org-rw/README.org

* Org-rw

A python library to parse, modify and save Org-mode files.

* Goals
- Reading org-mode files, with all the relevant information (format, dates, lists, links, metadata, ...).
- Modify these data and write it back to disk.
- Keep the original structure intact (indentation, spaces, format, ...).

** Principles
- Data structures should be exposed as it's read on Emacs's org-mode or when in doubt as raw as possible.
- Data in the objects should be modificable, as a way to update the document itself. *Consider this a Object-oriented design.*
- *Modification of the original text if there's no change is considered a bug (see [[id:7363ba38-1662-4d3c-9e83-0999824975b7][Known issues]]).*
** Safety mechanism
As this library is still in early development. Running it over files might
produce unexpected changes on them. For this reason it's heavily recommended to
have backup copies before using it on important files.

By default the library checks that the re-serialization of the loaded files will
not produce any change, and throw an error in case it does. But this cannot
guarantee that later changes to the document will not corrupt the output so be
careful.

Also, see [[id:76e77f7f-c9e0-4c83-ad2f-39a5a8894a83][Known issues:Structure modifications]] for cases when the structure is
not properly stored and can trigger this safety mechanism on a false-positive.

* Known issues
:PROPERTIES:
:ID:       7363ba38-1662-4d3c-9e83-0999824975b7
:END:
** Structure modifications
:PROPERTIES:
:ID:       76e77f7f-c9e0-4c83-ad2f-39a5a8894a83
:END:
- The exact format is not retained when saving dates/times. This might cause problems with the safety mechanism if you have dates that.
  Note that in both cases, doing ~C-c C-c~ on the date (from Emacs) will change it to the format that Org-rw serializes it to.
  - Use multiple dashes for hour ranges, like =<2020-12-01 10:00----11:00>=. It will get re-serialized as =<2020-12-01 10:00-11:00>=, thus triggering the safety mechanism as unexpected changes have happened.
  - Same in case hours are not two digits (with leading 0's if needed), like =<2020-12-01 9:00>=. It will get serialized as =<2020-12-01 9:00>=.

* Other python libraries for org-mode
- [[https://github.com/karlicoss/orgparse][orgparse]] :: More mature, but does not provide format support or writing back to disk.