Fixes #567 - Implement ods import

HISTORY.md

@@ -5,6 +5,8 @@
 ### Improvements
 
 - The html format now supports importing from HTML content (#243)
+- The ods format now supports importing from .ods files (#567). The support is
+  still a bit experimental.
 
 ### Changes
 

docs/formats.rst

@@ -145,12 +145,19 @@ If a title has been set, it will be exported as the table caption.
 ods
 ===
 
-Export data in OpenDocument Spreadsheet format. The ``ods`` format is currently
-export-only.
+Import/export data in OpenDocument Spreadsheet format.
+
+.. versionadded:: 3.6.0
+
+    Import functionality was added.
 
 This format is optional, install Tablib with ``pip install "tablib[ods]"`` to
 make the format available.
 
+The ``import_set()`` method also supports a ``skip_lines`` parameter that you
+can set to a number of lines that should be skipped before starting to read
+data.
+
 .. admonition:: Binary Warning
 
     :class:`Dataset.ods` contains binary data, so make sure to write in binary mode::

src/tablib/formats/_ods.py

@@ -2,10 +2,13 @@
 """
 
 import numbers
+from datetime import date, datetime
 from io import BytesIO
 
 from odf import opendocument, style, table, text
 
+import tablib
+
 bold = style.Style(name="bold", family="paragraph")
 bold.addElement(style.TextProperties(
     fontweight="bold",
@@ -49,6 +52,73 @@ def export_book(cls, databook):
         wb.save(stream)
         return stream.getvalue()
 
+    @classmethod
+    def import_sheet(cls, dset, sheet, headers=True, skip_lines=0):
+        """Populate dataset `dset` with sheet data."""
+
+        dset.title = sheet.getAttribute('name')
+
+        for i, row in enumerate(sheet.childNodes):
+            if i < skip_lines:
+                continue
+            row_vals = [cls.read_cell(cell) for cell in row.childNodes]
+            if i == skip_lines and headers:
+                dset.headers = row_vals
+            else:
+                if i > skip_lines and len(row_vals) < dset.width:
+                    row_vals += [''] * (dset.width - len(row_vals))
+                dset.append(row_vals)
+
+    @classmethod
+    def read_cell(cls, cell):
+        if not cell.childNodes:
+            if data := getattr(cell, 'data', None):
+                return data
+            value_type = cell.getAttribute('valuetype')
+            value = cell.getAttribute('value')
+            if value_type == 'float':
+                return float(value)
+            if value_type == 'date':
+                if 'T' in value:
+                    return datetime.strptime(value, "%Y-%m-%dT%H:%M:%S")
+                else:
+                    return datetime.strptime(value, "%Y-%m-%d").date()
+            else:
+                msg = f"value_type {value_type} not handled"
+                raise ValueError(msg)
+
+        for subnode in cell.childNodes:
+            value = cls.read_cell(subnode)
+            if value:
+                return value
+
+    @classmethod
+    def import_set(cls, dset, in_stream, headers=True, skip_lines=0):
+        """Populate dataset `dset` from ODS stream."""
+
+        dset.wipe()
+
+        ods_book = opendocument.load(in_stream)
+        for sheet in ods_book.spreadsheet.childNodes:
+            if sheet.qname[1] == 'table':
+                sheet = ods_book.spreadsheet.childNodes[0]
+                cls.import_sheet(dset, sheet, headers, skip_lines)
+
+    @classmethod
+    def import_book(cls, dbook, in_stream, headers=True):
+        """Populate databook `dbook` from ODS stream."""
+
+        dbook.wipe()
+
+        ods_book = opendocument.load(in_stream)
+
+        for sheet in ods_book.spreadsheet.childNodes:
+            if sheet.qname[1] != 'table':
+                continue
+            dset = tablib.Dataset()
+            cls.import_sheet(dset, sheet, headers)
+            dbook.add_sheet(dset)
+
     @classmethod
     def dset_sheet(cls, dataset, ws):
         """Completes given worksheet from given Dataset."""
@@ -66,6 +136,12 @@ def dset_sheet(cls, dataset, ws):
             for j, col in enumerate(row):
                 if isinstance(col, numbers.Number):
                     cell = table.TableCell(valuetype="float", value=col)
+                elif isinstance(col, datetime):
+                    cell = table.TableCell(
+                        valuetype="date", value=col.strftime('%Y-%m-%dT%H:%M:%S')
+                    )
+                elif isinstance(col, date):
+                    cell = table.TableCell(valuetype="date", value=col.strftime('%Y-%m-%d'))
                 else:
                     cell = table.TableCell(valuetype="string")
                     cell.addElement(text.P(text=str(col), stylename=style))

tests/test_tablib.py

@@ -1107,13 +1107,26 @@ def test_tsv_export(self):
 
 
 class ODSTests(BaseTestCase):
-    def test_ods_export_datatypes(self):
+    def test_ods_export_import_set(self):
+        date = datetime.date(2019, 10, 4)
         date_time = datetime.datetime(2019, 10, 4, 12, 30, 8)
-        data.append(('string', '004', 42, 21.55, Decimal('34.5'), date_time))
-        data.headers = ('string', 'start0', 'integer', 'float', 'decimal', 'date/time')
-        # ODS is currently write-only, just test that output doesn't crash.
-        assert data.ods is not None
-        assert len(data.ods)
+        data.append(('string', '004', 42, 21.55, Decimal('34.5'), date, date_time))
+        data.headers = ('string', 'start0', 'integer', 'float', 'decimal', 'date', 'date/time')
+        _ods = data.ods
+        data.ods = _ods
+        self.assertEqual(data.dict[0]['string'], 'string')
+        self.assertEqual(data.dict[0]['start0'], '004')
+        self.assertEqual(data.dict[0]['integer'], 42)
+        self.assertEqual(data.dict[0]['float'], 21.55)
+        self.assertEqual(data.dict[0]['decimal'], 34.5)
+        self.assertEqual(data.dict[0]['date'], date)
+        self.assertEqual(data.dict[0]['date/time'], date_time)
+
+    def test_ods_import_book(self):
+        ods_source = Path(__file__).parent / 'files' / 'book.ods'
+        with ods_source.open('rb') as fh:
+            dbook = tablib.Databook().load(fh, 'ods')
+        self.assertEqual(len(dbook.sheets()), 2)
 
 
 class XLSTests(BaseTestCase):