Merge pull request #189 from camelot-dev/fix-179

[MRG] Prevent taking max of an empty set

camelot/parsers/stream.py

@@ -121,6 +121,7 @@ class Stream(BaseParser):
         row_y = 0
         rows = []
         temp = []
+
         for t in text:
             # is checking for upright necessary?
             # if t.get_text().strip() and all([obj.upright for obj in t._objs if
@@ -131,7 +132,9 @@ class Stream(BaseParser):
                     temp = []
                     row_y = t.y0
                 temp.append(t)
+
         rows.append(sorted(temp, key=lambda t: t.x0))
+        if len(rows) > 1:
             __ = rows.pop(0)  # TODO: hacky
         return rows
 
@@ -345,6 +348,9 @@ class Stream(BaseParser):
         else:
             # calculate mode of the list of number of elements in
             # each row to guess the number of columns
+            if not len(elements):
+                cols = [(text_x_min, text_x_max)]
+            else:
                 ncols = max(set(elements), key=elements.count)
                 if ncols == 1:
                     # if mode is 1, the page usually contains not tables

tests/test_cli.py

@@ -160,8 +160,8 @@ def test_cli_output_format():
 
 def test_cli_quiet():
     with TemporaryDirectory() as tempdir:
-        infile = os.path.join(testdir, "blank.pdf")
+        infile = os.path.join(testdir, "empty.pdf")
-        outfile = os.path.join(tempdir, "blank.csv")
+        outfile = os.path.join(tempdir, "empty.csv")
         runner = CliRunner()
 
         result = runner.invoke(

tests/test_errors.py

@@ -55,15 +55,33 @@ def test_image_warning():
             )
 
 
-def test_no_tables_found():
+def test_lattice_no_tables_on_page():
-    filename = os.path.join(testdir, "blank.pdf")
+    filename = os.path.join(testdir, "empty.pdf")
     with warnings.catch_warnings():
         warnings.simplefilter("error")
         with pytest.raises(UserWarning) as e:
-            tables = camelot.read_pdf(filename)
+            tables = camelot.read_pdf(filename, flavor="lattice")
         assert str(e.value) == "No tables found on page-1"
 
 
+def test_stream_no_tables_on_page():
+    filename = os.path.join(testdir, "empty.pdf")
+    with warnings.catch_warnings():
+        warnings.simplefilter("error")
+        with pytest.raises(UserWarning) as e:
+            tables = camelot.read_pdf(filename, flavor="stream")
+        assert str(e.value) == "No tables found on page-1"
+
+
+def test_stream_no_tables_in_area():
+    filename = os.path.join(testdir, "only_page_number.pdf")
+    with warnings.catch_warnings():
+        warnings.simplefilter("error")
+        with pytest.raises(UserWarning) as e:
+            tables = camelot.read_pdf(filename, flavor="stream")
+        assert str(e.value) == "No tables found in table area 1"
+
+
 def test_no_tables_found_logs_suppressed():
     filename = os.path.join(testdir, "foo.pdf")
     with warnings.catch_warnings():
@@ -77,7 +95,7 @@ def test_no_tables_found_logs_suppressed():
 
 
 def test_no_tables_found_warnings_suppressed():
-    filename = os.path.join(testdir, "blank.pdf")
+    filename = os.path.join(testdir, "empty.pdf")
     with warnings.catch_warnings():
         # the test should fail if any warning is thrown
         warnings.simplefilter("error")