Cyclic0007/python-markdownify
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

allow tables with headers in first (or any) column

Browse Source
This commit is contained in:
AlexVonB
2021-05-17 12:36:48 +02:00
parent 7dac92e85e
commit e6da15c173
2 changed files with 30 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
markdownify/__init__.py
View File

@@ -281,20 +281,19 @@ class MarkdownConverter(object):
text_data = [] text_data = []
rendered_header = False rendered_header = False
for row in rows: for row in rows:
headers = row.find_all('th') cells = row.find_all(['td', 'th'])
columns = row.find_all('td') is_headrow = all([cell.name == 'th' for cell in cells])
if not rendered_header and len(headers) > 0: texts = [cell.text.strip() for cell in cells]
headers = [head.text.strip() for head in headers] if not rendered_header and is_headrow:
text_data.append('| ' + ' | '.join(headers) + ' |') text_data.append('| ' + ' | '.join(texts) + ' |')
text_data.append('| ' + ' | '.join(['---'] * len(headers)) + ' |') text_data.append('| ' + ' | '.join(['---'] * len(cells)) + ' |')
rendered_header = True rendered_header = True
elif len(columns) > 0: elif len(cells) > 0:
if not rendered_header: if not rendered_header:
text_data.append('| ' + ' | '.join([''] * len(columns)) + ' |') text_data.append('| ' + ' | '.join([''] * len(cells)) + ' |')
text_data.append('| ' + ' | '.join(['---'] * len(columns)) + ' |') text_data.append('| ' + ' | '.join(['---'] * len(cells)) + ' |')
rendered_header = True rendered_header = True
columns = [colm.text.strip() for colm in columns] text_data.append('| ' + ' | '.join(texts) + ' |')
text_data.append('| ' + ' | '.join(columns) + ' |')
else: else:
continue continue
return '\n'.join(text_data) return '\n'.join(text_data)

20
tests/test_conversions.py
View File

@@ -60,6 +60,25 @@ table = """<table>
</table>""" </table>"""
table_with_header_column = """<table>
<tr>
<th>Firstname</th>
<th>Lastname</th>
<th>Age</th>
</tr>
<tr>
<th>Jill</th>
<td>Smith</td>
<td>50</td>
</tr>
<tr>
<th>Eve</th>
<td>Jackson</td>
<td>94</td>
</tr>
</table>"""
table_head_body = """<table> table_head_body = """<table>
<thead> <thead>
<tr> <tr>
@@ -335,6 +354,7 @@ def test_div():
def test_table(): def test_table():
assert md(table) == '| Firstname | Lastname | Age |\n| --- | --- | --- |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |' assert md(table) == '| Firstname | Lastname | Age |\n| --- | --- | --- |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |'
assert md(table_with_header_column) == '| Firstname | Lastname | Age |\n| --- | --- | --- |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |'
assert md(table_head_body) == '| Firstname | Lastname | Age |\n| --- | --- | --- |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |' assert md(table_head_body) == '| Firstname | Lastname | Age |\n| --- | --- | --- |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |'
assert md(table_missing_text) == '| | Lastname | Age |\n| --- | --- | --- |\n| Jill | | 50 |\n| Eve | Jackson | 94 |' assert md(table_missing_text) == '| | Lastname | Age |\n| --- | --- | --- |\n| Jill | | 50 |\n| Eve | Jackson | 94 |'
assert md(table_missing_head) == '| | | |\n| --- | --- | --- |\n| Firstname | Lastname | Age |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |' assert md(table_missing_head) == '| | | |\n| --- | --- | --- |\n| Firstname | Lastname | Age |\n| Jill | Smith | 50 |\n| Eve | Jackson | 94 |'