Flattentool

Fixed

- flattening: Uses much less memory by storing data in a embedded ZODB database, using ijson and using write only mode in pyopenxl.
- use-titles: Use $ref'erring title if available https://github.com/OpenDataServices/flatten-tool/pull/368
- create-template --no-deprecated-fields: Did not work if deprecated element at same level as a $ref https://github.com/OpenDataServices/flatten-tool/issues/185#issuecomment-719587348

Fixed

- Don't fail if compiled translation files (.mo) don't exist, and Django isn't installed
- Fix some incorrect assumptions about types
- Allow XML IDs to be attributes

Fixed

- Don't fail if compiled translation files (.mo) don't exist

Added

- Add Spanish translation for all strings that could appear in the CoVE web UI https://github.com/OpenDataServices/flatten-tool/pull/362

Fixed

- Fix use of titles for sheet names, to avoid KeyError when a field is missing from the schema

Changed

- Include sheet name with duplicate heading warnings.

Changed

- When `--use-titles` is specified, also use the titles for sheet names (ie. Excel tabs and CSV filenames)