BinaryRelationEncoderDecoder.parse: never return remainings that are longer than 6

[ArneBinder]

Despite having #199 now, it still happens that token sequences get produced that are not allowed which breaks follow-up code of parsing which expects that everything that did not got parsed (a.k.a. the remaining sequence) is shorter than a complete tuple, i.e., 7 entries. Thus, we enforce the remaining to be shorter than 7 tokens directly in the parse method by trying to decode all chunks that grow up to 7 entries (instead of requiring that they also end with a none_label).

EDIT: Thinking about this again, it may be more correct to adjust the BinaryRelationEncoderDecoder.build_decoding_constraints to work with partial_encoding of length >= 7. This needs further investigation why this happens at all.

TODO:

• decide if the approach mentioned as EDIT would be better
• test in downstream projects

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.75%. Comparing base (00c88ca) to head (61a1ff3).
⚠️ Report is 20 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #200   +/-   ##
=======================================
  Coverage   95.75%   95.75%           
=======================================
  Files          68       68           
  Lines        6168     6168           
=======================================
  Hits         5906     5906           
  Misses        262      262           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.