v0.10.0

New features:

• connect method has attachments argument to give the option to include 'attachment' files like media files and canvas files in the graph. The behaviour from v0.9 is kept in this new release (via the default attachments=False).
• Information about media files & their filepaths is stored in the media_file_index attribute.
• New methods for metadata: get_canvas_file_metadata and get_media_file_metadata. The get_all_file_metadata method is new method that is best-placed to get all the metadata for files & notes in a vault.
• isolated_media_files and isolated_canvas_files attributes.
• nonexistent_media_files and nonexistent_media_files attributes.

Important API changes vs previous version:

• file_index attribute is now md_file_index, to avoid ambiguity from the extra support now for media files and canvas files.

Other improvements:

• Speed improvements for the gather() method and processing of HTML content.
• Tweaks to the code to address deprecation warnings from other packages.

v0.9.0

New features:

• Support for canvas files (the latest addition to Obsidian). See the Canvas files features notebook for detail on functionality.
• Nested tags are now supported. By default, Vault will not account for nested tags (as has previously been the case).
• Column for n_tags in get_note_metadata method.
• get_wikilink_counts method added.

Other improvements:

• Vault can handle duplicate filenames, better reflecting the 'Shortest path when possible' wikilink setting in Obsidian.
• More robust regex for tags.
• Fix bug where "tags" were being parsed from code block.
• More error handling for front matter.
• Fix bug where source_text gets cut off after a LaTeX block.
• Wikilinks are robust to the use of file extension md in them.

Package now requires Python 3.9 as a minimum.

Wiki has now been added to the Github repo, to cover detail for advanced users.

v0.8.1

Bug fixes:

• Fixed issue where markdown could not parse at all on some environments. md_mermaid extension has been removed. This issue was hard to reproduce so I have removed the extension for now.

v0.8.0

New features:

• The API now has two forms of text: 'source text' and 'readable text'. These have their own object attributes and methods, e.g. get_source_text() and get_readable_text(). The old text attributes and objects have been removed, but they were closest to the source text functionality. The readable text is essentially has a lot of formatting removed, while still retaining the context within notes, so it is in a form that can be used quite easily for NLP analysis. The source text best reflects how notes are formatted in the Obsidian app's 'source mode'.
• Applied multiple pymarkdown extensions to the logic, to reflect some of the main features of how Obsidian uses extended markdown. For example, mermaid diagrams, LaTeX equations and tables can now be parsed, tilde characters mark the deletion of text, etc.
• LaTeX equations can now be accessed for notes via get_math method and math_index attribute.
• More robust tag parsing.

Bug fixes:

• More robust paths for cross-platform support.
• Front matter parsing is more robust.

v0.7.0

New features:

• Support for tags
• Support for instantiating Vault on a filtered list of subdirectories
• 'Gather' functionality for note text: gather function; get_note function and notes_index attr

Fixes:

• Fix embedded files output where the pipe operator is used (e.g. to scale images): avoid backslashes appearing in output
• More robust processing of front matter

v0.6.0

New features:

• Support for front matter
• Support for embedded files
• More columns in metadata df
• get_md_links

v0.5.0

obsidiantools v0.5.0