Improve handling of tables

legalmd/latex_renderer.py

@@ -148,9 +148,7 @@ class LaTeXRenderer(mistletoe.latex_renderer.LaTeXRenderer):
 		align = ['X' for col in token.column_align]
 		if hasattr(token, 'header'):
 			for i, col in enumerate(token.header.children):
-				if '[' in col.children[0].content.split(' ')[0]:
-					param = col.children[0].content[col.children[0].content.index('['):col.children[0].content.index(']')+1]
-					align[i] += param
+				align[i] += '[{}]'.format(col.weight)
 		
 		result = [
 			format('{\\footnotesize\\begin{longtabu} to \\dimexpr\\columnwidth-<lmarg>\\relax [r] {'),
@@ -170,24 +168,22 @@ class LaTeXRenderer(mistletoe.latex_renderer.LaTeXRenderer):
 		return ''.join(result)
 	
 	def render_table_header(self, token):
-		cells1 = ['{\\bfseries Column ' + self.render_inner(child).split(' ')[0].split('[')[0] + '}' for child in token.children]
-		cells2 = ['{\\bfseries ' + ' '.join(self.render_inner(child).split(' ')[1:]) + '}' for child in token.children]
+		cells1 = ['{\\bfseries Column ' + child.label + '}' for child in token.children]
+		cells2 = ['{\\bfseries ' + self.render_inner(child) + '}' for child in token.children]
 		return ' & '.join(cells1) + ' \\\\\\midrule\n' + ' & '.join(cells2) + ' \\\\'
 	
 	def render_table_row(self, token):
-		cells = [self.render_table_cell(child, colnum) for colnum, child in enumerate(token.children)]
+		cells = [self.render_table_cell(child) for child in token.children]
 		return ' & '.join(cells) + ' \\\\'
 	
-	def render_table_cell(self, token, colnum):
-		inner = self.render_inner(token)
-		
-		if colnum == 0:
+	def render_table_cell(self, token):
+		if token.colnum == 0 and token.label:
 			return format(r'\makebox[2.2em][l]{<rownum>.}<content>',
-				rownum=inner.split(' ')[0],
-				content=' '.join(inner.split(' ')[1:])
+				rownum=token.label,
+				content=self.render_inner(token)
 			)
 		else:
-			return inner
+			return self.render_inner(token)
 	
 	def render_document(self, token):
 		template = r'''

legalmd/legal_token.py

@@ -14,6 +14,7 @@
 #    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/>.
 
+from itertools import zip_longest
 import re
 
 import mistletoe
@@ -155,3 +156,60 @@ mistletoe.block_token.add_token(Definition)
 def convert_leading_tabs(string):
 	return string
 mistletoe.block_token.Quote.convert_leading_tabs = convert_leading_tabs
+
+class Table(mistletoe.block_token.BlockToken):
+	def __init__(self, lines):
+		if '---' in lines[1]:
+			self.column_align = [mistletoe.block_token.Table.parse_align(column)
+				for column in mistletoe.block_token.Table.split_delimiter(lines[1])]
+			self.header = TableRow(lines[0], 0, self.column_align)
+			self.children = [TableRow(line, rownum + 1, self.column_align) for rownum, line in enumerate(lines[2:])]
+		else:
+			self.column_align = [None]
+			self.children = [TableRow(line, rownum) for rownum, line in enumerate(lines)]
+	
+	@staticmethod
+	def start(line):
+		return '|' in line
+	
+	@staticmethod
+	def read(lines):
+		lines.anchor()
+		line_buffer = [next(lines)]
+		while lines.peek() is not None and '|' in lines.peek():
+			line_buffer.append(next(lines))
+		if len(line_buffer) < 2 or '---' not in line_buffer[1]:
+			lines.reset()
+			return None
+		return line_buffer
+
+class TableRow(mistletoe.block_token.BlockToken):
+	def __init__(self, line, rownum, row_align=None):
+		self.rownum = rownum
+		self.row_align = row_align or [None]
+		cells = filter(None, line.strip().split('|'))
+		self.children = [TableCell(x[0].strip() if x[0] else '', self, colnum, x[1])
+		                 for colnum, x in enumerate(zip_longest(cells, self.row_align))]
+		self.label = self.children[0].label
+
+class TableCell(mistletoe.block_token.BlockToken):
+	pattern1 = re.compile(r'\s*([0-9A-Za-z]+\s+)?(\[[0-9]+\])?(.*)')
+	
+	def __init__(self, inner, row, colnum, align=None):
+		self.row = row
+		self.colnum = colnum
+		self.align = align
+		
+		if self.colnum == 0 or self.row.rownum == 0:
+			label, weight, content = self.pattern1.match(inner).group(1, 2, 3)
+			
+			self.label = label.strip() if label else None
+			self.weight = int(weight[1:-1]) if weight else 1
+			self.children = mistletoe.span_token.tokenize_inner(content)
+		else:
+			self.label = None
+			self.weight = None
+			self.children = mistletoe.span_token.tokenize_inner(inner)
+
+mistletoe.block_token.remove_token(mistletoe.block_token.Table)
+mistletoe.block_token.add_token(Table)

legalmd/rtf_renderer.py

@@ -155,13 +155,8 @@ class RTFRenderer(mistletoe.base_renderer.BaseRenderer):
 		result = [format(r'{\sa0\trowd\trgaph120\trleft<lmarg>')]
 		
 		# Get column weights
-		weights = []
 		if hasattr(token, 'header'):
-			for col in token.header.children:
-				if '[' in col.children[0].content.split(' ')[0]:
-					weights.append(int(col.children[0].content[col.children[0].content.index('[')+1:col.children[0].content.index(']')]))
-				else:
-					weights.append(1)
+			weights = [col.weight for col in token.header.children]
 		else:
 			weights = [1 for col in token.column_align]
 		
@@ -184,24 +179,22 @@ class RTFRenderer(mistletoe.base_renderer.BaseRenderer):
 		return ''.join(result)
 	
 	def render_table_header(self, token):
-		cells1 = ['{\\b Column ' + self.render_inner(child).split(' ')[0].split('[')[0] + '\\intbl\\cell}' for child in token.children]
-		cells2 = ['{\\b ' + ' '.join(self.render_inner(child).split(' ')[1:]) + '\\intbl\\cell}' for child in token.children]
+		cells1 = ['{\\b Column ' + child.label + '\\intbl\\cell}' for child in token.children]
+		cells2 = ['{\\b ' + self.render_inner(child) + '\\intbl\\cell}' for child in token.children]
 		return ''.join(cells1) + '\\row' + ''.join(cells2) + '\\row '
 	
 	def render_table_row(self, token):
-		return ''.join([self.render_table_cell(cell, i) for i, cell in enumerate(token.children)]) + '\\row '
+		return ''.join([self.render_table_cell(cell) for cell in token.children]) + '\\row '
 	
-	def render_table_cell(self, token, colnum):
-		inner = self.render_inner(token)
-		
-		if colnum == 0:
+	def render_table_cell(self, token):
+		if token.colnum == 0 and token.label:
 			return format(r'{<rownum>. <content>\intbl\cell}',
-				rownum=inner.split(' ')[0],
-				content=' '.join(inner.split(' ')[1:])
+				rownum=token.label,
+				content=self.render_inner(token)
 			)
 		else:
 			return format(r'{<content>\intbl\cell}',
-				content=inner
+				content=self.render_inner(token)
 			)
 	
 	def render_thematic_break(self, token):