geotiff: clean up flake8/isort style drift in tests and two source imports

[brendancol]

Reason or Problem

The geotiff subpackage has no flake8 or isort check in CI, so style drift piles up unnoticed. Running the project's own configured tooling (setup.cfg: flake8 max-line-length=100, isort line_length=100) turns up a small batch of real issues. The 33 source files are clean. Everything flagged lives in the test suite, apart from two import blocks in source files.

Proposal

Apply the project's configured flake8 + isort tooling to the geotiff subpackage and fix what it reports. Style-only cleanup, no behavioural change.

Findings (tool-confirmed):

pyflakes F-codes (dead code that can mask refactor leftovers):

• tests/unit/test_geotags.py:4: F401 struct imported but unused at module level
• tests/unit/test_geotags.py:256, :408: F811 redefinition of unused struct; a local import struct inside two functions shadows the dead top-level import
• tests/unit/test_metadata.py:1446: F841 local ascii_buf assigned but never used

Line length:

• tests/unit/test_metadata.py:31: E501 (101 > 100), part of a multi-line import block that isort reflows

Import ordering (isort):

• Source files: __init__.py and _writers/eager.py
• ~40 test files under tests/

Trailing blank line:

• tests/unit/test_geotags.py:833: W391

Value

Removes dead test imports and a dead local, and brings the subpackage in line with the project's own isort/flake8 config so future drift is easier to spot.

Drawbacks

The isort pass reorders imports across roughly 40 test files, which is churn. The diff is mechanical and reviewable file by file.

Additional Notes or Context

Scope guardrails: no black/autopep8/ruff-format, no config widening. Fixes limited to what flake8 + isort report. Verify with flake8 xrspatial/geotiff/ (expect 0) and isort --check-only xrspatial/geotiff/ (expect clean).

Found by /sweep-style.