refactor: move utils/structures.py into support/

[tmgbedu]

Summary

Audit of fastapi_startkit/src/fastapi_startkit/utils/structures.py (task #1133).

• Finding: not unused — data() is imported by configuration/Configuration.py, and load() is imported by loader/Loader.py.
• Action: moved the file to support/structures.py (support/ already exists from the earlier foundation/support consolidation, PR Relocate providers/helpers into foundation and support #162) via git mv, and updated both import sites (..utils.structures -> ..support.structures). No __init__.py re-export was added, matching the existing precedent for support/dataclass.py (imported by its full module path, not re-exported from support/__init__.py).
• utils/ had no __init__.py and nothing else referenced it, so the directory is gone now that its only file moved.

Test plan

• Grepped the whole monorepo (fastapi_startkit/, example/*, application/, tests) for utils.structures/utils/structures/utils import structures — no remaining references.
• uv run pytest --ignore=tests/masoniteorm/postgres — 1867 passed, 7 skipped, 4 subtests
• uv run ruff check . / uv run ruff format --check . — clean

[tmgbedu]

[Code Reviewer verdict: APPROVE — posting as a comment; gh blocks formal --approve since the reviewing account is also the PR author.]

Verified independently (not just trusting the PR description):

1) History-preserving move + no dangling references

• git diff shows support/structures.py as a 100% similarity RENAME from utils/structures.py — history preserved correctly.
• File content is byte-for-byte unchanged (diff confirms 0 additions/0 deletions on the moved file itself).
• Grepped the entire monorepo for utils.structures / utils/structures / utils import structures / any fastapi_startkit.utils / from ..utils / from .utils pattern — zero hits anywhere.
• utils/ is untracked by git after the move (git ls-files src/fastapi_startkit/utils/ returns nothing) — a fresh clone won't have the directory at all. (A local worktree showed an empty leftover utils/ folder on disk, but that's just a filesystem artifact of an in-place checkout, not something committed — non-issue.)

2) Consistency with support/ conventions (PR #162)

• Checked support/__init__.py: it re-exports Collection, collect, Str, Stringable, Uri, Uriable, Provider — but does not re-export anything from dataclass.py (the Dataclass class is imported by its full module path elsewhere). structures.py follows the same pattern (no re-export added). Precedent claim checks out.

3) Both import sites correctly updated + nothing else broken

• configuration/Configuration.py: from ..support.structures import data — confirmed data() is actually called (self._config = data(), return data(config_at_path)).
• loader/Loader.py: from ..support.structures import load — confirmed load() is actually called in 3 places.
• Relative import depth is unaffected: utils/ and support/ are both direct children of fastapi_startkit/, so structures.py's own internal from ..exceptions.exceptions import LoaderNotFound still resolves correctly post-move.

4) CI / test status

• Reran locally on the PR branch: uv run pytest --ignore=tests/masoniteorm/postgres -q → 1867 passed, 7 skipped, 4 subtests — matches the PR description exactly.
• uv run ruff check . → all checks passed. uv run ruff format --check . → clean (490 files already formatted).
• gh pr checks 186: Ruff and Pyright (non-blocking) both pass (on both workflow runs). The Pytest GitHub Actions job was still pending/queued at review time (not failed) — my local rerun already confirms the suite passes, so this isn't a correctness concern, just CI queue latency worth confirming turns green before merge.

No issues found. Clean, well-scoped refactor. Approving.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!