fix: support filename prefix search

[lylelllll]

Motivation

mineru find bench currently misses filenames such as bench1.pdf, even though the explicit wildcard form bench* works. This makes the default filename search behavior surprising and is the issue reported in #5281.

Modification

• Apply FTS5 prefix matching to the final sanitized token in filename searches when the query does not already end in *.
• Preserve explicit wildcard queries and keep content full-text search behavior unchanged.
• Add SQLite-backed regression coverage for implicit and explicit prefixes, exact tokens, multi-token queries, CJK filenames, and filename/content search isolation.

BC-breaking (Optional)

No. Filename lookup becomes more permissive for the final query token; public Python APIs, content search, indexing, and storage are unchanged.

Testing

• .venv/bin/python -m pytest -q -o addopts='' tests/unittest/test_doclib_cache_semantics.py tests/unittest/test_cli_next_command_design.py (307 passed)
• uvx ruff check mineru/doclib/core/fts.py
• uvx ruff check --ignore ANN001,E501 tests/unittest/test_doclib_cache_semantics.py
• Direct tokenization smoke test for implicit prefix, explicit wildcard, multi-token, and CJK queries

Model/GPU parsing E2E tests were not run because this change is isolated to SQLite FTS filename query construction.

Checklist

Before PR:

• Pre-commit or other linting tools are used to fix the potential lint issues.
• Bug fixes are fully covered by unit tests, the case that causes the bug should be added in the unit tests.
• The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
• Documentation impact was reviewed; no documentation change is needed for this targeted bug fix.

After PR:

• If the modification has potential influence on downstream or other related projects, this PR should be tested with those projects. (No downstream API or storage change.)
• CLA has been signed and all committers have signed the CLA in this PR.

Fixes #5281