Fix AttributeError when importing eyepieces from DeepskyLog

[brickbots]

Summary

Importing equipment from DeepskyLog on the web interface crashes with an "Internal Server Error" as soon as the import contains a new eyepiece. This is a v2.5.1 → v2.6.0 regression introduced by the Flask migration (#331), which changed the eyepiece-add call from

cfg.equipment.add_eyepiece(new_eyepiece)

to

cfg.equipment.eyepieces.add_eyepiece(new_eyepiece)

Equipment.eyepieces is a plain List[Eyepiece] with no add_eyepiece method, so the first new eyepiece raises AttributeError. Since cfg.save_equipment() comes after the import loop, nothing gets saved — telescopes already collected in the same request are lost too.

This PR restores the pre-migration call on Equipment itself (equipment.py:87), which appends the eyepiece and keeps the list sorted by focal length. The manual add-eyepiece route was unaffected (it already calls the correct method at server.py:736).

Regression tests

tests/test_server_dsl_import.py drives the real /equipment/import_from_deepskylog route through Flask's test client, with pydeepskylog and config.Config mocked out and a real Equipment instance so the plain-list bug reproduces:

• test_import_adds_new_eyepieces — new eyepieces are added (sorted by focal length, HTML entities decoded) and the config is saved. With the bad call reintroduced, this fails with AttributeError: 'list' object has no attribute 'add_eyepiece' — verified.
• test_import_skips_existing_eyepiece — an eyepiece already in the config is not duplicated.

Test plan

• New regression tests pass (and fail on the un-fixed code)
• pytest -m "unit or smoke" passes (785 passed)
• ruff check / ruff format --check pass

🤖 Generated with Claude Code

[mrosseel]

Seems like this cold use a unit test/integration test. Are these run regularly btw?

[brickbots]

Yes indeed! I'll make sure to add a test. Web interface tests don't run as part of the CI/CD, so it's something we'll need to do more of manually as part of a release 👍