fix: preserve embedding api version suffixes

[he-yufeng]

Summary

• preserve any /vN suffix when normalizing OpenAI-compatible embedding API bases
• keep the existing fallback that appends /v1 when a base URL has no version suffix
• add regression coverage for /v3, /v4, unversioned bases, and /embeddings suffix trimming

Why

Fixes #8732. Providers such as Volcengine Ark can use embedding endpoints ending in /v3. The old check only recognized /v1 and /v4, so https://.../v3 became https://.../v3/v1 and returned 404.

Verified

• python -m py_compile astrbot\core\provider\sources\openai_embedding_source.py tests\test_openai_embedding_source.py
• .\.venv\Scripts\python.exe -m ruff check astrbot\core\provider\sources\openai_embedding_source.py tests\test_openai_embedding_source.py
• git diff --check
• direct helper behavior check covering /v3, /v4, unversioned base URLs, and /v1/embeddings

I also tried .\.venv\Scripts\python.exe -m pytest tests\test_openai_embedding_source.py -q in a minimal local venv. Collection reached the shared tests/conftest.py dependency chain and stopped on missing full AstrBot runtime dependencies; I did not install the whole application dependency set just for this small provider normalization check.

Summary by Sourcery

Preserve and centralize normalization of OpenAI-compatible embedding API base URLs while ensuring a sensible default version is applied.

Bug Fixes:

• Fix incorrect normalization that appended '/v1' to embedding API bases already using other version suffixes like '/v3', which could cause 404 responses.

Enhancements:

• Extract API base URL normalization into a reusable helper that trims trailing slashes and '/embeddings' while respecting existing version suffixes.

Tests:

• Add regression tests covering versioned embedding API bases (including '/v3' and '/v4'), unversioned bases, and '/embeddings' suffix trimming behavior.

[gemini-code-assist]

Code Review

This pull request refactors the API base normalization logic for the OpenAI embedding source into a dedicated helper function _normalize_api_base and adds corresponding unit tests. The reviewer identified a potential AttributeError if embedding_api_base is explicitly configured as None or an empty string, and provided a code suggestion to safely fall back to the default URL using the or operator.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

If embedding_api_base is explicitly configured as None or an empty string in the provider configuration, provider_config.get("embedding_api_base", "https://api.openai.com/v1") will return None or "". This will cause _normalize_api_base to raise an AttributeError (since None has no strip method) or fail to apply the default base URL.

Using or instead of the get method's default argument ensures that we fall back to the default OpenAI API base URL if the configured value is falsy (such as None or "").

Suggested change

	api_base = _normalize_api_base(
	provider_config.get("embedding_api_base", "https://api.openai.com/v1")
	.strip()
	.removesuffix("/")
	.removesuffix("/embeddings")
	)
	api_base = _normalize_api_base(
	provider_config.get("embedding_api_base") or "https://api.openai.com/v1"
	)

[sourcery-ai]

Hey - I've found 1 issue

Prompt for AI Agents

Please address the comments from this code review:

## Individual Comments

### Comment 1
<location path="tests/test_openai_embedding_source.py" line_range="12-16" />
<code_context>
+    assert _normalize_api_base("https://example.test/v4") == "https://example.test/v4"
+
+
+def test_openai_embedding_api_base_adds_default_version():
+    assert _normalize_api_base("https://example.test/openai") == (
+        "https://example.test/openai/v1"
+    )
+    assert _normalize_api_base("https://example.test/v1/embeddings") == (
+        "https://example.test/v1"
+    )
</code_context>
<issue_to_address>
**suggestion:** Cover behavior for empty or whitespace-only bases and non-version `/embeddings` bases

The default-version behavior is only partially covered here. Please also add cases for:

- `"https://example.test/embeddings"` → `"https://example.test/v1"`, as described in the PR but not asserted.
- Empty / whitespace-only input (e.g., `""`, `"  "`), which currently returns an empty string.

These will lock in the `/embeddings` trimming and fallback behavior and help prevent regressions.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

suggestion: Cover behavior for empty or whitespace-only bases and non-version /embeddings bases

The default-version behavior is only partially covered here. Please also add cases for:

• "https://example.test/embeddings" → "https://example.test/v1", as described in the PR but not asserted.
• Empty / whitespace-only input (e.g., "", " "), which currently returns an empty string.

These will lock in the /embeddings trimming and fallback behavior and help prevent regressions.