fix: prevent RecursiveCharacterChunker overlap from exceeding chunk_size

[lxfight]

Fixes #9901

In RecursiveCharacterChunker.chunk(), when a split triggered a flush, the next chunk was unconditionally built as overlap_text + split without re-checking the combined length against chunk_size. Every flushed chunk could therefore be up to chunk_size + overlap long (about 562 chars with the default 512/50 config), and a larger configured overlap made the repeated overlap portion dominate the chunks, distorting embedding counts and retrieval quality.

Meanwhile _split_by_character() already rejects overlap >= chunk_size, so the two code paths disagreed on the same invariant.

Modifications / 改动点

• astrbot/core/knowledge_base/chunking/recursive.py:

  • After prepending the overlap to the next chunk, re-check the combined length; when overlap + split would exceed chunk_size, emit the overlap as a standalone chunk (bounded by overlap < chunk_size) and start the new chunk with the split only.
  • Add the same parameter validation as _split_by_character at chunk() entry (chunk_size > 0, 0 <= overlap < chunk_size) so both paths enforce the same invariant.

• tests/test_recursive_chunker.py: regression test proving no output chunk exceeds chunk_size (fails on the old code with 110-char chunks), plus small-text, normal-text, and invalid-overlap cases.

• This is NOT a breaking change. / 这不是一个破坏性变更。

Screenshots or Test Results / 运行截图或测试结果

$ uv run pytest tests/test_recursive_chunker.py -q
5 passed in 0.40s

# on the unfixed code the regression test fails as expected:
# FAILED test_overlap_never_pushes_chunks_over_chunk_size (110 > 100)

• uv run ruff format / uv run ruff check pass.
• Verification steps: upload a long document with the default chunker config and inspect the stored chunks — max chunk length is now within the configured chunk_size instead of chunk_size + overlap.

---

Checklist / 检查清单

• 😊 If there are new features added in the PR, I have discussed it with the authors through issues/emails, etc.
  / 如果 PR 中有新加入的功能，已经通过 Issue / 邮件等方式和作者讨论过。

• 👀 My changes have been well-tested, and "Verification Steps" and "Screenshots" have been provided above.
  / 我的更改经过了良好的测试，并已在上方提供了“验证步骤”和“运行截图”。

• 🤓 I have ensured that no new dependencies are introduced, OR if new dependencies are introduced, they have been added to the appropriate locations in requirements.txt and pyproject.toml.
  / 我确保没有引入新依赖库，或者引入了新依赖库的同时将其添加到 requirements.txt 和 pyproject.toml 文件相应位置。

• 😮 My changes do not introduce malicious code.
  / 我的更改没有引入恶意代码。

Summary by Sourcery

Ensure recursive character chunking respects configured size limits while validating overlap parameters consistently.

Bug Fixes:

• Prevent recursive chunking overlap from producing chunks larger than the configured chunk size.
• Reject invalid chunk size and overlap configurations consistently across recursive chunking paths.

Tests:

• Add regression and validation coverage for chunk size limits, small and normal inputs, and invalid overlap settings.