self.list_stack)-1), self.quote_depth)
def _inline_runs(children: list[Token] | None) -> list[tuple[str, dict[str, object]]]: def _inline_runs(children: list[Token] | None) -> list[tuple[str, dict[str, object]]]:
@@ -599,7 +612,7 @@ def _numbering_xml(blocks):
def _table_xml(block: TableBlock) -> str: def _table_xml(block: TableBlock) -> str:
marker = f' xmlns:mdh="urn:mdhelper:source" mdh:list-target="{block.list_target}" mdh:list-level="{block.list_level}"' if block.list_target else '' marker = f' xmlns:mdh="urn:mdhelper:source" mdh:list-target="{block.list_target}" mdh:list-level="{block.list_level}" mdh:quote-depth="{block.quote_depth}"' if block.list_target or block.quote_depth else ''
rows = [] rows = []
for row in block.rows: for row in block.rows:
cells = [] cells = []
+35 -13
View File
@@ -142,7 +142,7 @@ class WordStyles:
base = 'table' base = 'table'
style = self.new(identifier, 'paragraph', 'MDHelper ' + style_name(key, self.rules), base_id) style = self.new(identifier, 'paragraph', 'MDHelper ' + style_name(key, self.rules), base_id)
rule = dict(self.effective[key]) rule = dict(self.effective[key])
if key == 'list' or key in LIST_TARGETS: rule = absolute_indents(rule) if key in ('list', 'blockquote') or key in LIST_TARGETS: rule = absolute_indents(rule)
if rule.get('font_color') == 'inherit': rule['font_color'] = 'auto' if rule.get('font_color') == 'inherit': rule['font_color'] = 'auto'
_paragraph_properties(child(style, 'pPr'), rule) _paragraph_properties(child(style, 'pPr'), rule)
if key in ('table_head', 'table_cell'): remove_fields(child(style, 'pPr'), ('shd',)) if key in ('table_head', 'table_cell'): remove_fields(child(style, 'pPr'), ('shd',))
@@ -151,7 +151,7 @@ class WordStyles:
if base: if base:
for tag, writer in (('pPr', _paragraph_properties), ('rPr', _run_properties)): for tag, writer in (('pPr', _paragraph_properties), ('rPr', _run_properties)):
base_props = ET.Element(W + tag) base_props = ET.Element(W + tag)
base_rule = absolute_indents(self.effective[base]) if base == 'list' or base in LIST_TARGETS else self.effective[base] base_rule = absolute_indents(self.effective[base]) if base in ('list', 'blockquote') or base in LIST_TARGETS else self.effective[base]
writer(base_props, base_rule) writer(base_props, base_rule)
if tag == 'rPr': remove_fields(base_props, ('shd',)) if tag == 'rPr': remove_fields(base_props, ('shd',))
_delta(child(style, tag), base_props, normalize_rule(self.rules.get(key))) _delta(child(style, tag), base_props, normalize_rule(self.rules.get(key)))
@@ -393,7 +393,7 @@ class WordStyles:
child(child(style, 'pPr'), 'ind').attrib.update({W + key:'0' for key in ('firstLine','firstLineChars','hanging','hangingChars')}) child(child(style, 'pPr'), 'ind').attrib.update({W + key:'0' for key in ('firstLine','firstLineChars','hanging','hangingChars')})
return identifier return identifier
def quoted_block(self, base_id, target, depth): def quoted_block(self, base_id, target, depth, *, offset_depth=None):
identifier = base_id + 'Quote' + str(depth) identifier = base_id + 'Quote' + str(depth)
if identifier not in self.created: if identifier not in self.created:
style = self.new(identifier, 'paragraph', reference(self.created[base_id], 'name') + f'(引用第 {depth} 层)', base_id) style = self.new(identifier, 'paragraph', reference(self.created[base_id], 'name') + f'(引用第 {depth} 层)', base_id)
@@ -408,12 +408,13 @@ class WordStyles:
ind = base.find('w:pPr/w:ind', NS) ind = base.find('w:pPr/w:ind', NS)
if ind is not None: attributes.update(ind.attrib) if ind is not None: attributes.update(ind.attrib)
points = font_points(self.effective['paragraph']) points = font_points(self.effective['paragraph'])
offset_depth = depth if offset_depth is None else offset_depth
props = child(child(style, 'pPr'), 'ind') props = child(child(style, 'pPr'), 'ind')
for side, padding in (('left', points + 3), ('right', points)): for side, padding in (('left', points + 3), ('right', points)):
value = number(attributes.get(W + side, 0)) / 20 value = number(attributes.get(W + side, 0)) / 20
chars = number(attributes.get(W + side + 'Chars', 0)) chars = number(attributes.get(W + side + 'Chars', 0))
if chars: value = chars / 100 * font_points(self.effective[target]) if chars: value = chars / 100 * font_points(self.effective[target])
props.set(W + side, str(round((value + padding * depth) * 20))) props.set(W + side, str(round((value + padding * offset_depth) * 20)))
props.set(W + side + 'Chars', '0') props.set(W + side + 'Chars', '0')
return identifier return identifier
@@ -486,6 +487,10 @@ class WordStyles:
rpr = numbering_level.find(W + 'rPr') rpr = numbering_level.find(W + 'rPr')
if rpr is not None: remove_fields(rpr, ('rFonts',)) if rpr is not None: remove_fields(rpr, ('rFonts',))
def table_padding(self, side):
rule = self.effective['table']
return number(rule.get('cell_padding', font_points(rule) * (.76 if side in ('left', 'right') else .62)))
def table(self): def table(self):
identifier = style_id('table') identifier = style_id('table')
style = self.new(identifier, 'table', 'MDHelper ' + style_name('table', self.rules)) style = self.new(identifier, 'table', 'MDHelper ' + style_name('table', self.rules))
@@ -504,7 +509,7 @@ class WordStyles:
element.attrib.update({W + 'val': border if visible else 'none', W + 'sz': str(max(2, round(width * 8))), W + 'color': color.lstrip('#') if color != 'transparent' else 'auto'}) element.attrib.update({W + 'val': border if visible else 'none', W + 'sz': str(max(2, round(width * 8))), W + 'color': color.lstrip('#') if color != 'transparent' else 'auto'})
margins = child(props, 'tblCellMar') margins = child(props, 'tblCellMar')
for side in ('top', 'left', 'bottom', 'right'): for side in ('top', 'left', 'bottom', 'right'):
padding = number(rule.get('cell_padding', font_points(rule) * (.76 if side in ('left', 'right') else .62))) padding = self.table_padding(side)
child(margins, side).attrib.update({W + 'w': str(round(padding * 20)), W + 'type': 'dxa'}) child(margins, side).attrib.update({W + 'w': str(round(padding * 20)), W + 'type': 'dxa'})
if 'background_color' in rule: if 'background_color' in rule:
child(child(style, 'tcPr'), 'shd').attrib.update({W + 'val': 'clear', W + 'fill': str(rule['background_color']).lstrip('#') if rule['background_color'] != 'transparent' else 'auto'}) child(child(style, 'tcPr'), 'shd').attrib.update({W + 'val': 'clear', W + 'fill': str(rule['background_color']).lstrip('#') if rule['background_color'] != 'transparent' else 'auto'})
@@ -558,7 +563,7 @@ class WordStyles:
continuation = paragraph.attrib.pop(CONTINUATION, '') == '1' continuation = paragraph.attrib.pop(CONTINUATION, '') == '1'
list_target = paragraph.attrib.pop(LIST_TARGET, '') list_target = paragraph.attrib.pop(LIST_TARGET, '')
quote_depth = int(paragraph.attrib.pop(QUOTE_DEPTH, '0')) quote_depth = int(paragraph.attrib.pop(QUOTE_DEPTH, '0'))
paragraph.attrib.pop(LIST_LEVEL, None) context_level = int(paragraph.attrib.pop(LIST_LEVEL, '0'))
ppr = properties(paragraph, 'pPr') ppr = properties(paragraph, 'pPr')
original_id = reference(ppr, 'pStyle') or 'Normal' original_id = reference(ppr, 'pStyle') or 'Normal'
names = {reference(style, 'name') for style in self.chain(original_id)} names = {reference(style, 'name') for style in self.chain(original_id)}
@@ -571,7 +576,7 @@ class WordStyles:
names.add(base) names.add(base)
loose = bool(names & set(SOURCE_PARAGRAPH_STYLES)) loose = bool(names & set(SOURCE_PARAGRAPH_STYLES))
target = self.classify(paragraph, original_id) target = self.classify(paragraph, original_id)
if target.startswith('list_') and list_target in LIST_TARGETS: target = list_target if target.startswith('list_') and target != 'list_task' and list_target in LIST_TARGETS: target = list_target
identifier = self.paragraph(target) identifier = self.paragraph(target)
numpr = ppr.find(W + 'numPr') numpr = ppr.find(W + 'numPr')
if numpr is not None: if numpr is not None:
@@ -599,11 +604,20 @@ class WordStyles:
if marker and not continuation: self.apply_list_marker(numpr, target, marker) if marker and not continuation: self.apply_list_marker(numpr, target, marker)
if loose: identifier = self.list_paragraph(target, identifier, level) if loose: identifier = self.list_paragraph(target, identifier, level)
if QUOTED_LIST_PARAGRAPH_STYLE in names: identifier = self.quoted_list(target, identifier, level) if QUOTED_LIST_PARAGRAPH_STYLE in names: identifier = self.quoted_list(target, identifier, level)
elif target == 'blockquote' and list_target in LIST_TARGETS:
base = self.nested_list(list_target, context_level) if context_level else self.paragraph(list_target)
identifier = self.quoted_list(list_target, base, context_level)
marker = self.effective[list_target].get('list_marker')
if marker and not continuation: self.apply_list_marker(numpr, list_target, marker)
if continuation: if continuation:
remove_fields(ppr, ('numPr',)) remove_fields(ppr, ('numPr',))
identifier = self.continuation_style(identifier) identifier = self.continuation_style(identifier)
if quote_depth and (target == 'code_block' or re.fullmatch('h[1-6]', target)): if quote_depth:
identifier = self.quoted_block(identifier, target, quote_depth) # The legacy quote paragraph style already represents its first
# quote. Keep that base editable and add only deeper containers.
standalone_quote = target == 'blockquote' and list_target not in LIST_TARGETS
offset = quote_depth - 1 if standalone_quote else quote_depth
if offset: identifier = self.quoted_block(identifier, target, quote_depth, offset_depth=offset)
alignment = reference(ppr, 'jc') alignment = reference(ppr, 'jc')
if target in ('table_head','table_cell') and alignment in ('left','center','right','both'): if target in ('table_head','table_cell') and alignment in ('left','center','right','both'):
identifier = self.table_column(target, alignment) identifier = self.table_column(target, alignment)
@@ -613,11 +627,19 @@ class WordStyles:
for table in document.iter(W + 'tbl'): for table in document.iter(W + 'tbl'):
list_target = table.attrib.pop(LIST_TARGET, '') list_target = table.attrib.pop(LIST_TARGET, '')
level = int(table.attrib.pop(LIST_LEVEL, '0')) level = int(table.attrib.pop(LIST_LEVEL, '0'))
table.attrib.pop(QUOTE_DEPTH, None); table.attrib.pop(CONTINUATION, None) depth = int(table.attrib.pop(QUOTE_DEPTH, '0')); table.attrib.pop(CONTINUATION, None)
props = properties(table, 'tblPr') props = properties(table, 'tblPr')
if list_target in LIST_TARGETS: if list_target in LIST_TARGETS or depth:
rule = self.contextual_list_rule(list_target, {}, level) left = 0
child(props, 'tblInd').attrib.update({W+'type':'dxa', W+'w':str(round(rule['left_indent_value'] * 20))}) if list_target in LIST_TARGETS:
rule = self.contextual_list_rule(list_target, {}, level)
left = rule['left_indent_value']
left += (font_points(self.effective['paragraph']) + 3) * depth
# With no tblInd, Word aligns the table edge at the margin.
# An explicit tblInd instead positions the first cell's text.
# Retain the cell inset so the whole table moves by the context.
left += self.table_padding('left')
child(props, 'tblInd').attrib.update({W+'type':'dxa', W+'w':str(round(left * 20))})
remove_fields(props, ('tblBorders', 'tblCellMar', 'shd')) remove_fields(props, ('tblBorders', 'tblCellMar', 'shd'))
val(props, 'tblStyle', style_id('table')) val(props, 'tblStyle', style_id('table'))
child(props, 'tblLayout').set(W + 'type', 'autofit') child(props, 'tblLayout').set(W + 'type', 'autofit')
+5 -2
View File
@@ -50,7 +50,9 @@ local function list_paragraphs(list)
for _, item in ipairs(list.content) do for _, item in ipairs(list.content) do
-- Word numbers paragraphs, not tables. Give a table-first item its own -- Word numbers paragraphs, not tables. Give a table-first item its own
-- empty numbered paragraph; its following text remains a continuation. -- empty numbered paragraph; its following text remains a continuation.
if item[1] and item[1].t == 'Table' then local first = item[1]
while first and (first.t == 'BlockQuote' or first.t == 'Div') do first = first.content[1] end
if first and first.t == 'Table' then
item:insert(1, pandoc.Div({pandoc.Para({pandoc.Str('\u{200b}')})}, item:insert(1, pandoc.Div({pandoc.Para({pandoc.Str('\u{200b}')})},
pandoc.Attr('', {}, {['custom-style']='MDHelperSourceTableMarker'}))) pandoc.Attr('', {}, {['custom-style']='MDHelperSourceTableMarker'})))
end end
@@ -98,7 +100,8 @@ local function contexts(blocks, target, level, depth, item_state)
elseif block.t == 'Div' then elseif block.t == 'Div' then
block.content = contexts(block.content, target, level, depth, item_state) block.content = contexts(block.content, target, level, depth, item_state)
else else
if block.t == 'Header' or block.t == 'CodeBlock' or block.t == 'Table' then if block.t == 'Header' or block.t == 'CodeBlock' or block.t == 'Table' or
(depth > 0 and (block.t == 'Para' or block.t == 'Plain')) then
local continuation = item_state and not item_state.pending and '1' or '0' local continuation = item_state and not item_state.pending and '1' or '0'
output:insert(pandoc.RawBlock('openxml', '<mdh:context xmlns:mdh="urn:mdhelper:source"' .. output:insert(pandoc.RawBlock('openxml', '<mdh:context xmlns:mdh="urn:mdhelper:source"' ..
' mdh:list-target="' .. target .. '" mdh:list-level="' .. math.max(0,level) .. ' mdh:list-target="' .. target .. '" mdh:list-level="' .. math.max(0,level) ..
+1 -1
View File
@@ -238,7 +238,7 @@ def _font_names(value):
if len(tokens) == 1 and tokens[0].type == "string": if len(tokens) == 1 and tokens[0].type == "string":
names.append(tokens[0].value) names.append(tokens[0].value)
elif tokens and all(t.type == "ident" for t in tokens): elif tokens and all(t.type == "ident" for t in tokens):
if any(t.lower_value in ('inherit', 'initial', 'unset', 'revert', 'revert-layer', 'default') for t in tokens): if len(tokens) == 1 and tokens[0].lower_value in ('inherit', 'initial', 'unset', 'revert', 'revert-layer', 'default'):
raise ValueError raise ValueError
names.append(" ".join(t.value for t in tokens)) names.append(" ".join(t.value for t in tokens))
else: else:
+39
View File
@@ -0,0 +1,39 @@
"""Multiword font family names may contain words that are reserved alone."""
import pytest
from mdhelper.style_sync import normalize_payload, analyze_css
from mdhelper.style_dialog import DocumentStyleEditorDialog
from mdhelper.docx_exporter import export_markdown_to_docx
from mdhelper.docx_style import W, NS
from test_style_regressions import browser, parts
from docx_assertions import effective_document
@pytest.mark.parametrize('family', ['Initial Font', 'Default Sans', 'Inherit Serif', 'Unset Sans', 'Revert Layer'])
def test_valid_family_syncs_browser_and_graphical_controls(qt_application, tmp_path, family):
css = f'p{{font:italic 20pt {family}}}'
payload = normalize_payload({'css':css})
assert not analyze_css(css).notes
assert payload['rules']['paragraph']['english_font_family'] == family
result = browser(qt_application, tmp_path, payload, 'Text', "(()=>{const s=getComputedStyle(document.querySelector('p'));return [parseFloat(s.fontSize),s.fontStyle]})()")
assert result[0] == pytest.approx(20*4/3, abs=.01) and result[1] == 'italic'
dialog = DocumentStyleEditorDialog(payload)
try:
assert dialog.font_size_spin.value() == 20
dialog.font_size_spin.setValue(22)
rule = analyze_css(dialog.style_payload()['css']).rules['paragraph']
assert rule['font_size'] == 22 and rule['english_font_family'] == family and rule['italic']
dialog.extra_css_edit.setPlainText(css)
assert dialog.style_payload()['rules']['paragraph']['font_size'] == 20
assert dialog.font_size_spin.value() == 20
finally: dialog.close()
@pytest.mark.parametrize('engine', ['builtin', 'python-docx', 'pandoc'])
def test_valid_family_exports_editable_size_and_italic(tmp_path, engine):
rules = normalize_payload({'css':'p{font:italic 20pt Initial Font}'})['rules']
doc, styles = parts(export_markdown_to_docx('Text', tmp_path/'font.docx', engine=engine, style_rules=rules))
base = styles.find("w:style[@w:styleId='MDHelperParagraph']/w:rPr", NS)
assert base.find(W+'sz').get(W+'val') == '40'
assert base.find(W+'i').get(W+'val') == '1'
base.find(W+'sz').set(W+'val', '48')
assert effective_document(doc, styles).find('.//w:r/w:rPr/w:sz', NS).get(W+'val') == '48'
+3 -3
View File
@@ -106,9 +106,9 @@ def test_word_preserves_all_settings_and_content(tmp_path: Path, engine):
assert spacing.get(W + "before") == spacing.get(W + "after") == "0" assert spacing.get(W + "before") == spacing.get(W + "after") == "0"
assert spacing.get(W + "line") == "400" and spacing.get(W + "lineRule") == "exact" assert spacing.get(W + "line") == "400" and spacing.get(W + "lineRule") == "exact"
ind = ppr.find(W + "ind") ind = ppr.find(W + "ind")
# List geometry uses absolute lengths so Word's document grid cannot # List/quote geometry uses absolute lengths, independent of Word's grid.
# shift code/heading continuations away from their list's text column. absolute = ppr.find(W + 'numPr') is not None or ''.join(p.itertext()).startswith('Quote')
assert ind.get(W + "firstLineChars") == ("0" if ppr.find(W + 'numPr') is not None else "200") assert ind.get(W + "firstLineChars") == ("0" if absolute else "200")
assert ind.get(W + "firstLine") == "420" and ind.get(W + "right") == "120" assert ind.get(W + "firstLine") == "420" and ind.get(W + "right") == "120"
assert ind.get(W + "left") == "850" assert ind.get(W + "left") == "850"
for name, value in (("keepNext", "1"), ("keepLines", "1"), ("contextualSpacing", "1"), ("pageBreakBefore", "0"), ("widowControl", "0")): for name, value in (("keepNext", "1"), ("keepLines", "1"), ("contextualSpacing", "1"), ("pageBreakBefore", "0"), ("widowControl", "0")):
+29
View File
@@ -0,0 +1,29 @@
"""Empty and invalid directory history must not resolve to the working folder."""
from unittest.mock import MagicMock
import pytest
from PySide6.QtCore import QSettings
from mdhelper import app
@pytest.mark.parametrize('history', [None, '', ' ', 'missing', 'file', 'directory'])
def test_startup_history_requires_a_nonempty_directory(tmp_path, monkeypatch, history):
settings = QSettings(str(tmp_path/'settings.ini'), QSettings.Format.IniFormat)
folder = tmp_path/'project'; folder.mkdir()
file = folder/'note.md'; file.write_text('Text')
value = {'missing': str(tmp_path/'missing'), 'file': str(file), 'directory': str(folder)}.get(history, history)
if value is not None: settings.setValue('project/last_root', value)
window = MagicMock()
monkeypatch.setattr(app, 'QSettings', lambda *_: settings)
monkeypatch.setattr(app, 'QApplication', MagicMock())
monkeypatch.setattr(app, 'MainWindow', window)
monkeypatch.setattr(app, 'configure_process_runtime', lambda *_: None)
monkeypatch.setattr(app, 'setup_logging', lambda *_: None)
app.main([])
args = window.call_args.kwargs
assert args['project_root'] == (folder if history == 'directory' else None)
assert args['prompt_for_project'] == (history != 'directory')
assert args['initial_path'] is None
# Explicit startup files still take precedence over all history values.
app.main([str(file)])
assert window.call_args.kwargs['project_root'] == folder
assert window.call_args.kwargs['initial_path'] == file
+2 -1
View File
@@ -93,7 +93,8 @@ def test_quoted_list_keeps_additive_nested_indents(tmp_path, engine):
text=''.join(p.itertext()) text=''.join(p.itertext())
if text in ('Outer','Inner'): if text in ('Outer','Inner'):
ind=p.find('w:pPr/w:ind',NS) ind=p.find('w:pPr/w:ind',NS)
assert int(ind.get(W+'left'))/20 == (36 if text=='Outer' else 72) # Include the quote container's 12 pt padding and 3 pt border.
assert int(ind.get(W+'left'))/20 == (51 if text=='Outer' else 87)
assert ind.get(W+'hangingChars','0')=='0' assert ind.get(W+'hangingChars','0')=='0'
+68
View File
@@ -0,0 +1,68 @@
"""Nested list markers and quote containers remain shared, editable Word styles."""
import pytest
from mdhelper.docx_exporter import export_markdown_to_docx
from mdhelper.docx_style import W, NS
from test_style_regressions import parts
from docx_assertions import effective_document
ENGINES = ['builtin', 'python-docx', 'pandoc']
NESTED_TABLE = '1. Outer\n - | HeadOne | HeadTwo |\n |---|---|\n | CellOne | CellTwo |\n\n2. End'
QUOTE_TABLE = '1. > | HeadOne | HeadTwo |\n > |---|---|\n > | CellOne | CellTwo |\n\n2. End'
NESTED_FIRST = '1. - FirstChild\n - SecondChild\n\n2. End'
QUOTE_TEXT = '> OuterText\n>\n> > InnerText\n> >\n> > ### InnerHeading\n> >\n> > ```text\n> > InnerCode\n> > ```'
MIXED_QUOTE = '1. Outer\n\n > - Inner\n >\n > Continued\n >\n > ### Heading\n >\n > ```text\n > CodeLine\n > ```\n\n2. End'
@pytest.mark.parametrize('engine', ENGINES)
@pytest.mark.parametrize('source', [NESTED_TABLE, QUOTE_TABLE, NESTED_FIRST])
def test_nested_blocks_keep_parent_marker_and_content(tmp_path, engine, source):
doc, styles = parts(export_markdown_to_docx(source, tmp_path/'nested.docx', engine=engine))
paragraphs = doc.findall('w:body/w:p', NS)
first, last = paragraphs[0], paragraphs[-1]
assert first.find('w:pPr/w:numPr', NS) is not None
assert first.find('w:pPr/w:numPr/w:numId', NS).get(W+'val') == last.find('w:pPr/w:numPr/w:numId', NS).get(W+'val')
assert 'H2' not in first.find('w:pPr/w:pStyle', NS).get(W+'val')
if 'HeadOne' in source:
assert [''.join(c.itertext()) for c in doc.findall('.//w:tbl/w:tr/w:tc', NS)] == ['HeadOne','HeadTwo','CellOne','CellTwo']
assert '\u200b' not in ''.join(doc.itertext())
assert not any(a.startswith('{urn:mdhelper:source}') for el in doc.iter() for a in el.attrib)
@pytest.mark.parametrize('engine', ENGINES)
def test_parent_text_after_nested_list_is_a_continuation(tmp_path, engine):
source = NESTED_FIRST.replace('\n\n2. End', '\n\n Continued\n\n2. End')
doc, _ = parts(export_markdown_to_docx(source, tmp_path/'continued.docx', engine=engine))
p = next(p for p in doc.iter(W+'p') if ''.join(p.itertext()) == 'Continued')
assert p.find('w:pPr/w:numPr', NS) is None
@pytest.mark.parametrize('engine', ENGINES)
def test_nested_quote_text_keeps_depth_and_base_font_inheritance(tmp_path, engine):
doc, styles = parts(export_markdown_to_docx(QUOTE_TEXT, tmp_path/'quote.docx', engine=engine))
effective = effective_document(doc, styles)
def left(text):
p = next(p for p in effective.iter(W+'p') if ''.join(p.itertext()) == text)
return int(p.find('w:pPr/w:ind', NS).get(W+'left'))
assert left('InnerText') - left('OuterText') == 300
styles.find("w:style[@w:styleId='MDHelperBlockquote']/w:rPr/w:color", NS).set(W+'val','008000')
for p in effective_document(doc,styles).iter(W+'p'):
if ''.join(p.itertext()) in ('OuterText', 'InnerText'):
assert p.find('.//w:rPr/w:color',NS).get(W+'val') == '008000'
@pytest.mark.parametrize('engine', ENGINES)
def test_mixed_quote_containers_align_text_headings_and_code(tmp_path, engine):
doc, styles = parts(export_markdown_to_docx(MIXED_QUOTE, tmp_path/'mixed.docx', engine=engine))
positions = []
for p in effective_document(doc,styles).iter(W+'p'):
if ''.join(p.itertext()) in ('Inner','Continued','Heading','CodeLine'):
positions.append(int(p.find('w:pPr/w:ind',NS).get(W+'left')))
assert positions == [1500]*4
@pytest.mark.parametrize('engine', ENGINES)
@pytest.mark.parametrize('depth', [1, 2])
def test_quoted_tables_preserve_container_depth(tmp_path, engine, depth):
source = '\n'.join('> '*depth+line for line in ['| Head |','|---|','| Cell |'])
doc, _ = parts(export_markdown_to_docx(source, tmp_path/'table.docx', engine=engine))
assert doc.find('.//w:tblPr/w:tblInd', NS).get(W+'w') == str(round(depth*300 + 12*.76*20))
+1 -1
View File
@@ -45,7 +45,7 @@ print(folder)
if($text -eq 'Continuation' -and $range.ListFormat.ListString.Trim() -ne ''){throw "$engine continuation consumed a marker"} if($text -eq 'Continuation' -and $range.ListFormat.ListString.Trim() -ne ''){throw "$engine continuation consumed a marker"}
} }
} elseif($case -eq 'quoted') { } elseif($case -eq 'quoted') {
foreach($pair in @(@('Outer',36),@('Inner',72))) { foreach($pair in @(@('Outer',51),@('Inner',87))) {
$range=$doc.Content.Duplicate;[void]$range.Find.Execute($pair[0]) $range=$doc.Content.Duplicate;[void]$range.Find.Execute($pair[0])
if([Math]::Abs($range.ParagraphFormat.LeftIndent-$pair[1]) -gt 0.1){throw "$engine quoted $($pair[0]) indent: $($range.ParagraphFormat.LeftIndent)"} if([Math]::Abs($range.ParagraphFormat.LeftIndent-$pair[1]) -gt 0.1){throw "$engine quoted $($pair[0]) indent: $($range.ParagraphFormat.LeftIndent)"}
} }
+87
View File
@@ -0,0 +1,87 @@
param([string]$Python = (Join-Path $PSScriptRoot '..\mdhelper_env\mdhelper_env\Scripts\python.exe'))
$ErrorActionPreference='Stop'
Push-Location (Join-Path $PSScriptRoot '..')
$word=$null
try {
$folder = (@'
import sys,tempfile
from pathlib import Path
sys.path.insert(0,str(Path('tests').resolve()))
from test_style_revision_058 import *
from mdhelper.style_sync import normalize_payload
folder=Path(tempfile.mkdtemp(prefix='MDHelper-058-Word-'))
for engine in ENGINES:
for name,source in [('nested-table',NESTED_TABLE),('quote-table',QUOTE_TABLE),('nested-first',NESTED_FIRST),('quote-text',QUOTE_TEXT),('mixed',MIXED_QUOTE)]:
export_markdown_to_docx(source,folder/f'{engine}-{name}.docx',engine=engine)
for depth in range(3):
source='\n'.join('> '*depth+line for line in ['| Head |','|---|','| Cell |'])
export_markdown_to_docx(source,folder/f'{engine}-table-{depth}.docx',engine=engine)
rules=normalize_payload({'css':'p{font:italic 20pt Initial Font}'})['rules']
export_markdown_to_docx('Text',folder/f'{engine}-font.docx',engine=engine,style_rules=rules)
print(folder)
'@ | & $Python -B -).Trim()
if($LASTEXITCODE -ne 0){throw 'Fixture generation failed'}
$word=New-Object -ComObject Word.Application
$word.Visible=$false;$word.DisplayAlerts=0
foreach($engine in @('builtin','python-docx','pandoc')) {
foreach($case in @('nested-table','quote-table','nested-first','quote-text','mixed','table-0','table-1','table-2','font')) {
$doc=$word.Documents.Open((Join-Path $folder "$engine-$case.docx"),$false,$true,$false)
try {
if($case -in @('nested-table','quote-table','nested-first')) {
if($doc.Paragraphs.Item(1).Range.ListFormat.ListString.Trim() -ne '1.'){throw "$engine $case parent number missing"}
$r=$doc.Content.Duplicate;[void]$r.Find.Execute('End')
if($r.ListFormat.ListString.Trim() -ne '2.'){throw "$engine $case next number wrong"}
if($case -eq 'nested-table') {
$r=$doc.Content.Duplicate;[void]$r.Find.Execute('Outer')
if($r.Style.NameLocal -match '标题|Heading'){throw "$engine parent became a heading"}
}
} elseif($case -eq 'font') {
$r=$doc.Content.Duplicate;[void]$r.Find.Execute('Text')
if($r.Font.Size -ne 20 -or $r.Font.Italic -ne -1){throw "$engine font CSS lost"}
$doc.Styles.Item('MDHelper 正文').Font.Size=24
if($r.Font.Size -ne 24){throw "$engine font no longer follows paragraph style"}
} else {
$doc.ExportAsFixedFormat((Join-Path $folder "$engine-$case.pdf"),17)
if($case -eq 'quote-text') {
$doc.Styles.Item('MDHelper 引用').Font.Color=255
foreach($text in @('OuterText','InnerText')) {
$r=$doc.Content.Duplicate;[void]$r.Find.Execute($text)
if($r.Font.Color -ne 255){throw "$engine nested quote lost base inheritance"}
}
} elseif($case -eq 'mixed') {
$doc.Styles.Item('MDHelper 标题 3').Font.Size=20
$r=$doc.Content.Duplicate;[void]$r.Find.Execute('Heading')
if($r.Font.Size -ne 20){throw "$engine nested heading lost base inheritance"}
}
}
Write-Output "$engine $case PASS"
} finally {$doc.Close(0)}
}
}
@'
import sys
from pathlib import Path
from pypdf import PdfReader
folder=Path(sys.argv[1])
def positions(name):
result={}
def visit(text,cm,tm,font,size):
if text.strip():result[text.strip()]=tm[4]*cm[0]+tm[5]*cm[2]+cm[4]
for page in PdfReader(folder/name).pages:page.extract_text(visitor_text=visit)
return result
for engine in ('builtin','python-docx','pandoc'):
quote=positions(f'{engine}-quote-text.pdf')
assert abs(quote['InnerText']-quote['OuterText']-15)<.3,(engine,quote)
mixed=positions(f'{engine}-mixed.pdf')
values=[mixed[x] for x in ('Inner','Continued','Heading','CodeLine')]
assert max(values)-min(values)<.3,(engine,mixed)
tables=[positions(f'{engine}-table-{d}.pdf')['Head'] for d in range(3)]
assert all(abs(tables[d]-tables[d-1]-15)<.3 for d in (1,2)),(engine,tables)
print(engine,'actual Word quote/table layout PASS')
'@ | & $Python -B - $folder
if($LASTEXITCODE -ne 0){throw 'Word PDF layout failed'}
Write-Output "Fixtures: $folder"
} finally {
if($word){$word.Quit();[void][Runtime.InteropServices.Marshal]::FinalReleaseComObject($word)}
Pop-Location
}