Briq
/
camelot-py
6
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Add baseline test for hybrid 

Fix first split merge issue
pull/153/head
Frh 2020-05-04 17:41:57 -07:00
parent 77d289bd86
commit 79ea4adcd1
2 changed files with 12 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
camelot/parsers/hybrid.py
View File

@ -149,6 +149,11 @@ class Hybrid(BaseParser):
# Our boundary is fully after the split, move on # Our boundary is fully after the split, move on
idx_boundaries = idx_boundaries - 1 idx_boundaries = idx_boundaries - 1
previous_boundary = boundary previous_boundary = boundary
if idx_boundaries < 0:
# If this is the last boundary to the left, set its
# edge at the split
boundary[0] = split
idx_splits = idx_splits - 1
else: else:
# The split is inside our boundary: split it # The split is inside our boundary: split it
new_boundary = [split, boundary[1]] new_boundary = [split, boundary[1]]

7
tests/test_common.py
View File

@ -286,6 +286,13 @@ def test_network_layout_kwargs():
# Hybrid parser # Hybrid parser
def test_hybrid():
df = pd.DataFrame(data_stream)
filename = os.path.join(testdir, "health.pdf")
tables = camelot.read_pdf(filename, flavor="hybrid")
assert_frame_equal(df, tables[0].df)
def test_hybrid_vertical_header(): def test_hybrid_vertical_header():
"""Tests a complex table with a vertically text header. """Tests a complex table with a vertically text header.
""" """