coding-proxy 0.2.4a5__tar.gz → 0.3.0a1__tar.gz

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/CHANGELOG.md

@@ -4,6 +4,11 @@
 
 ## [Unreleased]
 
+- feat(native-api): 新增 `/api/{openai,gemini,anthropic}/**` 原生 LLM API 全量 catch-all 透传通道——客户端只需改 SDK `base_url` 即可复用 proxy 链路访问 OpenAI / Gemini / Anthropic 官方 API，认证完全透传不保管凭据；核心由 `NativeProxyHandler` + `OperationClassifier` + `NativeUsageExtractor` Registry 三件套组成，与既有 `/v1/messages` Claude Code 链路正交共存、零回归；首版覆盖 chat / completions / responses / embeddings / audio / image / count_tokens / moderations / cachedContents / messages / batches 等全量端点；
+- feat(usage): `usage_log` 新增 `client_category` / `operation` / `endpoint` / `extra_usage_json` 四列（全部 `DEFAULT`、幂等迁移），区分 Claude Code 场景（`'cc'`）与原生 API 场景（`'api'`），承载规范化操作名与非规范 token 字段（reasoning / audio / thoughts / server_tool_use 等）；`query_usage` 支持按新列过滤，`_PERIOD_SQL` 聚合追加 `client_category, operation` 维度；
+- feat(usage-parser): `parse_usage_from_chunk` 扩展识别 Gemini SSE `usageMetadata.*`（promptTokenCount / candidatesTokenCount / cachedContentTokenCount / thoughtsTokenCount / toolUsePromptTokenCount），新增 `gemini_usage_metadata` evidence kind，既有 Anthropic/OpenAI 分支行为零变更；
+- refactor(vendor-channels): 彻底收敛跨供应商兼容性逻辑——删除 `server/request_normalizer.py` 入口通用规范化层，将 `srvtoolu_*` ID 重写、`server_tool_use_delta` 私有块剥离全部迁入源→目标绑定通道（`prepare_zhipu_to_anthropic`、`prepare_zhipu_to_copilot`）；新增 `infer_source_vendor_from_body` 内容感知源推断，在无会话状态的首次请求场景下兜底识别源供应商；`_RouteExecutor._determine_source_vendor` 扩充为三级优先级（failed_tier → session_state → body inference），确保未注册转换对不触发任何清洗；
+- refactor(count-tokens): `/v1/messages/count_tokens` 端点移除无条件 `strip_thinking_blocks` 过度防御，改为基于 `infer_source_vendor_from_body` + `get_transition_channel` 的按需通道清洗，语义与 `/v1/messages` 对齐；
 - fix(request-normalizer): 重设计 zhipu→anthropic 跨供应商 tool_use/tool_result 配对修复——以单遍自包含 `enforce_anthropic_tool_pairing` 替代原有多步串联管线（剥离→重定位→孤儿修复），消除步骤间隐式依赖导致的孤儿 tool_use 漏修问题，彻底根治 `tool_use ids were found without tool_result blocks` 400 异常;
 - refactor(vendor-channels): 将供应商转换通道从「目标 vendor 专属」重构为「源→目标绑定」模型——注册表键从 `target_vendor` 改为 `(source, target)` 二元组，通道函数从 `prepare_for_X` 重命名为 `prepare_X_to_Y`，触发逻辑从 `_needs_vendor_channel` 替换为 `_determine_source_vendor`（基于请求内 `failed_tier_name` 和会话历史推断源 vendor），未注册的转换对（如 anthropic→zhipu）不触发任何通道;
 - feat(vendor-channels): 新增 zhipu→anthropic、zhipu→copilot、copilot→zhipu 三条源→目标绑定转换通道，在跨供应商故障转移时自动清理源 vendor 产物（thinking 块、cache_control 字段、thinking 参数、tool_use/tool_result 配对），消除 `likely format incompatibility (400 + tool_results)` 错误;

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: coding-proxy
-Version: 0.2.4a5
+Version: 0.3.0a1
 Summary: A High-Availability, Transparent, and Smart Multi-Vendor Proxy for Claude Code. Support Claude Plans, GitHub Copilot, Google Antigravity, ZAI/GLM, MiniMax, Qwen, Xiaomi, Kimi, Doubao...
 Project-URL: Source Code, https://github.com/ThreeFish-AI/coding-proxy
 Project-URL: User Guide, https://github.com/ThreeFish-AI/coding-proxy/blob/master/docs/user-guide.md
@@ -57,7 +57,7 @@ When you're deeply immersed in your coding "zone" with **Claude Code** (or any A
 ## 🌟 Core Features
 
 <div align="center">
-    <img src="assets/dashboard-v0.2.3.png">
+    <img src="assets/dashboard-v0.2.4.png">
 </div>
 
 - **⛓️ N-tier Chained Failover**: Autonomous descending sequence, supporting Claude's official plans, as well as Coding Plans from GitHub Copilot, Google Antigravity, Z AI, MiniMax, Alibaba Qwen, Xiaomi, Kimi, Doubao, etc.
@@ -119,13 +119,13 @@ claude
 
 `coding-proxy` comes equipped with a badass suite of CLI tools to help you boss around your proxy state.
 
-| Command  | Description                                                                                                                                        | Example Usage                                 |
-| :------- | :------------------------------------------------------------------------------------------------------------------------------------------------- | :-------------------------------------------- |
-| `start`  | **Fire up the proxy server.** Supports custom ports and configuration paths.                                                                       | `coding-proxy start -p 8080 -c ~/config.yaml` |
+| Command  | Description                                                                                                                                          | Example Usage                                 |
+| :------- | :--------------------------------------------------------------------------------------------------------------------------------------------------- | :-------------------------------------------- |
+| `start`  | **Fire up the proxy server.** Supports custom ports and configuration paths.                                                                         | `coding-proxy start -p 8080 -c ~/config.yaml` |
 | `auth`   | **Manage OAuth credentials.** Sub-commands: `login` (browser OAuth), `status` (token validity), `reauth` (re-authenticate), `logout` (clear tokens). | `coding-proxy auth login -p github`           |
-| `status` | **Check proxy health.** Shows circuit breaker states (OPEN/CLOSED) and quota status across all tiers.                                              | `coding-proxy status`                         |
-| `usage`  | **Token Stats Dashboard.** Stalks every single token consumed, failovers triggered, and latency across day/vendor/model dimensions.                | `coding-proxy usage -d 7 -v anthropic`        |
-| `reset`  | **The emergency flush button.** Force-reset all circuit breakers and quotas instantly when you've confirmed the main vendor is back from the dead. | `coding-proxy reset`                          |
+| `status` | **Check proxy health.** Shows circuit breaker states (OPEN/CLOSED) and quota status across all tiers.                                                | `coding-proxy status`                         |
+| `usage`  | **Token Stats Dashboard.** Stalks every single token consumed, failovers triggered, and latency across day/vendor/model dimensions.                  | `coding-proxy usage -d 7 -v anthropic`        |
+| `reset`  | **The emergency flush button.** Force-reset all circuit breakers and quotas instantly when you've confirmed the main vendor is back from the dead.   | `coding-proxy reset`                          |
 
 ---
 

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/README.md

@@ -30,7 +30,7 @@ When you're deeply immersed in your coding "zone" with **Claude Code** (or any A
 ## 🌟 Core Features
 
 <div align="center">
-    <img src="assets/dashboard-v0.2.3.png">
+    <img src="assets/dashboard-v0.2.4.png">
 </div>
 
 - **⛓️ N-tier Chained Failover**: Autonomous descending sequence, supporting Claude's official plans, as well as Coding Plans from GitHub Copilot, Google Antigravity, Z AI, MiniMax, Alibaba Qwen, Xiaomi, Kimi, Doubao, etc.
@@ -92,13 +92,13 @@ claude
 
 `coding-proxy` comes equipped with a badass suite of CLI tools to help you boss around your proxy state.
 
-| Command  | Description                                                                                                                                        | Example Usage                                 |
-| :------- | :------------------------------------------------------------------------------------------------------------------------------------------------- | :-------------------------------------------- |
-| `start`  | **Fire up the proxy server.** Supports custom ports and configuration paths.                                                                       | `coding-proxy start -p 8080 -c ~/config.yaml` |
+| Command  | Description                                                                                                                                          | Example Usage                                 |
+| :------- | :--------------------------------------------------------------------------------------------------------------------------------------------------- | :-------------------------------------------- |
+| `start`  | **Fire up the proxy server.** Supports custom ports and configuration paths.                                                                         | `coding-proxy start -p 8080 -c ~/config.yaml` |
 | `auth`   | **Manage OAuth credentials.** Sub-commands: `login` (browser OAuth), `status` (token validity), `reauth` (re-authenticate), `logout` (clear tokens). | `coding-proxy auth login -p github`           |
-| `status` | **Check proxy health.** Shows circuit breaker states (OPEN/CLOSED) and quota status across all tiers.                                              | `coding-proxy status`                         |
-| `usage`  | **Token Stats Dashboard.** Stalks every single token consumed, failovers triggered, and latency across day/vendor/model dimensions.                | `coding-proxy usage -d 7 -v anthropic`        |
-| `reset`  | **The emergency flush button.** Force-reset all circuit breakers and quotas instantly when you've confirmed the main vendor is back from the dead. | `coding-proxy reset`                          |
+| `status` | **Check proxy health.** Shows circuit breaker states (OPEN/CLOSED) and quota status across all tiers.                                                | `coding-proxy status`                         |
+| `usage`  | **Token Stats Dashboard.** Stalks every single token consumed, failovers triggered, and latency across day/vendor/model dimensions.                  | `coding-proxy usage -d 7 -v anthropic`        |
+| `reset`  | **The emergency flush button.** Force-reset all circuit breakers and quotas instantly when you've confirmed the main vendor is back from the dead.   | `coding-proxy reset`                          |
 
 ---
 

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/docs/arch/testing.md

@@ -65,6 +65,7 @@
 | `test_convert_sse.py`              | Gemini SSE→Anthropic SSE 流适配（单/多 chunk、各 finishReason、边界情况） |
 | `test_copilot_convert_request.py`  | Anthropic→OpenAI 请求格式转换                                             |
 | `test_copilot_convert_response.py` | OpenAI→Anthropic 响应格式转换                                             |
+| `test_vendor_channels.py`          | 源→目标通道：zhipu/copilot 兼容性清洗、tool_use 配对、内容感知源推断      |
 
 ### 2.5 数据模型（model）
 
@@ -95,7 +96,6 @@
 | 测试文件                     | 覆盖范围                                                |
 | ---------------------------- | ------------------------------------------------------- |
 | `test_app_routes.py`         | FastAPI 路由端点测试                                    |
-| `test_request_normalizer.py` | 请求标准化：私有块清洗、tool_use_id 重写、fatal_reasons |
 | `test_cli_usage.py`          | CLI 用量查询命令                                        |
 | `test_banner.py`             | CLI Banner 显示                                         |
 | `test_logging_dual_write.py` | 日志双写机制                                            |

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/docs/framework.md

@@ -69,7 +69,6 @@ graph TD
         App["<code>app.py</code><br/>应用工厂 + lifespan"]
         Routes["<code>routes.py</code><br/>路由注册"]
         Factory["<code>factory.py</code><br/>Vendor/Tier 构建工厂"]
-        Normalizer["<code>request_normalizer.py</code><br/>请求标准化"]
         Dashboard["<code>dashboard.py</code><br/>状态面板"]
     end
 
@@ -179,15 +178,8 @@ graph TD
 ```mermaid
 flowchart TD
     Client["Client POST /v1/messages"] --> Server["server.routes.messages()"]
-
-    subgraph Normalize ["请求标准化"]
-        Body["body = await request.json()"]
-        Norm["normalize_anthropic_request(body)<br/>清洗私有块 + 重写 tool_use_id"]
-        Body --> Norm
-    end
-
-    Server --> Normalize
-    Norm --> RouteType{"stream?"}
+    Server --> Body["body = await request.json()"]
+    Body --> RouteType{"stream?"}
 
     RouteType -- "true" --> StreamRoute["route_stream()"]
     RouteType -- "false" --> MsgRoute["route_message()"]
@@ -401,7 +393,6 @@ flowchart TD
 | [`app.py`](../src/coding/proxy/server/app.py)                               | FastAPI 应用工厂 `create_app()` + `lifespan` 生命周期管理 |
 | [`factory.py`](../src/coding/proxy/server/factory.py)                       | Vendor/Tier 构建工厂 + 凭证解析                           |
 | [`routes.py`](../src/coding/proxy/server/routes.py)                         | 路由端点按职责分组注册                                    |
-| [`request_normalizer.py`](../src/coding/proxy/server/request_normalizer.py) | 入站请求标准化（清洗供应商私有块）                        |
 | [`responses.py`](../src/coding/proxy/server/responses.py)                   | 响应辅助工具（JSON error / stream error 构建）            |
 | [`dashboard.py`](../src/coding/proxy/server/dashboard.py)                   | 状态面板（Web Dashboard）                                 |
 

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/docs/zh-CN/README.md

@@ -30,7 +30,7 @@
 ## 🌟 核心特性 (Core Features)
 
 <div align="center">
-    <img src="../../assets/dashboard-v0.2.3.png">
+    <img src="../../assets/dashboard-v0.2.4.png">
 </div>
 
 - **⛓️ N-tier 链式故障转移 (Failover)**：自主降序序列，支持 Claude 官方 Plans，以及 GitHub Copilot、Google Antigravity、智谱、MiniMax、阿里千问、小米、Kimi、豆包等的 Coding Plan。
@@ -92,13 +92,13 @@ claude
 
 `coding-proxy` 附带了强大的 CLI 工具套件，帮助您全面掌控代理状态。
 
-| 指令     | 说明                                                                              | 示例用法                                      |
-| :------- | :-------------------------------------------------------------------------------- | :-------------------------------------------- |
-| `start`  | **启动代理服务器**。支持自定义端口与配置路径。                                    | `coding-proxy start -p 8080 -c ~/config.yaml` |
+| 指令     | 说明                                                                                                                            | 示例用法                                      |
+| :------- | :------------------------------------------------------------------------------------------------------------------------------ | :-------------------------------------------- |
+| `start`  | **启动代理服务器**。支持自定义端口与配置路径。                                                                                  | `coding-proxy start -p 8080 -c ~/config.yaml` |
 | `auth`   | **管理 OAuth 登录凭证**。子命令：`login`（浏览器 OAuth 登录）、`status`（令牌状态）、`reauth`（重认证）、`logout`（清除令牌）。 | `coding-proxy auth login -p github`           |
-| `status` | **查看代理健康状态**。展示各层级熔断器（OPEN/CLOSED）与配额状态。                 | `coding-proxy status`                         |
-| `usage`  | **Token 统计看板**。按天/供应商/模型维度追踪每一次的 Token 消耗、故障转移及耗时。 | `coding-proxy usage -d 7 -v anthropic`        |
-| `reset`  | **强制一键重置**。人工确认主供应商恢复可用后，立刻初始化所有熔断器和配额状态。    | `coding-proxy reset`                          |
+| `status` | **查看代理健康状态**。展示各层级熔断器（OPEN/CLOSED）与配额状态。                                                               | `coding-proxy status`                         |
+| `usage`  | **Token 统计看板**。按天/供应商/模型维度追踪每一次的 Token 消耗、故障转移及耗时。                                               | `coding-proxy usage -d 7 -v anthropic`        |
+| `reset`  | **强制一键重置**。人工确认主供应商恢复可用后，立刻初始化所有熔断器和配额状态。                                                  | `coding-proxy reset`                          |
 
 ---
 

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "coding-proxy"
-version = "0.2.4a5"
+version = "0.3.0a1"
 description = "A High-Availability, Transparent, and Smart Multi-Vendor Proxy for Claude Code. Support Claude Plans, GitHub Copilot, Google Antigravity, ZAI/GLM, MiniMax, Qwen, Xiaomi, Kimi, Doubao..."
 readme = "README.md"
 requires-python = ">=3.12"

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/src/coding/proxy/config/config.default.yaml

@@ -331,6 +331,12 @@ model_mapping:
 # 未配置定价的模型在统计中 Cost 列显示 "-"
 pricing:
   # ── Anthropic Claude ──
+  - vendor: anthropic
+    model: claude-opus-4-7
+    input_cost_per_mtok: $5.0
+    output_cost_per_mtok: $25.0
+    cache_write_cost_per_mtok: $6.25
+    cache_read_cost_per_mtok: $0.50
   - vendor: anthropic
     model: claude-opus-4-6
     input_cost_per_mtok: $5.0

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/src/coding/proxy/config/schema.py

@@ -19,6 +19,7 @@ from typing import Any
 
 from pydantic import BaseModel, Field, model_validator
 
+from ..native_api.config import NativeApiConfig  # noqa: F401
 from .auth_schema import AuthConfig  # noqa: F401
 from .resiliency import (  # noqa: F401
     CircuitBreakerConfig,
@@ -143,6 +144,14 @@ class ProxyConfig(BaseModel):
             "未配置时回退到 vendors 列表原始顺序。"
         ),
     )
+    # 原生 LLM API 透传通道 — 与 /v1/messages（Claude Code）链路正交
+    native_api: NativeApiConfig = Field(
+        default_factory=NativeApiConfig,
+        description=(
+            "OpenAI / Gemini / Anthropic 原生 API 透传配置。"
+            "三个 provider 默认 enabled=False，显式启用才暴露 /api/{provider}/* 端点。"
+        ),
+    )
 
     @model_validator(mode="before")
     @classmethod
@@ -320,4 +329,6 @@ __all__ = [
     "DoubaoConfig",
     "XiaomiConfig",
     "AlibabaConfig",
+    # native api passthrough
+    "NativeApiConfig",
 ]

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/src/coding/proxy/convert/vendor_channels.py

@@ -16,6 +16,7 @@ from __future__ import annotations
 
 import copy
 import logging
+import re
 from collections.abc import Callable
 from typing import Any
 
@@ -23,6 +24,17 @@ logger = logging.getLogger(__name__)
 
 _THINKING_BLOCK_TYPES = {"thinking", "redacted_thinking"}
 
+# ── Anthropic 工具块 ID 规范 ───────────────────────────────────
+_ANTHROPIC_TOOL_USE_ID_RE = re.compile(r"^toolu_[A-Za-z0-9_]+$")
+_ANTHROPIC_SERVER_TOOL_USE_ID_RE = re.compile(r"^srvtoolu_[A-Za-z0-9_]+$")
+
+# Zhipu 流式响应中出现的非标准供应商私有 content block 类型.
+# Anthropic API 拒绝这些块，需要在跨 vendor 请求体中剥离.
+_ZHIPU_VENDOR_BLOCK_TYPES = {"server_tool_use_delta"}
+
+# Zhipu 内联输出非标准 content block 类型的标识（用于源供应商推断）.
+_ZHIPU_SERVER_TOOL_USE_TYPES = {"server_tool_use", "server_tool_use_delta"}
+
 # ── 转换通道注册表 ─────────────────────────────────────────────
 # (source_vendor, target_vendor) → (body) → (prepared_body, adaptations)
 VENDOR_TRANSITIONS: dict[
@@ -250,6 +262,138 @@ def _strip_cache_control(body: dict[str, Any]) -> int:
     return removed
 
 
+def _remove_vendor_blocks(body: dict[str, Any], block_types: set[str]) -> int:
+    """从 messages[].content[] 中就地移除指定 type 的内容块.
+
+    用于剥离 vendor 私有 content block 类型（如 zhipu 的 ``server_tool_use_delta``），
+    Anthropic API 会拒绝这些非标准块。
+
+    Returns:
+        被移除的块数量。
+    """
+    removed = 0
+    for message in body.get("messages", []):
+        if not isinstance(message, dict):
+            continue
+        content = message.get("content")
+        if not isinstance(content, list):
+            continue
+        new_content: list[Any] = []
+        for block in content:
+            if isinstance(block, dict) and block.get("type") in block_types:
+                removed += 1
+                continue
+            new_content.append(block)
+        if removed:
+            message["content"] = new_content
+    return removed
+
+
+def _rewrite_srvtoolu_ids(body: dict[str, Any]) -> tuple[int, dict[str, str]]:
+    """将 zhipu 的 ``server_tool_use`` + ``srvtoolu_*`` ID 改写为标准 Anthropic 形式.
+
+    Anthropic API 要求 tool_use 类型与 ``toolu_*`` 格式的 ID。Zhipu 的
+    ``server_tool_use`` + ``srvtoolu_*`` 在上游 Anthropic 兼容端点可用，但无法
+    透传至其他供应商；同时还需重写紧随其后 user 消息中 ``tool_result.tool_use_id``
+    引用，保持配对关系。
+
+    Returns:
+        (rewritten_count, id_map) — 重写次数与 {原 ID: 新 ID} 映射。
+    """
+    id_map: dict[str, str] = {}
+    counter = 0
+
+    def next_id() -> str:
+        nonlocal counter
+        counter += 1
+        return f"toolu_normalized_{counter}"
+
+    for message in body.get("messages", []):
+        if not isinstance(message, dict):
+            continue
+        content = message.get("content")
+        if not isinstance(content, list):
+            continue
+        role = message.get("role")
+        for block in content:
+            if not isinstance(block, dict):
+                continue
+            block_type = block.get("type")
+            block_id = block.get("id")
+
+            # Case A: assistant 消息里的 server_tool_use / srvtoolu_* → 改写
+            if role == "assistant" and block_type in {"tool_use", "server_tool_use"}:
+                if isinstance(block_id, str) and _ANTHROPIC_SERVER_TOOL_USE_ID_RE.match(
+                    block_id
+                ):
+                    new_id = next_id()
+                    id_map[block_id] = new_id
+                    block["id"] = new_id
+                    block["type"] = "tool_use"
+                elif (
+                    isinstance(block_id, str)
+                    and block_id
+                    and not _ANTHROPIC_TOOL_USE_ID_RE.match(block_id)
+                    and block.get("name")
+                ):
+                    # 非标准 ID（非 toolu_ / srvtoolu_），且具备 name 可改写
+                    new_id = next_id()
+                    id_map[block_id] = new_id
+                    block["id"] = new_id
+                    block["type"] = "tool_use"
+                elif block_type == "server_tool_use" and isinstance(block_id, str):
+                    # 兜底: 类型是 server_tool_use 但 ID 已是标准 toolu_ 形式，仅纠正类型
+                    block["type"] = "tool_use"
+
+            # Case B: user 消息里的 tool_result.tool_use_id 同步重写
+            if block_type == "tool_result":
+                tool_use_id = block.get("tool_use_id")
+                if isinstance(tool_use_id, str) and tool_use_id in id_map:
+                    block["tool_use_id"] = id_map[tool_use_id]
+
+    return len(id_map), id_map
+
+
+def infer_source_vendor_from_body(body: dict[str, Any]) -> str | None:
+    """从请求 body 内容推断源供应商（仅在无会话上下文时作为兜底）.
+
+    启发式（按置信度排序）:
+    - 出现 ``srvtoolu_*`` 格式的 ``tool_use.id`` → zhipu
+    - 出现 ``server_tool_use`` / ``server_tool_use_delta`` 类型的 content block → zhipu
+
+    原则: 只读扫描不修改 body；无匹配返回 None（视作纯净无需跨供应商清洗）。
+
+    Args:
+        body: Anthropic Messages 请求体。
+
+    Returns:
+        推断的源供应商名称（当前仅支持 ``"zhipu"``），无法推断返回 None。
+    """
+    for message in body.get("messages", []):
+        if not isinstance(message, dict):
+            continue
+        content = message.get("content")
+        if not isinstance(content, list):
+            continue
+        for block in content:
+            if not isinstance(block, dict):
+                continue
+            block_type = block.get("type")
+            if block_type in _ZHIPU_SERVER_TOOL_USE_TYPES:
+                return "zhipu"
+            block_id = block.get("id")
+            if isinstance(block_id, str) and _ANTHROPIC_SERVER_TOOL_USE_ID_RE.match(
+                block_id
+            ):
+                return "zhipu"
+            tool_use_id = block.get("tool_use_id")
+            if isinstance(tool_use_id, str) and _ANTHROPIC_SERVER_TOOL_USE_ID_RE.match(
+                tool_use_id
+            ):
+                return "zhipu"
+    return None
+
+
 # ── copilot → zhipu 转换通道 ─────────────────────────────────────
 
 
@@ -316,17 +460,27 @@ def prepare_zhipu_to_copilot(
     prepared = copy.deepcopy(body)
     adaptations: list[str] = []
 
-    # Step 1: 剥离 thinking/redacted_thinking 块
+    # Step 1: 剥离 zhipu 私有 content block 类型
+    removed_vendor_blocks = _remove_vendor_blocks(prepared, _ZHIPU_VENDOR_BLOCK_TYPES)
+    if removed_vendor_blocks:
+        adaptations.append(f"removed_{removed_vendor_blocks}_zhipu_vendor_blocks")
+
+    # Step 2: 改写 srvtoolu_* ID 与 server_tool_use 类型
+    rewritten, _ = _rewrite_srvtoolu_ids(prepared)
+    if rewritten:
+        adaptations.append(f"rewritten_{rewritten}_srvtoolu_ids")
+
+    # Step 3: 剥离 thinking/redacted_thinking 块
     stripped = strip_thinking_blocks(prepared)
     if stripped:
         adaptations.append(f"stripped_{stripped}_thinking_blocks")
 
-    # Step 2: 移除 cache_control 字段
+    # Step 4: 移除 cache_control 字段
     removed_cc = _strip_cache_control(prepared)
     if removed_cc:
         adaptations.append(f"removed_{removed_cc}_cache_control_fields")
 
-    # Step 3: 强制 tool_use/tool_result 配对
+    # Step 5: 强制 tool_use/tool_result 配对
     pairing_fixes = enforce_anthropic_tool_pairing(prepared.get("messages", []))
     if pairing_fixes:
         adaptations.extend(pairing_fixes)
@@ -343,14 +497,18 @@ def prepare_zhipu_to_anthropic(
     """zhipu → anthropic 转换: 清理 zhipu 产物以适配 Anthropic API.
 
     Anthropic API 要求:
+    - tool_use 类型与 ``toolu_*`` 格式 ID（zhipu 的 ``server_tool_use``/``srvtoolu_*`` 不兼容）
     - 每个 tool_use 必须在紧随的 user 消息中有对应 tool_result
     - thinking blocks 的 signature 必须是 Anthropic 签发（zhipu 签发的无效）
+    - 不接受 ``server_tool_use_delta`` 等 zhipu 私有流式块类型
 
-    此通道执行两项变换:
-    1. enforce_anthropic_tool_pairing: 单遍正向扫描修复配对
-    2. strip_thinking_blocks: 移除非 Anthropic 签发的 thinking 块
+    此通道按顺序执行:
+    1. 剥离 zhipu 私有 block 类型（``server_tool_use_delta``）
+    2. 改写 ``srvtoolu_*`` ID 与 ``server_tool_use`` 类型为标准 Anthropic 形式
+    3. 强制 tool_use/tool_result 配对（单遍正向扫描）
+    4. 剥离 thinking blocks（signature 无效）
 
-    两项变换均为幂等操作，安全地在已清理的请求体上重复执行。
+    所有变换均为幂等操作，安全地在已清理的请求体上重复执行。
 
     Returns:
         (prepared_body, adaptations) — adaptations 为应用的变换描述列表。
@@ -358,12 +516,22 @@ def prepare_zhipu_to_anthropic(
     prepared = copy.deepcopy(body)
     adaptations: list[str] = []
 
-    # Step 1: 强制 tool_use/tool_result 配对
+    # Step 1: 剥离 zhipu 私有 content block 类型（如 server_tool_use_delta）
+    removed_vendor_blocks = _remove_vendor_blocks(prepared, _ZHIPU_VENDOR_BLOCK_TYPES)
+    if removed_vendor_blocks:
+        adaptations.append(f"removed_{removed_vendor_blocks}_zhipu_vendor_blocks")
+
+    # Step 2: 改写 srvtoolu_* ID 与 server_tool_use 类型
+    rewritten, _ = _rewrite_srvtoolu_ids(prepared)
+    if rewritten:
+        adaptations.append(f"rewritten_{rewritten}_srvtoolu_ids")
+
+    # Step 3: 强制 tool_use/tool_result 配对
     pairing_fixes = enforce_anthropic_tool_pairing(prepared.get("messages", []))
     if pairing_fixes:
         adaptations.extend(pairing_fixes)
 
-    # Step 2: 剥离 thinking blocks（zhipu signature 无效）
+    # Step 4: 剥离 thinking blocks（zhipu signature 无效）
     stripped = strip_thinking_blocks(prepared)
     if stripped:
         adaptations.append(f"stripped_{stripped}_thinking_blocks")

{coding_proxy-0.2.4a5 → coding_proxy-0.3.0a1}/src/coding/proxy/logging/db.py

@@ -166,7 +166,11 @@ CREATE TABLE IF NOT EXISTS usage_log (
     success BOOLEAN NOT NULL DEFAULT 1,
     failover BOOLEAN NOT NULL DEFAULT 0,
     failover_from TEXT DEFAULT NULL,
-    request_id TEXT DEFAULT ''
+    request_id TEXT DEFAULT '',
+    client_category TEXT NOT NULL DEFAULT 'cc',
+    operation TEXT NOT NULL DEFAULT '',
+    endpoint TEXT NOT NULL DEFAULT '',
+    extra_usage_json TEXT NOT NULL DEFAULT '{}'
 );
 CREATE TABLE IF NOT EXISTS usage_evidence (
     id INTEGER PRIMARY KEY AUTOINCREMENT,
@@ -188,6 +192,8 @@ CREATE TABLE IF NOT EXISTS usage_evidence (
 _CREATE_INDEXES = """
 CREATE INDEX IF NOT EXISTS idx_usage_ts ON usage_log(ts);
 CREATE INDEX IF NOT EXISTS idx_usage_vendor ON usage_log(vendor);
+CREATE INDEX IF NOT EXISTS idx_usage_client_category ON usage_log(client_category);
+CREATE INDEX IF NOT EXISTS idx_usage_operation ON usage_log(operation);
 CREATE INDEX IF NOT EXISTS idx_usage_evidence_request_id ON usage_evidence(request_id);
 CREATE INDEX IF NOT EXISTS idx_usage_evidence_vendor ON usage_evidence(vendor);
 """
@@ -196,25 +202,29 @@ CREATE INDEX IF NOT EXISTS idx_usage_evidence_vendor ON usage_evidence(vendor);
 
 _PERIOD_SQL: dict[TimePeriod, tuple[str, str, str]] = {
     # (date_expr, group_by, order_by_expr)
+    # group_by / order_by 附带 client_category, operation：
+    # 1) 新列加入 SELECT 后必须对应 GROUP BY，否则 SQLite 会返回任意未确定值；
+    # 2) 历史行默认 ('cc', '') → 对既有聚合口径零回归（同值无额外拆分）；
+    # 3) 当 client_category='api' 场景产生混合数据时，按类别 + 操作拆行可直观区分。
     TimePeriod.DAY: (
         "local_date(ts) AS date",
-        "local_date(ts), vendor, model_served",
-        "local_date(ts) DESC, vendor, model_served",
+        "local_date(ts), vendor, model_served, client_category, operation",
+        "local_date(ts) DESC, vendor, model_served, client_category, operation",
     ),
     TimePeriod.WEEK: (
         "local_week(ts) AS date",
-        "local_week(ts), vendor, model_served",
-        "local_week(ts) DESC, vendor, model_served",
+        "local_week(ts), vendor, model_served, client_category, operation",
+        "local_week(ts) DESC, vendor, model_served, client_category, operation",
     ),
     TimePeriod.MONTH: (
         "local_month(ts) AS date",
-        "local_month(ts), vendor, model_served",
-        "local_month(ts) DESC, vendor, model_served",
+        "local_month(ts), vendor, model_served, client_category, operation",
+        "local_month(ts) DESC, vendor, model_served, client_category, operation",
     ),
     TimePeriod.TOTAL: (
         "NULL AS date",
-        "vendor, model_served",
-        "vendor, model_served",
+        "vendor, model_served, client_category, operation",
+        "vendor, model_served, client_category, operation",
     ),
 }
 
@@ -236,6 +246,7 @@ class TokenLogger:
         # 迁移必须在建索引之前执行，确保 vendor 列已存在
         await self._migrate_rename_backend_to_vendor()
         await self._migrate_add_failover_from()
+        await self._migrate_add_native_columns()
         await self._db.executescript(_CREATE_INDEXES)
         # 注册时区感知的日期函数：将 UTC 时间戳转为本地时间维度
         await self._db.create_function("local_date", 1, _local_date_udf)
@@ -255,6 +266,26 @@ class TokenLogger:
             )
             logger.info("Migration: added failover_from column to usage_log")
 
+    async def _migrate_add_native_columns(self) -> None:
+        """幂等迁移：为已有数据库添加原生 API 透传通道所需的四列.
+
+        历史行自动得到：client_category='cc'、operation=''、endpoint=''、extra_usage_json='{}'。
+        """
+        if not self._db:
+            return
+        cursor = await self._db.execute("PRAGMA table_info(usage_log)")
+        columns = {row["name"] for row in await cursor.fetchall()}
+        specs = [
+            ("client_category", "TEXT NOT NULL DEFAULT 'cc'"),
+            ("operation", "TEXT NOT NULL DEFAULT ''"),
+            ("endpoint", "TEXT NOT NULL DEFAULT ''"),
+            ("extra_usage_json", "TEXT NOT NULL DEFAULT '{}'"),
+        ]
+        for name, ddl in specs:
+            if name not in columns:
+                await self._db.execute(f"ALTER TABLE usage_log ADD COLUMN {name} {ddl}")
+                logger.info("Migration: added %s column to usage_log", name)
+
     async def _migrate_rename_backend_to_vendor(self) -> None:
         """幂等迁移：重命名 backend 列为 vendor."""
         if not self._db:
@@ -284,6 +315,10 @@ class TokenLogger:
         failover: bool = False,
         failover_from: str | None = None,
         request_id: str = "",
+        client_category: str = "cc",
+        operation: str = "",
+        endpoint: str = "",
+        extra_usage_json: str = "{}",
     ) -> None:
         if not self._db:
             return
@@ -292,8 +327,9 @@ class TokenLogger:
                (vendor, model_requested, model_served,
                 input_tokens, output_tokens,
                 cache_creation_tokens, cache_read_tokens,
-                duration_ms, success, failover, failover_from, request_id)
-               VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)""",
+                duration_ms, success, failover, failover_from, request_id,
+                client_category, operation, endpoint, extra_usage_json)
+               VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)""",
             (
                 vendor,
                 model_requested,
@@ -307,6 +343,10 @@ class TokenLogger:
                 failover,
                 failover_from,
                 request_id,
+                client_category,
+                operation,
+                endpoint,
+                extra_usage_json,
             ),
         )
         await self._db.commit()
@@ -376,6 +416,9 @@ class TokenLogger:
         count: int = 7,
         vendor: str | list[str] | None = None,
         model: str | list[str] | None = None,
+        client_category: str | list[str] | None = None,
+        operation: str | list[str] | None = None,
+        endpoint: str | list[str] | None = None,
     ) -> list[dict]:
         """按指定时间维度聚合 Token 使用统计.
 
@@ -385,6 +428,9 @@ class TokenLogger:
                    ``TOTAL`` 维度下忽略此参数。
             vendor: 过滤供应商，支持单个字符串或字符串列表（多 vendor 过滤）。
             model: 过滤实际服务模型（model_served），支持单个字符串或字符串列表。
+            client_category: 过滤客户端类别（``'cc'`` / ``'api'``）。
+            operation: 过滤规范化操作名（``'chat'`` / ``'embedding'`` ...）。
+            endpoint: 过滤原始上游路径（``'/v1/chat/completions'`` ...）。
         """
         if not self._db:
             return []
@@ -394,6 +440,8 @@ class TokenLogger:
         sql = f"""SELECT {date_expr}, vendor,
                    GROUP_CONCAT(DISTINCT model_requested) AS model_requested,
                    model_served,
+                   client_category,
+                   operation,
                    COUNT(*) AS total_requests,
                    SUM(input_tokens) AS total_input,
                    SUM(output_tokens) AS total_output,
@@ -420,6 +468,25 @@ class TokenLogger:
             placeholders = ",".join("?" * len(models))
             sql += f" AND model_served IN ({placeholders})"
             params.extend(models)
+        if client_category:
+            cats = (
+                [client_category]
+                if isinstance(client_category, str)
+                else client_category
+            )
+            placeholders = ",".join("?" * len(cats))
+            sql += f" AND client_category IN ({placeholders})"
+            params.extend(cats)
+        if operation:
+            ops = [operation] if isinstance(operation, str) else operation
+            placeholders = ",".join("?" * len(ops))
+            sql += f" AND operation IN ({placeholders})"
+            params.extend(ops)
+        if endpoint:
+            eps = [endpoint] if isinstance(endpoint, str) else endpoint
+            placeholders = ",".join("?" * len(eps))
+            sql += f" AND endpoint IN ({placeholders})"
+            params.extend(eps)
 
         sql += f" GROUP BY {group_clause} ORDER BY {order_clause}"