def _extract_attrs(x, n):
"""Extracts attributes for an image. n is the index where the
attributes begin. Extracted elements are deleted from the element
list x. Attrs are returned in pandoc format.
"""
try:
return extract_attrs(x, n)
except (ValueError, IndexError):
if PANDOCVERSION < '1.16':
# Look for attributes attached to the image path, as occurs with
# image references for pandoc < 1.16 (pandoc-fignos Issue #14).
# See http://pandoc.org/MANUAL.html#images for the syntax.
# Note: This code does not handle the "optional title" for
# image references (search for link_attributes in pandoc's docs).
assert x[n - 1]['t'] == 'Image'
image = x[n - 1]
s = image['c'][-1][0]
if '%20%7B' in s:
path = s[:s.index('%20%7B')]
attrs = unquote(s[s.index('%7B'):])
image['c'][-1][0] = path # Remove attr string from the path
return PandocAttributes(attrs.strip(), 'markdown').to_pandoc()
raise
def action(key, value, fmt, meta): # pylint: disable=unused-argument
"""Processes elements."""
global has_lettrine # pylint: disable=global-statement
if key == 'Span':
attrs = PandocAttributes(value[0], 'pandoc')
if 'lettrine' in attrs.classes:
has_lettrine = True
firstword = value[1][0]['c']
content = value[1][1:]
# Replace span in para with the elements
ret = [PRE, Str(firstword[0]), MID]
if len(firstword) > 1:
ret.append(Str(firstword[1:]))
ret.append(POST)
ret += content
return ret
return None
def action(key, value, fmt, meta): # pylint: disable=unused-argument
"""Processes elements."""
global replaced_figure_env # pylint: disable=global-statement
if is_figure(key, value):
attrs = PandocAttributes(value[0]['c'][0], 'pandoc')
# Convert figures with `marginfigure` class to marginfigures
if 'marginfigure' in attrs.classes:
if 'documentclass' in meta and \
get_meta(meta, 'documentclass') in \
['tufte-book', 'tufte-handout']:
replaced_figure_env = True
# Get the marginfigure options
offset = attrs['offset'] if 'offset' in attrs else '0pt'
# LaTeX used to apply the environment
pre = RawBlock('tex', r'\begin{marginfigure_}[%s]' % offset)
post = RawBlock('tex', r'\end{marginfigure_}')
return [pre, Para(value), post]
if warninglevel:
STDERR.write(WARNING)
return None
def action(key, value, fmt, meta): # pylint: disable=unused-argument
"""Processes elements."""
if key == 'Div':
attrs = PandocAttributes(value[0], 'pandoc')
if 'marginnote' in attrs.classes:
offset = attrs['offset'] if 'offset' in attrs else '0pt'
if 'documentclass' in meta and \
get_meta(meta, 'documentclass') in \
['tufte-book', 'tufte-handout']:
pre = RawInline('tex', r'\marginnote[%s]{'%offset)
post = RawInline('tex', r'}')
assert value[1][0]['t'] == 'Para'
assert value[1][-1]['t'] == 'Para'
value[1][0]['c'].insert(0, pre)
value[1][-1]['c'].append(post)
elif warninglevel:
STDERR.write(WARNING)
def _process_equation(value, fmt):
"""Processes the equation. Returns a dict containing eq properties."""
# pylint: disable=global-statement
global Nreferences # Global references counter
global cursec # Current section
global has_unnumbered_equations # Flags that unnumbered eqs were found
# Initialize the return value
eq = {
'is_unnumbered': False,
'is_unreferenceable': False,
'is_tagged': False
}
# Parse the equation
attrs = eq['attrs'] = PandocAttributes(value[0], 'pandoc')
# Bail out if the label does not conform to expectations
if not LABEL_PATTERN.match(attrs.id):
eq.update({'is_unnumbered': True, 'is_unreferenceable': True})
return eq
# Identify unreferenceable equations
if attrs.id == 'eq:': # Make up a unique description
attrs.id += str(uuid.uuid4())
eq['is_unreferenceable'] = True
# Update the current section number
if attrs['secno'] != cursec: # The section number changed
cursec = attrs['secno'] # Update the global section tracker
Nreferences = 1 # Resets the global reference counter
# Pandoc's --number-sections supports section numbering latex/pdf, html,
# epub, and docx
if numbersections:
# Latex/pdf supports equation numbers by section natively. For the
# other formats we must hard-code in equation numbers by section as
# tags.
if fmt in ['html', 'html5', 'epub', 'epub2', 'epub3', 'docx'] and \
'tag' not in attrs:
attrs['tag'] = str(cursec + secoffset) + '.' + str(Nreferences)
Nreferences += 1
# Save reference information
eq['is_tagged'] = 'tag' in attrs
if eq['is_tagged']: # ... then save the tag
# Remove any surrounding quotes
if attrs['tag'][0] == '"' and attrs['tag'][-1] == '"':
attrs['tag'] = attrs['tag'].strip('"')
elif attrs['tag'][0] == "'" and attrs['tag'][-1] == "'":
attrs['tag'] = attrs['tag'].strip("'")
references[attrs.id] = pandocxnos.Target(attrs['tag'], cursec, attrs.id
in references)
else:
references[attrs.id] = pandocxnos.Target(Nreferences, cursec, attrs.id
in references)
Nreferences += 1 # Increment the global reference counter
return eq
def _process_theorem(value, fmt):
"""Processes the theorem. Returns a dict containing theorem properties."""
# pylint: disable=global-statement
global Ntargets # Global targets counter
global cursec # Current section
# Initialize the return value
thm = {'is_unreferenceable': False, 'is_tagged': False}
# Parse the theorem attributes
attrs = thm['attrs'] = PandocAttributes(value[0], 'pandoc')
# Bail out if the label does not conform to expectations
assert LABEL_PATTERN and LABEL_PATTERN.match(attrs.id)
# Identify unreferenceable theorems
if attrs.id[-1] == ':': # Make up a unique description
attrs.id += str(uuid.uuid4())
thm['is_unreferenceable'] = True
counter = attrs.id.split(':')[0]
if sharedcounter:
counter = 'shared'
# Update the current section number
if attrs['secno'] != cursec: # The section number changed
cursec = attrs['secno'] # Update the global section tracker
for key, nref in Ntargets.items(): # pylint: disable=unused-variable
Ntargets[key] = 1 # Resets the global target counter
# Pandoc's --number-sections supports section numbering latex/pdf, html,
# epub, and docx
if numbersections:
# Latex/pdf supports theorems numbers by section natively. For the
# other formats we must hard-code in theorem numbers by section as
# tags.
if fmt in ['html', 'html5', 'epub', 'epub2', 'epub3', 'docx'] and \
'tag' not in attrs:
attrs['tag'] = str(cursec+secoffset) + '.' + \
str(Ntargets[counter])
Ntargets[counter] += 1
# Save reference information
thm['is_tagged'] = 'tag' in attrs
if thm['is_tagged']: # ... then save the tag
# Remove any surrounding quotes
if attrs['tag'][0] == '"' and attrs['tag'][-1] == '"':
attrs['tag'] = attrs['tag'].strip('"')
elif attrs['tag'][0] == "'" and attrs['tag'][-1] == "'":
attrs['tag'] = attrs['tag'].strip("'")
targets[attrs.id] = pandocxnos.Target(attrs['tag'], cursec, attrs.id
in targets)
else:
targets[attrs.id] = pandocxnos.Target(Ntargets[counter], cursec,
attrs.id in targets)
Ntargets[counter] += 1 # Increment the global reference counter
return thm
def test_kvs(self):
"""Tests PandocAttributes.kvs."""
attrs = PandocAttributes(['', [], [['tag', 'B.1']]], 'pandoc')
kvs = attrs.kvs
# Ensure that changing the kvs changes attrs too
kvs['tag'] = 'B.3'
self.assertEqual(attrs['tag'], 'B.3')
def action(key, value, fmt, meta): # pylint: disable=unused-argument
"""Processes elements."""
if key == 'Div':
attrs = PandocAttributes(value[0], 'pandoc')
if 'noindent' in attrs.classes:
return [PRE, Div(*value), POST]
return None
def action(key, value, fmt, meta): # pylint: disable=unused-argument
"""Processes elements."""
if key == 'Span':
attrs = PandocAttributes(value[0], 'pandoc')
if 'newthought' in attrs.classes:
if 'documentclass' in meta and \
get_meta(meta, 'documentclass') in \
['tufte-book', 'tufte-handout']:
pre = RawInline('tex', r'\newthought{')
post = RawInline('tex', r'}')
value[1].insert(0, pre)
value[1].append(post)
elif warninglevel:
STDERR.write(WARNING)
def processor(meta, blocks):
"""Document processor."""
has_epigraph = False # Flags that an epigraph was found
noindent = False # Flags that next para should not be indented
# Process the blocks
for block in blocks:
if block['t'] == 'Div':
attrs = PandocAttributes(block['c'][0])
# Process epigraph divs
if 'epigraph' in attrs.classes:
# Insert tex into div content
content = block['c'][1]
content[0]['c'].insert(0, PRE)
content[-2]['c'].append(MID)
for el in content[-1]['c']:
content[-2]['c'].append(el)
content[-2]['c'].append(POST)
del content[-1]
# Set flags and continue
has_epigraph = True
noindent = True
continue
# Don't indent the first non-empty paragraph after an epigraph
if block['t'] == 'Para' and noindent:
content = stringify(quotify(copy.deepcopy(block['c'])))
if content.strip():
block['c'].insert(0, RawInline('tex', r'\noindent '))
noindent = False
if has_epigraph:
pandocxnos.add_to_header_includes(
meta, 'tex', textwrap.dedent(r'''
\usepackage{epigraph}
'''), warninglevel)
def _process_equation(value, fmt):
"""Processes the equation. Returns a dict containing eq properties."""
# pylint: disable=global-statement
global Nreferences # Global references counter
global cursec # Current section
# Parse the equation
attrs = value[0]
# Initialize the return value
eq = {'is_unnumbered': False,
'is_unreferenceable': False,
'is_tagged': False,
'attrs': attrs}
# Bail out if the label does not conform
if not LABEL_PATTERN.match(attrs[0]):
eq['is_unnumbered'] = True
eq['is_unreferenceable'] = True
return eq
# Process unreferenceable equations
if attrs[0] == 'eq:': # Make up a unique description
attrs[0] = attrs[0] + str(uuid.uuid4())
eq['is_unreferenceable'] = True
unreferenceable.append(attrs[0])
# For html, hard-code in the section numbers as tags
kvs = PandocAttributes(attrs, 'pandoc').kvs
if numbersections and fmt in ['html', 'html5'] and 'tag' not in kvs:
if kvs['secno'] != cursec:
cursec = kvs['secno']
Nreferences = 1
kvs['tag'] = cursec + '.' + str(Nreferences)
Nreferences += 1
# Save to the global references tracker
eq['is_tagged'] = 'tag' in kvs
if eq['is_tagged']:
# Remove any surrounding quotes
if kvs['tag'][0] == '"' and kvs['tag'][-1] == '"':
kvs['tag'] = kvs['tag'].strip('"')
elif kvs['tag'][0] == "'" and kvs['tag'][-1] == "'":
kvs['tag'] = kvs['tag'].strip("'")
references[attrs[0]] = kvs['tag']
else:
Nreferences += 1
references[attrs[0]] = Nreferences
# Adjust equation depending on the output format
if fmt in ['latex', 'beamer']:
if not eq['is_unreferenceable']: # Code in the tags
value[-1] += r'\tag{%s}\label{%s}' % \
(references[attrs[0]].replace(' ', r'\ '), attrs[0]) \
if eq['is_tagged'] else r'\label{%s}'%attrs[0]
elif fmt in ('html', 'html5'):
pass # Insert html in process_equations() instead
else: # Hard-code in the number/tag
if isinstance(references[attrs[0]], int): # Numbered reference
value[-1] += r'\qquad (%d)' % references[attrs[0]]
else: # Tagged reference
assert isinstance(references[attrs[0]], STRTYPES)
text = references[attrs[0]].replace(' ', r'\ ')
if text.startswith('$') and text.endswith('$'): # Math
tag = text[1:-1]
else: # Text
tag = r'\text{%s}' % text
value[-1] += r'\qquad (%s)' % tag
return eq
def _process_figure(value, fmt):
"""Processes the figure. Returns a dict containing figure properties."""
# pylint: disable=global-statement
global Nreferences # Global references counter
global has_unnumbered_figures # Flags unnumbered figures were found
global cursec # Current section
# Parse the image
attrs, caption = value[0]['c'][:2]
# Initialize the return value
fig = {
'is_unnumbered': False,
'is_unreferenceable': False,
'is_tagged': False,
'attrs': attrs
}
# Bail out if the label does not conform
if not LABEL_PATTERN.match(attrs[0]):
has_unnumbered_figures = True
fig['is_unnumbered'] = True
fig['is_unreferenceable'] = True
return fig
# Process unreferenceable figures
if attrs[0] == 'fig:': # Make up a unique description
attrs[0] = attrs[0] + str(uuid.uuid4())
fig['is_unreferenceable'] = True
unreferenceable.append(attrs[0])
# For html, hard-code in the section numbers as tags
kvs = PandocAttributes(attrs, 'pandoc').kvs
if numbersections and fmt in ['html', 'html5'] and not 'tag' in kvs:
if kvs['secno'] != cursec:
cursec = kvs['secno']
Nreferences = 1
kvs['tag'] = cursec + '.' + str(Nreferences)
Nreferences += 1
# Save to the global references tracker
fig['is_tagged'] = 'tag' in kvs
if fig['is_tagged']:
# Remove any surrounding quotes
if kvs['tag'][0] == '"' and kvs['tag'][-1] == '"':
kvs['tag'] = kvs['tag'].strip('"')
elif kvs['tag'][0] == "'" and kvs['tag'][-1] == "'":
kvs['tag'] = kvs['tag'].strip("'")
references[attrs[0]] = kvs['tag']
else:
Nreferences += 1
references[attrs[0]] = Nreferences
# Adjust caption depending on the output format
if fmt in ['latex', 'beamer']: # Append a \label if this is referenceable
if not fig['is_unreferenceable']:
value[0]['c'][1] += [RawInline('tex', r'\label{%s}' % attrs[0])]
else: # Hard-code in the caption name and number/tag
if type(references[attrs[0]]) is int: # Numbered reference
value[0]['c'][1] = [Str(captionname), Space(),
Str('%d:'%references[attrs[0]]), Space()] + \
list(caption)
else: # Tagged reference
assert type(references[attrs[0]]) in STRTYPES
text = references[attrs[0]]
if text.startswith('$') and text.endswith('$'): # Math
math = text.replace(' ', r'\ ')[1:-1]
els = [Math({"t": "InlineMath", "c": []}, math), Str(':')]
else: # Text
els = [Str(text + ':')]
value[0]['c'][1] = [Str(captionname), Space()]+ els + [Space()] + \
list(caption)
return fig
def _process_figure(key, value, fmt):
"""Processes a figure. Returns a dict containing figure properties.
Parameters:
key - 'Para' (for a normal figure) or 'Div'
value - the content of the figure
fmt - the output format ('tex', 'html', ...)
"""
# pylint: disable=global-statement
global cursec # Current section being processed
global Ntargets # Number of targets in current section (or document)
global has_unnumbered_figures # Flags that unnumbered figures were found
# Initialize the return value
fig = {
'is_unnumbered': False,
'is_unreferenceable': False,
'is_tagged': False
}
# Bail out if there are no attributes
if key == 'Para' and len(value[0]['c']) == 2:
has_unnumbered_figures = True
fig.update({'is_unnumbered': True, 'is_unreferenceable': True})
return fig
# Parse the figure
attrs = fig['attrs'] = \
PandocAttributes(value[0]['c'][0] if key == 'Para' else value[0],
'pandoc')
fig['caption'] = value[0]['c'][1] if key == 'Para' else None
# Bail out if the label does not conform to expectations
if not LABEL_PATTERN.match(attrs.id):
has_unnumbered_figures = True
fig.update({'is_unnumbered': True, 'is_unreferenceable': True})
return fig
# Identify unreferenceable figures
if attrs.id == 'fig:':
attrs.id += str(uuid.uuid4())
fig['is_unreferenceable'] = True
# Update the current section number
if attrs['secno'] != cursec: # The section number changed
cursec = attrs['secno'] # Update the global section tracker
Ntargets = 1 # Resets the global target counter
# Pandoc's --number-sections supports section numbering latex/pdf, html,
# epub, and docx
if numbersections:
# Latex/pdf supports equation numbers by section natively. For the
# other formats we must hard-code in figure numbers by section as
# tags.
if fmt in ['html', 'html5', 'epub', 'epub2', 'epub3', 'docx'] and \
'tag' not in attrs:
attrs['tag'] = str(cursec + secoffset) + '.' + str(Ntargets)
Ntargets += 1
# Update the global targets tracker
fig['is_tagged'] = 'tag' in attrs
if fig['is_tagged']: # ... then save the tag
# Remove any surrounding quotes
if attrs['tag'][0] == '"' and attrs['tag'][-1] == '"':
attrs['tag'] = attrs['tag'].strip('"')
elif attrs['tag'][0] == "'" and attrs['tag'][-1] == "'":
attrs['tag'] = attrs['tag'].strip("'")
targets[attrs.id] = pandocxnos.Target(attrs['tag'], cursec, attrs.id
in targets)
else: # ... then save the figure number
targets[attrs.id] = pandocxnos.Target(Ntargets, cursec, attrs.id
in targets)
Ntargets += 1 # Increment the global targets counter
return fig
def _process_equation(value, fmt):
"""Processes the equation. Returns a dict containing eq properties."""
# pylint: disable=global-statement
global Ntargets # Global targets counter
global cursec # Current section
global has_unnumbered_equations # Flags that unnumbered eqs were found
# Initialize the return value
eq = {
'is_unnumbered': False,
'is_unreferenceable': False,
'is_tagged': False
}
# Parse the equation
attrs = eq['attrs'] = PandocAttributes(value[0], 'pandoc')
import sys
# Bail out if the label does not conform to expectations
if not LABEL_PATTERN.match(attrs.id):
eq.update({'is_unnumbered': True, 'is_unreferenceable': True})
return eq
# Identify unreferenceable equations
if attrs.id == 'eq:': # Make up a unique description
attrs.id += str(uuid.uuid4())
eq['is_unreferenceable'] = True
sublabels = _get_sublabels(attrs)
# Update the current section number
if attrs['secno'] != cursec: # The section number changed
cursec = attrs['secno'] # Update the global section tracker
if numbersections:
Ntargets = 0 # Resets the global targets counter
# Pandoc's --number-sections supports section numbering latex/pdf, html,
# epub, and docx
section_prefix = ""
if numbersections:
# Latex/pdf supports equation numbers by section natively. For the
# other formats we must hard-code in equation numbers by section as
# tags.
if fmt in ['html', 'html4', 'html5', 'epub', 'epub2', 'epub3', 'docx']:
section_prefix = str(cursec + secoffset) + '.'
eq['is_tagged'] = 'tag' in attrs
# Save reference information
if eq['is_tagged']: # ... then save the tag
# Remove any surrounding quotes
if attrs['tag'][0] == '"' and attrs['tag'][-1] == '"':
attrs['tag'] = attrs['tag'].strip('"')
elif attrs['tag'][0] == "'" and attrs['tag'][-1] == "'":
attrs['tag'] = attrs['tag'].strip("'")
targets[attrs.id] = pandocxnos.Target(attrs['tag'], cursec, attrs.id
in targets)
for sublabel in sublabels:
# Sublabel number Ntargets will not be shown, fill in a dummy value
targets[sublabel] = pandocxnos.Target(Ntargets, cursec, sublabel
in targets)
else: # Not tagged = numbered
if not sublabels:
Ntargets += 1
targets[attrs.id] = pandocxnos.Target(
section_prefix + str(Ntargets), cursec, attrs.id in targets)
else:
if eq['is_unreferenceable']:
# Main label does not increase number
# If there are sublabels, treat each sublabel as an equation with a new number.
targets[attrs.id] = pandocxnos.Target(
section_prefix + str(Ntargets), cursec, attrs.id
in targets)
for sublabel in sublabels:
Ntargets += 1
targets[sublabel] = pandocxnos.Target(
section_prefix + str(Ntargets), cursec, sublabel
in targets)
else: # Main equation has one number, increased by one
Ntargets += 1
targets[attrs.id] = pandocxnos.Target(
section_prefix + str(Ntargets), cursec, attrs.id
in targets)
for i, sublabel in enumerate(sublabels):
Nsubtargets = chr(ord('a') + i)
targets[sublabel] = pandocxnos.Target(
section_prefix + str(Ntargets) + Nsubtargets, cursec,
sublabel in targets)
return eq
def _process_table(value, fmt):
"""Processes the table. Returns a dict containing table properties."""
# pylint: disable=global-statement
global cursec # Current section being processed
global Nreferences # Number of refs in current section (or document)
global has_unnumbered_tables # Flags unnumbered tables were found
# Initialize the return value
table = {
'is_unnumbered': False,
'is_unreferenceable': False,
'is_tagged': False
}
# Bail out if there are no attributes
if len(value) == 5:
has_unnumbered_tables = True
table.update({'is_unnumbered': True, 'is_unreferenceable': True})
return table
# Parse the table
attrs = table['attrs'] = PandocAttributes(value[0], 'pandoc')
table['caption'] = value[1]
# Bail out if the label does not conform to expectations
if not LABEL_PATTERN.match(attrs.id):
has_unnumbered_tables = True
table.update({'is_unnumbered': True, 'is_unreferenceable': True})
return table
# Identify unreferenceable tables
if attrs.id == 'tbl:': # Make up a unique description
attrs.id = 'tbl:' + str(uuid.uuid4())
table['is_unreferenceable'] = True
# Update the current section number
if attrs['secno'] != cursec: # The section number changed
cursec = attrs['secno'] # Update the global section tracker
Nreferences = 1 # Resets the global reference counter
# Pandoc's --number-sections supports section numbering latex/pdf, html,
# epub, and docx
if numbersections:
if fmt in ['html', 'html5', 'epub', 'epub2', 'epub3', 'docx'] and \
'tag' not in attrs:
attrs['tag'] = str(cursec + secoffset) + '.' + str(Nreferences)
Nreferences += 1
# Save reference information
table['is_tagged'] = 'tag' in attrs
if table['is_tagged']:
# Remove any surrounding quotes
if attrs['tag'][0] == '"' and attrs['tag'][-1] == '"':
attrs['tag'] = attrs['tag'].strip('"')
elif attrs['tag'][0] == "'" and attrs['tag'][-1] == "'":
attrs['tag'] = attrs['tag'].strip("'")
references[attrs.id] = pandocxnos.Target(attrs['tag'], cursec, attrs.id
in references)
else: # ... then save the table number
references[attrs.id] = pandocxnos.Target(Nreferences, cursec, attrs.id
in references)
Nreferences += 1 # Increment the global reference counter
return table
def _is_theorem(item):
"""Returns True if item is a theorem; false otherwise."""
if item[0][0]['t'] != 'Span':
return False
attrs = PandocAttributes(item[0][0]['c'][0], 'pandoc')
return bool(LABEL_PATTERN.match(attrs.id)) if LABEL_PATTERN else False