Sort tables extracted on a page by their top position

Fixes #336
h/t @gqh1995 for reporting

pdfplumber/table.py

@@ -450,6 +450,11 @@ def __init__(self, page, settings={}):
         )
         self.cells = intersections_to_cells(self.intersections)
         self.tables = [Table(self.page, t) for t in cells_to_tables(self.cells)]
+        # If multiple tables are found, sort them by their vertical placement.
+        # Top to bottom, left to right.
+        self.tables = sorted(
+            self.tables, key=lambda table: (table.bbox[1], table.bbox[0])
+        )
 
     def get_edges(self):
         settings = self.settings

tests/test_table.py

@@ -84,3 +84,15 @@ def test_explicit_desc_decimalization(self):
     def test_text_without_words(self):
         assert table.words_to_edges_h([]) == []
         assert table.words_to_edges_v([]) == []
+
+    def test_order(self):
+        """
+        See issue #336
+        """
+        path = os.path.join(HERE, "pdfs/issue-336-example.pdf")
+        with pdfplumber.open(path) as pdf:
+            tables = pdf.pages[0].extract_tables()
+            assert len(tables) == 3
+            assert len(tables[0]) == 8
+            assert len(tables[1]) == 11
+            assert len(tables[2]) == 2