Odfdo Release Notes
[3.13.9] - 2025-03-22
Technical update: migrate from poetry build environment to uv, and reorganize tests.
Changed
- Migrate from
poetrybuild environment touv - Organize tests with more fixtures and sub directories.
[3.13.8] - 2025-03-15
Technical update: update poetry to version 2.1, reformat pyproject.toml, move source code in the src directory.
Changed
- The source code is now in the
srcsub directory. - Build of the library requires
poetryversion 2.0 or more.
[3.13.7] - 2025-03-08
Improved filter on allowed characters in a table name (for compatibility with allowed names in LibreOffice). Forbidden characters are now \n \ / * ? : [ ], and "'" as first or last character. Previously forbidden characters were simply \n / \ "'".
Added lxml version 5.3.1 to the test suite.
Changed
- The regex for unauthised characters in table names is now:
r"^\'|[\n\\/\*\?:\][]|\'$".
Added
- Version 5.3.1 of
lxmladded totox.ini.
[3.13.6] - 2025-02-04
Fix corrupted ODF file due to incoherent manifest files, especially in .odg files (issue 54).
Fixed
- Fixed
Document.save()check formanifest.rdflisted inmanifest.xml
[3.13.5] - 2025-01-25
Add the option formatted to script odfdo-replace and method Element.replace() to interpret the <space>, <tab> and <CR> of the replacement string.
Added
- Add
Element.replace()optional argumentformatted - Add
odfdo-replaceoptional option--formatted
[3.13.4] - 2025-01-22
- Fix a bug on
Document.insert_style()and related methods that prevented the loading of styles from the familynumber:*, likenumber:currency-style(issue #53).
Fixed
- Fixed
Document.insert_style()andCONTEXT_MAPPINGinstyles.pyfor a better detection of styles pseudo-families.
[3.13.3] - 2025-01-01
- Refactor cell properties to get easier and more consistent access to cell values. Some new properties added: Cell.decimal, Cell.int, Cell.bool, Cell.duration, Cell.datetime, Cell.date.
- Add a
.rdfmanifest file if it is missing (this should only happen in very rare manipulations).
Changed
- Refactoring of Cell properties.
- Refactoring of Container class to ensure creation of manifest file.
Added
- Add properties:
Cell.decimal,Cell.int,Cell.bool,Cell.duration,Cell.datetime,Cell.date.
[3.13.2] - 2024-12-15
- Refactor Table and Row caching, removing useless
_cachingattribute from many classes. - Remove module scriptutils.
Changed
- Refactoring of Table and Row caching.
- Replace in
container.pyand scriptsshow.pyandstyles.pythe previous functions of scriptutils.
Removed
odfdo.scriptutils.pyremoved.
[3.13.1] - 2024-12-14
- Allow XML export of base64 encoded images (preparing for flat ODF export).
- Update XML propertires to ODF 1.2.
Changed
- Refactoring of Document.add_file() and export to XML format.
[3.13.0] - 2024-12-07
- The
Metaclass which manages themeta.xmlpart has two new methodsas_dict()andas_json()to export its content. - Improved “pretty” export of documents.
Added
- Add methods:
Meta.as_dict(),Meta.as_json(),MetaTemplate.as_dict(),MetaAutoReload.as_dict(),MetaHyperlinkBehaviour.as_dict().
Changed
- Small XML file formatting changes when saving with “pretty=True”.
[3.12.1] - 2024-11-30
Fix some small rendering issues for Markdown export.
Fixed
- Better Markdown export for strike style, non break space, successive tags, line breaks, footnotes
[3.12.0] - 2024-11-30
- Change in
str(Paragraph)which now includes a'\n'at the end of the string. - The
odfdo-to-mdscript is renamed toodfdo-markdownand should be functional. Markdown export of .odt files supports all standard formatting features (including tables) except quoted text (no clear semantic equivalent in the ODF standard). - Improved
__str__methods for many classes: Document.body, Paragraph, Span, Link, Unit, Note, Annotation. - Some added methods:
Document.get_parent_style(),Document.get_list_style(),Style.get_list_style_properties(),Style.get_text_properties(). - The new
Element.inner_textproperty is now the preferred way to access an element’s inner text.
Added
- Add methods:
Document.get_parent_style(),Document.get_list_style(),Style.get_list_style_properties(),Style.get_text_properties(). - Add propterty
Element.inner_text.
Changed
- Script
odfdo-to-mdrenamed toodfdo-markdown. str(Paragraph)now includes a'\n'at the end of the string.- Output of the str method modified for many elements.
[3.11.0] - 2024-11-23
- New script
odfdo-to-mdto export text document in markdown format to stdout (experimental, do not export images links neither tables). - Fix
VarTimeinitialization: class can now be initialized without mandatory time argument.
Added
- Add script
odfdo-to-md.
Changed
odfdo-folderscript now writes XML files with the “pretty” option by default.
Fixed
- Fix
VarTimeinitialization.
[3.10.1] - 2024-11-23
The HTML documentation in /doc (mostly auto generated) contains now all recipes, sorted by relevance.
Changed
- Improvement of documentation.
[3.10.0] - 2024-11-23
- Fix a bug of
Paragraph.set_span()when using an offset argument of zero (the Span was not created). Added 3 methods related to searching strings in paragraphs:search_first(),search_all()andtext_at(). These methods permit to search some string with regex in a paragraph and get their position,text_at()returns the text content at a given position. - Fix the “pretty” option of
Document.save(). “pretty” is now the default for odfdo-folder.
Added a new recipe showing several methods to change the style of a paragraph or words in a pragraph with the use of Paragraph.style = style.name and Paragraph.set_span().
Added
- Added
Element.search_first(),Element.search_all(),Element.text_at(). - Added
change_paragraph_styles_or_spans.pyrecipe (issue #21).
Changed
odfdo-folderscript now writes XML files with the “pretty” option by default.
Fixed
- Fix
Paragraph.set_span()when using an offset argument of zero (issue #21). - Fix the “pretty” option of
Document.save()(issue #28).
[3.9.4] - 2024-11-06
Fix a performance bug on huge .ods tables when number of rows is a large (several thousand). See issue #46 for a table of about ~83k. Table.traverse() on such a table is expected to be ~2 sec.
Changed
- Rewrite the method Table.traverse().
Fixed
- Fix the performance bug on huge .ods tables (issue #46).
[3.9.3] - 2024-10-14
Add support for Python 3.13 final in test suite.
Added
- Add support for Python3.13 in tox.ini
[3.9.2] - 2024-10-05
Add support for Python 3.13.0.rc3 in test suite.
Added
- Add support for Python3.13.0.rc3 in tox.ini, add requirement for lxml version 5.3 or higher for Python 3.13.
[3.9.1] - 2024-09-29
When creating a Document() allow alias “odt” for “Text”, “ods” for “spreadsheet”.
Add a recipe showing how to remove parts from a text document.
Added
-
Aliases “odt”, “ods”, “odp” and “odg” for Document creation.
-
Add recipe
delete_parts_of_a_text_document.py.
[3.9.0] - 2024-09-22
Two changes in this version:
- Fix of the broken
Table.displayedproperty. - Fix the way spaces are represented for better compliance with the ODF standard and word processors.
The Table.displayed property was broken and is removed. The functionality is replaced by the Document.get_table_displayed and Document.set_table_displayed methods. This change should not affect anyone since the previous implementation was unusable.
In previous version 3 spaces were translated into 1 space followed by '<text:s text:c="2"/>' unconditionally. However, the standard specifies that at the beginning and end of a paragraph spaces must be discarded by word processors, so 3 spaces should be coded '<text:s text:c="3"/>' and a single space as '<text:s/>'. This change should fix the bug of “disappearing” spaces at the beginning of paragraphs.
Added
-
Methods Document.get_table_displayed(), Document.set_table_displayed(), Document.get_table_style().
-
The Spacer() class has 2 new properties: Spacer.length and Spacer.text.
Changed
-
XML generation of spaces at beginning and end of Paragraph content.
-
Update of dependency versions.
Removed
- Table.displayed property.
Fixed
- Fix the “disappearing” spaces at the beginning of paragraphs bug.
[3.8.0] - 2024-08-25
Changed the default behavior for appending text to a Paragraph: the behavior of the Paragraph.append_plain_text() method is now the default. A "formatted" argument is added, True by default, which applies the recognition of “\n”, “\t” or a sequence of several spaces and converts them to ODF tags (text:line-break, text:tab, text:s)). To ignore this text formatting, set "formatted=False".
This change affects you if you create paragraphs from text containing line breaks or tabs and you don’t want them to appear. In this case, add the argument "formatted=False"
Details:
-
Paragraph("word1 word2")-
previous behavior:
- product XML:
'<text:p>word1 word2</text:p>' - expected display:
word1 word2(single space, the ODF standard does not recognize space sequences)
- product XML:
-
new behavior:
- product XML:
'<text:p>word1 <text:s text:c="4"/>word2</text:p>' - expected display:
word1 word2(5 spaces)
- product XML:
-
-
Paragraph("word1 word2", formatted=False)- new behavior:
- product XML:
'<text:p>word1 word2</text:p>' - expected display:
word1 word2
- product XML:
- new behavior:
-
Paragraph("word1\nword2")-
previous behavior:
- product XML:
'<text:p>word1\nword2</text:p>' - expected display:
word1 word2(single space, the ODF standard does not recognize “\n” in XML content)
- product XML:
-
new behavior:
- product XML:
'<text:p>word1<text:line-break/>word2</text:p>' - expected display:
word1 word2
- product XML:
-
-
Paragraph("word1\nword2", formatted=False)- new behavior:
- product XML:
'<text:p>word1 word2</text:p>' - expected display:
word1 word2
- product XML:
- new behavior:
On the same principle the "formatted" argument is available for Pararaph.append(text), Header(text), Span(text).
The Paragraph.append_plain_text(text) method is retained for compatibility with previous versions and has the same behavior as Paragraph.append(text, formatted=True), the default.
Changed
-
Paragraph(),Paragraph.append()and subclassesHeader()andSpan()have a new"formatted"argument True by default that translates into ODF format “\n”, “\t” and multiples spaces. -
Updating dependency versions.
[3.7.13] - 2024-08-17
- Fix parsing of Date and Datetime for a better compliance with ISO8601.
Changed
- Updating dependency versions.
Fixed
-
Fix datetime encoding/decoding for ISO8601 compliance and different Python versions.
-
Move from strptime() to date.isoformat() for class Date and DateTime.
[3.7.12] - 2024-08-11
- Update dependencies and test suite, support of
lxmlversion 5.3.0.
Changed
- Updating dependency versions.
Fixed
-
Fix a type hint in element.py
-
Fix missing .venv in gitconfig
[3.7.11] - 2024-05-25
- New script
odfdo-userfieldto show or set the user-field content in an ODF file.
Added
- Add script
odfdo-userfield.
Changed
- Updating dependency versions.
[3.7.10] - 2024-05-4
-
Refactor to add property getter for some common methods. Original get_* method is still available and permits detailed requests with parameters.
- Body.tables -> Body.get_tables() - Element.tocs -> Element.get_tocs() - Element.toc -> Element.get_toc() - Element.text_changes -> Element.get_text_changes() - Element.tracked_changes -> Element.get_tracked_changes() - Element.user_defined_list -> Element.get_user_defined_list() - Element.images -> Element.get_images() - Element.frames -> Element.get_frames() - Element.lists -> Element.get_lists() - Element.headers -> Element.get_headers() - Element.spans -> Element.get_spans() - Element.paragraphs -> Element.get_paragraphs() - Element.sections -> Element.get_sections() - Table.rows -> Table.get_rows() - Table.cells -> Table.get_cells() - Table.columns -> Table.get_columns() - Row.cells -> Row.get_cells() - Document.parts -> Document.get_parts() - Container.parts -> Container.get_parts() -
Refactor to add property getter/setter for some common methods. Original get_* and set_* methods are still available and permit detailed requests with parameters.
- Column.default_cell_style -> Column.get/set_default_cell_style()
Added
- Added
Body.tables - Added
Element.tocs - Added
Element.toc - Added
Element.text_changes - Added
Element.tracked_changes - Added
Element.images - Added
Element.frames - Added
Element.lists - Added
Element.headers - Added
Element.spans - Added
Element.paragraphs - Added
Element.sections - Added
Column.default_cell_style - Added
Table.rows - Added
Table.cells - Added
Table.columns - Added
Row.cells - Added
Document.parts - Added
Container.parts
[3.7.9] - 2024-05-3
-
Refactor the Body access methods, creating relevant a Body class and related sub-classes. Moved some access method from the Element class to relevant Body sub-classes.
-
Refactor metadata methods to permit access throuh @property (the legacy get_* and set_* methods are still available).
-
Added a few metadata elements from the ODF standard (hyperlink-behaviour, auto-reload, template, print-dateprinted-by)
Added
- Added
MetaAutoReloadclass - Added
MetaHyperlinkBehaviourclass - Added
MetaTemplateclass - Added
DcCreatorMixinclass - Added
DcDateMixinclass - Added
Bodyclass - Added
Chartclass - Added
Databaseclass - Added
Drawingclass - Added
Imageclass - Added
Presentationclass - Added
Spreadsheetclass - Added
Textclass (renaming the previous internalTextclass toEText)
[3.7.8] - 2024-05-2
Fix embedded chart analysis in documents, see recipe change_values_of_a_chart_inside_a_document.py.
Added
- Added
change_values_of_a_chart_inside_a_document.pyrecipe
Changed
-
The “pretty” setting when saving the file always defaults to False. This setting should only be used for debugging purposes
-
meta.generatorcan be used via a @property accessor -
(Internal change) move body() definition to xmlpart
-
(Internal change) refactoring for future XML feature
Fixed
-
Fix parsing of Table when parent uses “table:table-rows” kind of wrapper
-
Fix a bug when a Cell contains the valid ‘NaN’ Decimal number
[3.7.7] - 2024-04-1
Improvement of the lxml dependency support.
Added
-
Added a
CHANGES.mdfile -
Automatic tests for ubuntu-latest, macos-latest, windows-latest
Changed
-
Now supports a wider range of
lxmlversions:-
python 3.9: lxml version 4.8.0 to 4.9.4
-
python 3.10: lxml version 4.8.0 to 5.1.1
-
python 3.11: lxml version 4.9.4 to 5.2.0 and beyond
-
python 3.12: lxml version 4.9.4 to 5.2.0 and beyond
-
-
autogenerated documentation now uses
mkdocs
Fixed
-
Use
sys.executableto ensure all tests can pass in a github virtualenv on Windows. -
Remove import of
lxmlinternal\_ElementUnicodeResultand\_ElementUnicodeResultclasses.
[3.7.6] - 2024-03-30
Quick fix for the crash with new lxml version 5.1.1
Fixed
- Fix crash with `lxml` 5.1.1 by restricting version do 5.1.0
[3.7.5] - 2024-03-23
Add the method get_cell_background_color to retrieve the background color of a cell in a table.
Added
-
Tables: some users need to easily access the background color of cells, including cells without “value” content. That was requiring a complex parsing of styles. So a new method:
Document.get_cell_background_color(sheet_id, cell_coords). -
See the corresponding recipe
recipes/get_cell_background_color.pyfor an exemple of usage. -
Tables: (related to previous). It is often useful to reduce the table size before working on it, especially if styles apply to whole rows. A method called
Table.rstrip()already permitted to remove empty bottom rows and empty right columns. However, aCellmays have no value but a style (color background for example), andrstrip()was removing such cells. So an new clever method is provided:Table.optimize_width()that shrink the table size, still keeping styled empty cells. -
To test the actual result of this method, you can use the new script
odfdo-table-shrinkwhich is basically a wrapper upon this method. (Note: all this stuff aims to facilitate some feature for the related github projectodsparsator). -
repr()method forCell,RowandColumn. -
Ancillary methods related to above features.
Fixed
Document(path)now accepts astrpath starting with~as the path relative to the user home.
Changed
-
Tables: (related to previous), change the
Cell.is_empty()test. A cell is now considered as not empty if part of aspan(a cell spanned on several rows or columns). This may induce some changes for parsing scripts. Before that, only the first cell of the span (which actually contains the value) was considered as non empty. Now other cells of the span are not empty (but contain a null value). -
Minor refactor of code, version updates of dependencies.
[3.7.4] - 2024-03-17
Add a recipe as example of programmatically setting text styles for headers and paragraphs, with basic font and color properties.
Added
-
Add recipe
create_basic_text_styles. -
All style fields related to color accept a color name from the CSS list of color.
Changed
- Updating dependency versions.
[3.7.3] - 2024-03-10
Internal maintenance release.
Fixed
- Fix logo link on
Pypipage.
Changed
-
Technical updates from
optparsetoargparse. -
Updating dependency versions.
[3.7.2] - 2024-03-9
Internal maintenance release.
Changed
-
Use
pdocfor autogenerated documentation. -
Refactor some recipes to use them in a test suit.
-
Code refactor, Updating dependency versions.
[3.7.1] - 2024-03-3
Minor performance improvement of script odfdo-headers.
Changed
- Use better algorithm for script
odfdo-headers.
[3.7.0] - 2024-03-2
New script odfdo-headers to print the headers of a ODF file.
Added
- Add script
odfdo-headers.
Changed
- Updating dependency versions.
[3.6.0] - 2024-02-25
New script odfdo-highlight to highlight the text matching a pattern (regex) in a ODF file.
Added
- Add script
odfdo-highlight.
Changed
- Updating dependency versions.
[3.5.1] - 2024-02-20
Fix the update method of Table of Content and add a recipe to show how to update a TOC.
Added
- Add recipe
update_a_text_document_with_a_table_of_content.
Changed
-
Refactor of TOC related code.
-
Updating dependency versions.
[3.5.0] - 2024-01-27
2024 release, updated ODF templates and better test suit.
Changed
-
Update
ODFtemplates. -
Refactor many Python files for use of type hints.
-
Updates for year 2024, updating dependency versions.
[3.4.7] - 2024-01-15
Updade to lxml version 5.
Changed
- Update `lxml` from version 4 to 5.
[3.4.6] - 2023-12-25
Add script odfdo-replace to find a pattern (regex) in an ODF file and replace by some string.
Fixed
- Fix reading content from a
BytesIO.
Changed
- Add script
odfdo-replace.
[3.4.5] - 2023-12-24
Add recipes showing how to save/read document from io.BytesIO.
Added
- Add recipes
read_document_from_bytesio.pyandsave_document_as_bytesio.py.
Changed
- Refactoring of code.