oadmin/sub2api
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity
1,361 Commits 2 Branches 0 Tags
e1a68497d63775222772cf435a08fef057d9bfdb
Commit Graph

110 Commits

Author SHA1 Message Date
erio
e1a68497d6 refactor: simplify sticky session rate limit handling — switch immediately on any rate limit 
Remove threshold-based waiting in both sticky session and antigravity
pre-check paths. When a model is rate-limited, immediately clear the
sticky session and switch accounts instead of waiting for short durations.
 2026-02-07 17:06:49 +08:00
erio
2656320d04 fix(antigravity): fetch default mapping from API and sync Redis on rate limit 
1. Frontend: replace hardcoded antigravityDefaultMappings with async
   fetch from GET /admin/accounts/antigravity/default-model-mapping,
   eliminating the duplicate data source that caused frontend/backend
   mapping inconsistency.

2. Backend: convert handleSmartRetry and antigravityRetryLoop from
   standalone functions to AntigravityGatewayService methods, enabling
   Redis cache sync (updateAccountModelRateLimitInCache) after both
   rate-limit write paths — long-delay branch and retry-exhausted branch.
 2026-02-07 15:59:27 +08:00
erio
de0927289e fix(antigravity): support upstream accounts and custom model_mapping in scheduling 
- GetAccessToken: add upstream branch to read api_key from credentials
- shouldTriggerAntigravitySmartRetry: relax check from IsOAuth to Platform-based
- isModelSupportedByAccount/WithContext: replace IsAntigravityModelSupported
  whitelist with mapAntigravityModel for unified scheduling/forwarding logic
- mapAntigravityModel: fix edge case where wildcard target equals request model
- Update tests for new behavior and add custom model_mapping test cases
 2026-02-07 14:32:08 +08:00
erio
5e98445b22 feat(antigravity): comprehensive enhancements - model mapping, rate limiting, scheduling & ops 
Key changes:
- Upgrade model mapping: Opus 4.5 → Opus 4.6-thinking with precise matching
- Unified rate limiting: scope-level → model-level with Redis snapshot sync
- Load-balanced scheduling by call count with smart retry mechanism
- Force cache billing support
- Model identity injection in prompts with leak prevention
- Thinking mode auto-handling (max_tokens/budget_tokens fix)
- Frontend: whitelist mode toggle, model mapping validation, status indicators
- Gemini session fallback with Redis Trie O(L) matching
- Ops: enhanced concurrency monitoring, account availability, retry logic
- Migration scripts: 049-051 for model mapping unification
 2026-02-07 12:31:10 +08:00
erio
8917afab2a fix(antigravity): reduce 429 fallback cooldown from 5min to 30s 
The default fallback cooldown when rate limit reset time cannot be
parsed was 5 minutes, which is too aggressive and causes accounts
to be unnecessarily locked out. Reduce to 30 seconds for faster
recovery. Config override still works (unit remains minutes).
 2026-02-07 11:54:00 +08:00
shaw
5299f3dcf6 fix: ix: antigravity 添加 aude-opus-4-6-thinking 模型支持 2026-02-07 10:38:10 +08:00
shaw
39e05a2dad feat: 新增全局错误透传规则功能 
支持管理员配置上游错误如何返回给客户端：
- 新增 ErrorPassthroughRule 数据模型和 Ent Schema
- 实现规则的 CRUD API（/admin/error-passthrough-rules）
- 支持按错误码、关键词匹配，支持 any/all 匹配模式
- 支持按平台过滤（anthropic/openai/gemini/antigravity）
- 支持透传或自定义响应状态码和错误消息
- 实现两级缓存（Redis + 本地内存）和多实例同步
- 集成到 gateway_handler 的错误处理流程
- 新增前端管理界面组件
- 新增单元测试覆盖核心匹配逻辑

优化：
- 移除 refreshLocalCache 中的冗余排序（数据库已排序）
- 后端 Validate() 增加匹配条件非空校验
 2026-02-05 21:52:54 +08:00
Wesley Liddick
804b6f2282 Merge pull request #468 from s-Joshua-s/fix/thinking-block-modification-error 
fix(api): 修复 thinking 块被意外修改导致的 400 错误
 2026-02-03 22:21:06 +08:00
song
7cb5444dbb fix: update tests for group fallback 2026-02-03 16:48:52 +08:00
song
3bede6e65f merge upstream main 2026-02-03 16:21:58 +08:00
JIA-ss
ad90bb4645 fix(api): 修复 thinking 块被意外修改导致的 400 错误 
问题描述：
使用扩展思考功能时，偶现以下错误：
"thinking or redacted_thinking blocks in the latest assistant message cannot be modified"

根因分析：
当代理服务修改请求体中的某些字段时（如 metadata.user_id、model），
使用 map[string]any 解析整个 JSON 后重新序列化，导致：
1. 字段顺序改变（Go map 序列化按字母排序）
2. 数字格式变化（如 1.0 → 1）
3. Unicode 转义变化

Claude API 对 thinking 块进行字节级验证，任何变化都会触发错误。

修复内容：
1. identity_service.go - RewriteUserID/RewriteUserIDWithMasking
   使用 json.RawMessage 保留其他字段的原始字节

2. gateway_service.go - replaceModelInBody
   使用 json.RawMessage 保留其他字段的原始字节

3. gateway_service.go - normalizeClaudeOAuthRequestBody
   保留 messages 的原始字节，跳过包含 thinking 块的消息修改

4. gateway_service.go - isThinkingBlockSignatureError
   添加 "cannot be modified" 错误检测，触发自动重试

5. antigravity_gateway_service.go - isSignatureRelatedError
   添加 "cannot be modified" 错误检测

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-02-03 16:15:37 +08:00
song
2220fd18ca merge upstream main 2026-02-03 15:36:17 +08:00
Wesley Liddick
6e54eda41f Merge pull request #464 from touwaeriol/pr/antigravity-scope-ratelimit 
feat(antigravity): 支持按配额域(scope)级别限流
 2026-02-03 15:02:15 +08:00
liuxiongfeng
56949a58bc feat(antigravity): 默认开启按配额域限流，避免整个账号被锁定 
将 GATEWAY_ANTIGRAVITY_429_SCOPE_LIMIT 的默认值从关闭改为开启。
当 Gemini 模型触发 429 限流时，只会限制对应的配额域（gemini_text），
而 Claude 和 gemini_image 仍可继续使用，提高账号利用率。
 2026-02-03 14:25:30 +08:00
liuxiongfeng
7d256879c5 feat(antigravity): map all gemini-2.5 to gemini-3 series 
Antigravity 上游不再支持 gemini-2.5 系列，统一映射到 gemini-3：
- gemini-2.5-flash → gemini-3-flash
- gemini-2.5-flash-lite → gemini-3-flash
- gemini-2.5-flash-thinking → gemini-3-flash
- gemini-2.5-flash-image → gemini-3-pro-image
- gemini-2.5-pro → gemini-3-pro-high
- gemini-2.5-pro-preview → gemini-3-pro-high
- gemini-2.5-pro-exp → gemini-3-pro-high
 2026-02-03 14:23:47 +08:00
liuxiongfeng
beb63cb152 feat(antigravity): map gemini-2.5-pro to gemini-3-pro-high 
Add prefix mapping rules for gemini-2.5-pro variants:
- gemini-2.5-pro -> gemini-3-pro-high
- gemini-2.5-pro-preview -> gemini-3-pro-high
- gemini-2.5-pro-exp -> gemini-3-pro-high
 2026-02-03 14:23:47 +08:00
song
3ecadf4aad chore: apply stashed changes 2026-02-02 22:20:08 +08:00
song
0170d19fa7 merge upstream main 2026-02-02 22:13:50 +08:00
song
7ade9baa15 fix(gateway): 过滤 Gemini 请求中 parts 为空的消息 
Gemini API 不接受 contents 数组中 parts 为空的消息，会返回 400 INVALID_ARGUMENT 错误。
添加 filterEmptyPartsFromGeminiRequest 函数在转发前过滤这类消息。

影响范围：ForwardGemini (antigravity) 和 ForwardNative (gemini)
 2026-01-29 21:09:33 +08:00
song
5b787334c8 antigravity: 转发优先 daily 2026-01-28 11:17:39 +08:00
song
f761afb1ef antigravity: 区分切换后重试次数 2026-01-28 00:01:03 +08:00
song
877c17251d feat(group): 添加 MCP XML 注入开关 
- Group 新增 mcp_xml_inject 字段，控制 Antigravity 平台的 MCP XML 协议注入
- 默认启用，可在分组设置中关闭
- 修复 GetByKeyForAuth 遗漏查询 mcp_xml_inject 字段导致认证缓存值始终为 false 的问题
 2026-01-27 13:09:56 +08:00
song
fd0370c07a Add invalid-request fallback routing 2026-01-23 22:24:46 +08:00
0xff26b9a8
4f0c2b794c style: gofmt antigravity_gateway_service.go 2026-01-22 14:38:55 +08:00
0xff26b9a8
e756064c19 fix(antigravity): 修复非流式 Claude To Antigravity 响应内容为空的问题 
- 修复 TransformGeminiToClaude 的 JSON 解析逻辑，当 V1InternalResponse
  解析成功但 candidates 为空时，尝试直接解析为 GeminiResponse 格式
- 修复 handleClaudeStreamToNonStreaming 收集流式响应的逻辑，累积所有
  chunks 的内容而不是只保留最后一个（最后一个 chunk 通常 text 为空）
- 新增 mergeCollectedPartsToResponse 函数，合并所有类型的 parts
  （text、thinking、functionCall、inlineData），保持原始顺序
- 连续的普通 text parts 合并为一个，thinking/functionCall/inlineData 保持原样
 2026-01-22 14:17:59 +08:00
song
207e09500a feat(antigravity): 支持按模型类型配置重试次数 
新增环境变量：
- GATEWAY_ANTIGRAVITY_MAX_RETRIES_CLAUDE
- GATEWAY_ANTIGRAVITY_MAX_RETRIES_GEMINI_TEXT
- GATEWAY_ANTIGRAVITY_MAX_RETRIES_GEMINI_IMAGE

未设置时回退到平台级 GATEWAY_ANTIGRAVITY_MAX_RETRIES
 2026-01-21 20:48:36 +08:00
0xff26b9a8
71f8b9e473 refactor(antigravity): 提取并同步 Schema 清理逻辑至 schema_cleaner.go 
主要变更：
1. 重构代码结构：
   - 将 CleanJSONSchema 及其相关辅助函数从 request_transformer.go 提取到独立的 schema_cleaner.go 文件中，实现逻辑解耦。

2. 逻辑优化与修正：
   - 参考 Antigravity-Manager (json_schema.rs) 的实现逻辑，修正了 Schema 清洗策略。
 2026-01-21 12:08:16 +08:00
0xff26b9a8
da48df06d2 refactor(antigravity): 提取并同步 Schema 清理逻辑至 schema_cleaner.go 
主要变更：
1. 重构代码结构：
   - 将 CleanJSONSchema 及其相关辅助函数从 request_transformer.go 提取到独立的 schema_cleaner.go 文件中，实现逻辑解耦。

2. 逻辑优化与修正：
   - 参考 Antigravity-Manager (json_schema.rs) 的实现逻辑，修正了 Schema 清洗策略。
 2026-01-20 23:41:53 +08:00
song
549c134bb8 chore: gofmt antigravity gateway service 2026-01-20 19:16:43 +08:00
song
d206721fc1 feat: make antigravity max retries configurable 2026-01-20 19:12:19 +08:00
song
86d63f919d feat(antigravity): 支持秒级 fallback 冷却时间 2026-01-20 11:38:40 +08:00
song
c43aa22cdb feat(antigravity): 支持按映射模型计费 2026-01-20 11:02:08 +08:00
song
d1a6303e49 fix(antigravity): 修复 Claude 非流式响应丢失 2026-01-20 00:52:27 +08:00
song
8b071cc665 fix(antigravity): restore signature retry and base order 2026-01-17 22:50:50 +08:00
song
959f6c538a fix(antigravity): remove thinking sanitation 2026-01-17 22:21:48 +08:00
song
217b3b59c0 fix(antigravity): drop MarkUnavailable 2026-01-17 21:59:32 +08:00
song
ec916a3197 fix(antigravity): remove signature retry 2026-01-17 21:56:57 +08:00
song
22eb72e0f9 fix(antigravity): restore url fallback behavior 2026-01-17 21:50:09 +08:00
song
07ba64c666 fix(antigravity): handle url-level 429 without failover 2026-01-17 21:37:32 +08:00
song
f22bc59fe3 fix(antigravity): route signature retry through url fallback 2026-01-17 21:15:33 +08:00
song
5e9f5efbe3 chore: log antigravity signature retry 429 2026-01-17 18:22:53 +08:00
song
a7a0017aa8 chore: gofmt antigravity gateway service 2026-01-17 18:22:43 +08:00
song
b9b4db3df5 Merge upstream/main 2026-01-17 18:00:07 +08:00
song
5a6f60a954 fix(antigravity): 区分 URL 级别和账户配额级别的 429 限流 
- "Resource has been exhausted" → URL 级别限流，立即切换 URL
- "exhausted your capacity on this model" → 账户配额限流，重试 3 次（指数退避）后标记限流
 2026-01-17 11:11:18 +08:00
song
31933c8a60 fix: 删除未使用的字段修复 lint 错误 2026-01-17 10:40:28 +08:00
song
78bccd032d refactor(antigravity): 提取公共重试循环函数减少重复代码 
- 新增 antigravityRetryLoop 函数统一处理 Forward 和 ForwardGemini 的重试逻辑
- 429 日志增加 base_url 字段便于调试
- 删除重复的 shouldRetryUpstreamError 方法
 2026-01-17 10:28:31 +08:00
song
ac7503d95f fix(antigravity): 429 时也切换 URL 重试 
- 429 优先切换到下一个 URL 重试
- 只有所有 URL 都返回 429 时才限流账户并返回错误
- 与 client.go 中的逻辑保持一致
 2026-01-17 02:14:57 +08:00
song
69c4b17a9b feat(antigravity): 动态 URL 排序，最近成功的优先使用 
- URLAvailability 新增 lastSuccess 字段追踪最近成功的 URL
- GetAvailableURLs 返回列表时优先放置 lastSuccess
- 所有 Antigravity API 调用成功后调用 MarkSuccess 更新优先级
 2026-01-17 01:54:14 +08:00
song
34d6b0a601 feat(gateway): 账户切换次数和 Antigravity 限流时间可配置 
- gateway.max_account_switches: 账户切换最大次数，默认 10
- gateway.max_account_switches_gemini: Gemini 账户切换次数，默认 3
- gateway.antigravity_fallback_cooldown_minutes: Antigravity 429 fallback 限流时间，默认 5 分钟
- Antigravity 429 不再重试，直接标记账户限流
 2026-01-16 20:18:30 +08:00
song
2055a60bcb fix(antigravity): 429 重试3次后限流账户 
- 收到429后重试最多3次（指数退避）
- 3次都失败后调用 handleUpstreamError 限流账户
- 移除无效的 URL fallback 逻辑（当前只有一个URL）
 2026-01-16 18:51:07 +08:00