Add ODS support

ODS support is based on ezodf [1] library.

It looks, that ezodf has a bug [2], where number of sheet rows and cells are
calculated incorrectly. Probably this is the reason, why I get extra rows and
cols in tests.

[1] https://github.com/T0ha/ezodf
[2] T0ha/ezodf#12

See: frictionlessdata#28

README.md

@@ -10,7 +10,7 @@ Consistent interface for stream reading and writing tabular data (csv/xls/json/e
 
 ## Features
 
-- supports various formats: csv/tsv/xls/xlsx/json/ndjson/native/etc
+- supports various formats: csv/tsv/xls/xlsx/json/ndjson/ods/native/etc
 - reads data from variables, filesystem or Internet
 - streams data instead of using a lot of memory
 - processes data via simple user processors

setup.py

@@ -30,6 +30,8 @@ def read(*paths):
  'linear-tsv>=1.0,<2.0a',
  'unicodecsv>=0.14,<1.0a',
  'jsonlines>=1.1,<1.2',
+ 'ezodf>=0.3,<0.4',
+ 'lxml', # required by ezodf
 ]
 TESTS_REQUIRE = [
  'pylama',

tabulator/config.py

@@ -31,6 +31,7 @@
  'tsv': 'tabulator.parsers.tsv.TSVParser',
  'xls': 'tabulator.parsers.excel.ExcelParser',
  'xlsx': 'tabulator.parsers.excelx.ExcelxParser',
+ 'ods': 'tabulator.parsers.ods.ODSParser',
 }
 
 WRITERS = {

tabulator/parsers/ods.py

@@ -0,0 +1,67 @@
+# -*- coding: utf-8 -*-
+from __future__ import division
+from __future__ import print_function
+from __future__ import absolute_import
+from __future__ import unicode_literals
+
+import ezodf
+from six import BytesIO
+
+from .. import helpers
+from . import api
+
+
+# Module API
+
+class ODSParser(api.Parser):
+ """Parser to parse ODF Spreadsheets.
+
+ Args:
+ sheet (int or str): sheet number or name
+ First sheet's number is 1.
+
+ """
+
+ # Public
+
+ options = [
+ 'sheet',
+ ]
+
+ def __init__(self, sheet=1):
+ self.__index = sheet - 1 if isinstance(sheet, int) else sheet
+ self.__loader = None
+ self.__bytes = None
+ self.__book = None
+ self.__sheet = None
+ self.__extended_rows = None
+
+ @property
+ def closed(self):
+ return self.__bytes is None or self.__bytes.closed
+
+ def open(self, source, encoding, loader):
+ self.close()
+ self.__loader = loader
+ self.__bytes = loader.load(source, encoding, mode='b')
+ self.__book = ezodf.opendoc(BytesIO(self.__bytes.read()))
+ self.__sheet = self.__book.sheets[self.__index]
+ self.reset()
+
+ def close(self):
+ if not self.closed:
+ self.__bytes.close()
+
+ def reset(self):
+ helpers.reset_stream(self.__bytes)
+ self.__extended_rows = self.__iter_extended_rows()
+
+ @property
+ def extended_rows(self):
+ return self.__extended_rows
+
+ # Private
+
+ def __iter_extended_rows(self):
+ for number, row in enumerate(self.__sheet.rows(), start=1):
+ yield number, None, [cell.value for cell in row]

tests/parsers/test_ods.py

@@ -0,0 +1,51 @@
+# -*- coding: utf-8 -*-
+from __future__ import division
+from __future__ import print_function
+from __future__ import absolute_import
+from __future__ import unicode_literals
+
+import io
+from mock import Mock
+from tabulator import Stream
+from tabulator.parsers.ods import ODSParser
+
+
+# Tests
+
+def test_excelx_parser():
+
+ source = 'data/table.ods'
+ encoding = None
+ loader = Mock()
+ loader.load = Mock(return_value=io.open(source, 'rb'))
+ parser = ODSParser()
+
+ assert parser.closed
+ parser.open(source, encoding, loader)
+ assert not parser.closed
+
+ assert list(parser.extended_rows) == [
+ (1, None, ['id', 'name', None]),
+ (2, None, [1.0, 'english', None]),
+ (3, None, [2.0, '中国人', None]),
+ (4, None, [None, None, None]),
+ (5, None, [None, None, None]),
+ ]
+
+ assert len(list(parser.extended_rows)) == 0
+ parser.reset()
+ assert len(list(parser.extended_rows)) == 5
+
+ parser.close()
+ assert parser.closed
+
+
+def test_stream_ods():
+ with Stream('data/table.ods', headers=1) as stream:
+ assert stream.headers == ['id', 'name', None]
+ assert stream.read(keyed=True) == [
+ {'id': 1.0, 'name': 'english', None: None},
+ {'id': 2.0, 'name': '中国人', None: None},
+ {'id': None, 'name': None, None: None},
+ {'id': None, 'name': None, None: None},
+ ]