feat(provider): add Dots chat completion provider

[wcqqq1214]

Refs #10094.

Add an opt-in Dots provider that converts native <dots_function_call> blocks into standard tool calls, keeping protocol markup out of assistant replies.

This PR covers text, tool calls, and streaming. Dots-specific audio/video adaptation and validation are not included.

Modifications / 改动点

• Add Dots configuration and per-request api-key authentication, preserving key rotation, bearer authentication, and proxy settings.

• Support native XML/JSON calls and incremental streaming. Preserve argument whitespace, exclude reasoning examples from execution, and prefer standard calls over duplicate native blocks.

• Add regression tests and English/Chinese setup and migration guides. Existing configurations are unchanged; migration is manual.

• 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

ruff format . && ruff check .
Passed

python -m pytest tests/test_dots_source.py tests/test_openai_source.py tests/test_tool_loop_agent_runner.py tests/test_openai_thinking_tags.py -q
393 passed

cd docs && npm run docs:build
Passed

The docs build used npm with existing dependencies because local pnpm blocked esbuild's install script. No dependencies changed.

WebUI verified: Providers → Chat Completion → Add → Dots, model configuration controls, and Dots selection under Config → AI → Model. Both guides match the current controls; no existing navigation or screenshots require replacement.

The original issue's full group-chat environment remains unverified. Cross-turn search deduplication, search-service rate limits, and model-generated off-topic replies are outside this PR's scope.

---

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 checked the affected WebUI instructions and screenshots in docs/zh and docs/en against the changed navigation, page structure, and labels, and updated them in this PR (or explained why no documentation update is needed). For renamed, moved, or merged entry points, I included an old entry → new entry mapping in the documentation and changelog.
  / 我已对照变化后的 WebUI 入口、页面结构和术语，核对并在本 PR 中更新 docs/zh 和 docs/en 的相关操作说明与截图（或说明无需更新文档的原因）。入口改名、移动或合并时，已在文档和 changelog 中补充 旧入口 → 新入口 对照。

• 🤓 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

Add an opt-in Dots chat completion provider with native tool-call compatibility, safe streaming support, WebUI integration, documentation, and regression tests.

New Features:

• Add an opt-in Dots chat completion provider with default configuration, model discovery, authentication, proxy, and key rotation support.
• Convert Dots-native XML and JSON function-call blocks into validated standard tool calls for both regular and streaming responses.
• Add Dots provider setup, migration, tool-calling, and streaming documentation in English and Chinese.
• Add a Dots provider icon to the WebUI.

Bug Fixes:

• Prevent native call markup and reasoning examples from appearing in assistant replies or triggering unintended tool execution.
• Preserve argument whitespace and prioritize standard tool calls when native and standard calls are duplicated.

Enhancements:

• Buffer ambiguous native markers during tool-enabled streaming while releasing safe assistant text promptly and preserving reasoning, usage, and failure handling.

Documentation:

• Document Dots provider setup, migration from OpenAI Compatible configurations, supported tool-call behavior, and streaming limitations in both supported languages.

Tests:

• Add comprehensive regression coverage for authentication rotation, XML/JSON parsing, schema validation, reasoning exclusion, standard-call precedence, streaming, and agent tool-loop behavior.

[sourcery-ai]

Hey - I've found 2 issues

Prompt for AI Agents

Please address the comments from this code review:

## Individual Comments

### Comment 1
<location path="astrbot/core/provider/sources/dots_source.py" line_range="76-79" />
<code_context>
+
+        Raises:
+            ValueError: The block, tool name, or arguments are invalid.
+        """
+        calls = []
+        if body.lstrip().startswith("{"):
+            try:
+                call = json.loads(body)
+            except json.JSONDecodeError as exc:
+                raise ValueError("Invalid Dots JSON tool call") from exc
+            calls.append(
+                (call.get("name"), call.get("arguments", call.get("parameters", {})))
+            )
+        else:
+            invoke_pattern = re.compile(
</code_context>
<issue_to_address>
**issue (bug_risk):** A syntactically valid JSON value that is not an object, such as `[]` or `null`, causes `call.get(...)` to raise `AttributeError` instead of the documented `ValueError`, so malformed native calls escape the provider's validation path.

**Triggers:** When a Dots JSON function-call body parses to a non-object JSON value.

**Suggested fix:** Check `isinstance(call, dict)` immediately after `json.loads` and raise `ValueError` for other JSON types.

```suggestion
            try:
                call = json.loads(body)
            except json.JSONDecodeError as exc:
                raise ValueError("Invalid Dots JSON tool call") from exc
            if not isinstance(call, dict):
                raise ValueError("Invalid Dots JSON tool call")
```
</issue_to_address>

### Comment 2
<location path="astrbot/core/provider/sources/dots_source.py" line_range="207-210" />
<code_context>
+            normalized = completion.model_dump()
+            normalized["choices"][0]["message"]["content"] = (
+                re.sub(
+                    r"<dots_function_call>.*?</dots_function_call>"
+                    r"|<dots_function_call.*\Z|</dots_function_call>",
+                    "",
+                    content,
+                    flags=re.DOTALL,
+                ).strip()
</code_context>
<issue_to_address>
**issue (bug_risk):** When a standard tool call is present alongside an incomplete `<dots_function_call` marker, the fallback regex removes the marker and every subsequent character, including legitimate assistant text after the marker; standard-call precedence therefore silently truncates the reply.

**Triggers:** When a response contains an authoritative standard tool call, a malformed native-call opening marker, and ordinary text after that marker.

**Suggested fix:** Remove only the native markup span, or preserve text after an incomplete marker rather than applying a `.*\Z` suffix deletion.
</issue_to_address>

Sourcery assessment

Needs a human reviewer. 2 findings to address first, and this adds a new external API integration that transmits provider keys and can turn model-generated Dots markup into executable tool calls, so a parsing or endpoint mistake could expose credentials or trigger external actions before a revert. Reverting stops future requests, but credentials already sent and tool effects that already occurred cannot be undone.

Blocking findings: astrbot/core/provider/sources/dots_source.py:79, astrbot/core/provider/sources/dots_source.py:210

---

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

• X
• Mastodon
• LinkedIn
• Facebook

[sourcery-ai]

issue (bug_risk): A syntactically valid JSON value that is not an object, such as [] or null, causes call.get(...) to raise AttributeError instead of the documented ValueError, so malformed native calls escape the provider's validation path.

Triggers: When a Dots JSON function-call body parses to a non-object JSON value.

Suggested fix: Check isinstance(call, dict) immediately after json.loads and raise ValueError for other JSON types.

Suggested change

	try:
	call = json.loads(body)
	except json.JSONDecodeError as exc:
	raise ValueError("Invalid Dots JSON tool call") from exc
	try:
	call = json.loads(body)
	except json.JSONDecodeError as exc:
	raise ValueError("Invalid Dots JSON tool call") from exc
	if not isinstance(call, dict):
	raise ValueError("Invalid Dots JSON tool call")

[sourcery-ai]

issue (bug_risk): When a standard tool call is present alongside an incomplete <dots_function_call marker, the fallback regex removes the marker and every subsequent character, including legitimate assistant text after the marker; standard-call precedence therefore silently truncates the reply.

Triggers: When a response contains an authoritative standard tool call, a malformed native-call opening marker, and ordinary text after that marker.

Suggested fix: Remove only the native markup span, or preserve text after an incomplete marker rather than applying a .*\Z suffix deletion.

[wcqqq1214]

@codex review