State bookmarks should use ids instead of repo names

Nobody has claimed this yet.

Following the same logic as what is described in #110 and the linked PR, the state bookmarks should use stable IDs for their context, as opposed to the current `org / repo`:
If a repo is renamed between 2 runs of the tap, the later run would ignore the bookmark, as the repo name would not match anymore. While not a disaster, it would essentially force a full resync of the history of the stream, which is a waste of time/energy/api quotas.

Example of a state bookmark right now:
```json
{
"context": { "org": "scipy", "repo": "scipy" }, // this should use a stable id
"replication_key": "updated_at",
"replication_key_value": "2022-05-06T14:29:18Z",
"replication_key_signpost": "2022-06-22T09:07:14.047100+00:00",
"starting_replication_value": "2022-05-06T14:29:18Z",
"progress_markers": {
"Note": "Progress is not resumable if interrupted.",
"replication_key": "updated_at",
"replication_key_value": "2022-06-22T07:55:32Z"
}
},
```

Start by reviewing issue #110 and its linked pull request, then locate the state bookmark serialization and matching logic in the tap. Confirm how repository rename handling currently behaves; done means bookmarks use a stable repository identifier and a renamed repository can resume without a full history resync.

Written by the indexing model from the issue text.