An open-source document parsing model with SOTA accuracy and speed, using an RSI loop during post-training.
from oxr.pipeline.models import Block, BlockType, ContentFormat
from oxr.pipeline.postprocess.table_postprocess import postprocess_table
from oxr.utils.table import html_table_to_markdown
def _table_block() -> Block:
return Block(0, BlockType.TABLE, [0, 0, 10, 10], "", ContentFormat.TEXT)
def test_html_table_to_markdown_expands_rowspan_and_colspan():
html = """
<table>
<tr><th rowspan="2">Name</th><th colspan="2">Score</th></tr>
<tr><td>Math</td><td>English</td></tr>
<tr><td>Ada</td><td>98</td><td>96</td></tr>
</table>
"""
assert html_table_to_markdown(html) == "\n".join([
"| Name | Score | Score |",
"| --- | --- | --- |",
"| Name | Math | English |",
"| Ada | 98 | 96 |",
])
def test_html_table_to_markdown_escapes_pipes_and_strips_fences():
html = """```html
<table><tr><th>A|B</th></tr><tr><td>x|y</td></tr></table>
```"""
assert html_table_to_markdown(html) == "\n".join([
"| A\\|B |",
"| --- |",
"| x\\|y |",
])
def test_html_table_to_markdown_handles_thead_tbody_and_th():
html = """
<table>
<thead>
<tr><th>City</th><th>Population</th></tr>
</thead>
<tbody>
<tr><td>Shanghai</td><td>24M</td></tr>
<tr><td>Beijing</td><td>21M</td></tr>
</tbody>
</table>
"""
assert html_table_to_markdown(html) == "\n".join([
"| City | Population |",
"| --- | --- |",
"| Shanghai | 24M |",
"| Beijing | 21M |",
])
def test_html_table_to_markdown_falls_back_to_cleaned_text():
assert html_table_to_markdown("<p>not a table</p>") == "not a table"
def test_postprocess_table_sets_markdown_format_when_requested():
block = postprocess_table(
_table_block(),
"<table><tr><th>A</th></tr><tr><td>B</td></tr></table>",
output_format="markdown",
)
assert block.format == ContentFormat.MARKDOWN
assert block.content == "| A |\n| --- |\n| B |"