Prepare for SYNERGY+

[timovdk]

• Support two sources (backwards compatibility for SYN1), you can set an environment variable to switch to SYN1, this is also used to decide between what code to use when creating the dataset.
• Fix concept code, SYN1 had old openalex work objects, but concepts are deprecated, so we have to use the topics/domain/field system now. (but concept logic was retained for SYN1)
• Fix concepts in metadata.json files (same as above, but then for creating review_metadata.csv and other CLI commands)
• Add extensive tests
• Add Pydantic validation (Created Models based on the OpenAlex API description to validate)
• Add "extended" option to get command
• Add flattening to complex fields (authorships is split into lists of authors etc.)
• Set up train/test split
• Setup URLs to the correct repositories

[EmilyWes]

Looks very good! Nice work!
Added 2 comments, but after that I would say good to go :)

[timovdk]

After a chat with Jonathan I decided to reduce the number of options that the new version of the package will support. This improves maintainability and makes it simpler to 'just' download synergy for a quick benchmark.

[J535D165]

Can you mark this ready for review?