Merge pull request #175 from OpenDataServices/174-hashcomments

[#174] Add hashcomments command to xlsx

Author: Bjwebb

flattentool/input.py

@@ -574,7 +574,8 @@ def get_sheet_headings(self, sheet_name):
             sheet_configuration = {}
 
         skip_rows = sheet_configuration.get("skipRows", 0)
-        if sheet_configuration.get("ignore"):
+        if (sheet_configuration.get("ignore") or
+            (sheet_configuration.get("hashcomments") and sheet_name.startswith('#'))):
             # returning empty headers is a proxy for no data in the sheet.
             return []
 
@@ -617,7 +618,13 @@ def get_sheet_lines(self, sheet_name):
             header_row = worksheet.rows[skip_rows + configuration_line]
             remaining_rows = worksheet.rows[skip_rows + configuration_line + header_rows:]
 
-        coli_to_header = ({i: x.value for i, x in enumerate(header_row) if x.value is not None})
+        coli_to_header = {}
+        for i, header in enumerate(header_row):
+            if header.value is None:
+                continue
+            if sheet_configuration.get("hashcomments") and str(header.value).startswith('#'):
+                continue
+            coli_to_header[i] = header.value
         for row in remaining_rows:
             yield OrderedDict((coli_to_header[i], x.value) for i, x in enumerate(row) if i in coli_to_header)
 

flattentool/lib.py

@@ -15,4 +15,6 @@ def parse_sheet_configuration(configuration_list):
             configuration['headerRows'] = max(int(parts[1]), 1)
         if (len(parts) == 1 and parts[0].lower() == "ignore"):
             configuration['ignore'] = True
+        if (len(parts) == 1 and parts[0].lower() in ("hashcomments", "hashcomment")):
+            configuration['hashcomments'] = True
     return configuration

flattentool/tests/fixtures/xlsx/commands_hashcomments.xlsx

@@ -1240,3 +1240,20 @@ def test_commands_ignore(tmpdir):
     unflattened = json.load(tmpdir.join('command_single_unflattened.json'))
 
     assert unflattened == {'main': [{'actual': 'actual', 'headings': 'data', 'some': 'some'}]}
+
+def test_commands_hashcomments(tmpdir):
+
+    unflatten(
+        'flattentool/tests/fixtures/xlsx/commands_hashcomments.xlsx',
+        input_format='xlsx',
+        output_name=tmpdir.join('commands_hashcomments_unflattened.json').strpath,
+        cell_source_map=tmpdir.join('commands_hashcomments_source_map.json').strpath,
+        heading_source_map=tmpdir.join('commands_hashcomments_heading_source_map.json').strpath,
+        metatab_name='Meta',
+        metatab_vertical_orientation=True
+        )
+
+    unflattened = json.load(tmpdir.join('commands_hashcomments_unflattened.json'))
+
+    assert unflattened == {'main': [{'actual': 'actual', 'headings': 'data', 'some': 'some'}, {'actual': 'actual', 'headings': 'Other data', 'some': 'some'}],
+                           'some': 'data'}