docs: 更新文档至 v0.6.1(覆盖 v0.4.2-v0.6.1 全量检查)#1
Conversation
更新内容: - 新增 Langfuse LLM 可观测性集成文档 - 新增 changelog v0.5.3 ~ v0.6.1 全部版本记录 - 新增环境变量:ENABLE_ENDPOINT_CIRCUIT_BREAKER、Langfuse 系列配置 - 更新供应商管理:定时启停字段、批量操作撤销功能 - 更新 Webhook 通知:缓存率异常告警 - 更新数据导入导出:ledgerOnly 仅账本导出 - 更新用户管理:客户端限制重构、登录鉴权重构说明 - 更新代理设置:Swap 缓存 TTL 计费说明 - 更新监控日志:决策链溯源 - 修复:移除不存在的 ENABLE_MULTI_PROVIDER_TYPES 环境变量引用 - 更新 FETCH_BODY_TIMEOUT/FETCH_HEADERS_TIMEOUT 默认值
📝 WalkthroughWalkthrough该 PR 主要为文档与导航更新:新增 v0.6.1/0.6.0 发布说明、Langfuse 集成页面与导航项,扩展环境变量与提供者字段(定时启停、groupPriorities、缓存 TTL 计费、批量撤销、ledgerOnly 导出、缓存率告警、客户端黑名单等)。仅文档与一处导航变更,无运行逻辑修改。 Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes 🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! 此拉取请求旨在将文档更新至 v0.6.0 和 v0.6.1 版本,详细记录了自 v0.5.3 以来的所有新功能、优化和修复。核心更新包括引入 Langfuse LLM 可观测性集成,增强了系统监控和追踪能力,并对供应商管理、数据处理和用户认证等多个模块进行了功能扩展和安全性改进,确保文档与最新版本的功能保持一致。 Highlights
Changelog
Activity
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
|
|
||
| ### 新增 | ||
|
|
||
| - 扩展只读密钥访问权限,支持更多 API 端点访问 (#704) [@AptS:1547](https://github.com/AptS1547) |
There was a problem hiding this comment.
| | 性能指标 | `durationMs`, `ttfbMs` | 总耗时和首字节时间 | | ||
| | 错误信息 | `errorMessage`, `errorStack`, `errorCause`, `statusCode` | 错误详情、堆栈、原因和 HTTP 状态码 | | ||
| | 供应商决策 | `providerChain` | 供应商选择决策链(JSONB) | | ||
| | 决策链溯源 | origin chain | 被复用请求可查看 Session 原始决策过程(v0.6.0+) | |
There was a problem hiding this comment.
| ``` | ||
| 客户端请求 | ||
| ↓ | ||
| 创建 Langfuse Trace(关联 request-id、user-id、model 等元数据) |
There was a problem hiding this comment.
Actionable comments posted: 6
🧹 Nitpick comments (1)
src/app/docs/monitoring/logs/page.md (1)
62-62: 中文语法建议(可选)静态分析工具提示第 62 行的表述"被复用请求可查看"可能存在语法问题,建议改为"可被复用的请求查看"或"复用的请求可以查看"。不过,原文在上下文中是可以理解的,此建议仅供参考。
类似的提示也出现在第 712 行的"被复用的请求还可以"。
如果您认为当前表述已足够清晰,可以忽略此建议。
📝 可选的改写方案
第 62 行:
-| 决策链溯源 | origin chain | 被复用请求可查看 Session 原始决策过程(v0.6.0+) | +| 决策链溯源 | origin chain | 复用的请求可以查看 Session 原始决策过程(v0.6.0+) |第 712 行:
-A: 使用 Session ID 筛选,按时间排序。查看 `requestSequence` 了解请求顺序,`providerChain` 了解供应商选择过程。自 v0.6.0 起,被复用的请求还可以通过决策链溯源功能查看 Session 的原始决策过程。 +A: 使用 Session ID 筛选,按时间排序。查看 `requestSequence` 了解请求顺序,`providerChain` 了解供应商选择过程。自 v0.6.0 起,复用的请求还可以通过决策链溯源功能查看 Session 的原始决策过程。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/app/docs/monitoring/logs/page.md` at line 62, Update the Chinese wording in the markdown table cell that currently reads "被复用请求可查看 Session 原始决策过程(v0.6.0+)" to a grammatically clearer form such as "复用的请求可以查看 Session 原始决策过程(v0.6.0+)" (and likewise change the similar phrase at the other occurrence "被复用的请求还可以" to "复用的请求还可以" or "可被复用的请求还可以") so both the table entry containing "决策链溯源 | origin chain | 被复用请求可查看 Session 原始决策过程(v0.6.0+)" and the later occurrence with "被复用的请求还可以" use the improved phrasing.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@src/app/docs/changelog/page.md`:
- Around line 79-236: Update the version headings that currently use the plain
"## vX.Y.Z (YYYY-MM-DD)" format (e.g. "## v0.5.8 (2026-02-15)", "## v0.5.7
(2026-02-14)", "## v0.5.6 (2026-02-12)", "## v0.5.5 (2026-02-11)", "## v0.5.4
(2026-02-07)", "## v0.5.3 (2026-02-03)") to the clickable release link format
shown elsewhere: "## [vX.Y.Z](release-url) - YYYY-MM-DD" (replace release-url
with the actual release permalink and keep the original date after a hyphen),
ensuring every version heading follows this unified pattern.
- Line 161: The changelog entry "Vacuum Filter has 热路径性能优化,降低 API Key 负向短路成本
(`#757`) [`@tesgth032`]" mixes English and Chinese awkwardly; replace that exact
line with a coherent phrasing—either fully Chinese (e.g., "Vacuum Filter
在热路径上进行了性能优化,降低 API Key 触发的负向短路成本") or a clear bilingual sentence (e.g., "Vacuum
Filter: hot-path performance optimizations, reducing API Key negative
short‑circuit cost") so the wording flows naturally.
In `@src/app/docs/providers/crud/page.md`:
- Line 41: 当前文件中出现的标记 "{% /table %}" 表示此处可能是表格结束或表格行被误写;如果这是表格行,请把该行改为与表格其它行一致的
Markdown 表格格式——以竖线开头和结尾并包含 3 列(例如 "| col1 | col2 | col3 |"),保证列数与表头分隔行一致;如果 "{%
/table %}" 是表格结束标记,请恢复为正确的结束标记并检查前一行是否为合法的表格行(补上缺失的前导/尾随管道符和缺失的列分隔符)以修复列数不匹配问题。
In `@src/app/docs/reference/env-variables/page.md`:
- Around line 478-485: Update the table entry for the `true` value so it
explicitly states the switch affects both provider-type-level and endpoint-level
circuit breakers (e.g., "启用供应商类型级和端点级熔断器,连续失败的端点/供应商会被临时屏蔽(默认 3 次失败后熔断 5 分钟)")
and keep the existing {% callout type="note" %} text but remove redundant
phrasing so both places consistently say "双层生效:供应商类型级和端点级". Apply this change by
editing the table row containing `true` and the callout block to use the unified
wording.
In `@src/app/docs/system/langfuse/page.md`:
- Around line 78-90: The fenced code block in
src/app/docs/system/langfuse/page.md is missing a language identifier which can
trigger markdownlint MD040; update the fenced block starting at the diagram so
it explicitly specifies a language (e.g., change the opening fence from ``` to
```text) to silence the lint warning and ensure consistent rendering.
In `@src/app/docs/users/access-restrictions/page.md`:
- Line 113: 在文档中已新增字段 blockedClients,但未在 USER_FIELD_PERMISSIONS
和示例验证规则中同步补全,导致前后不一致;请在 USER_FIELD_PERMISSIONS 的权限映射中为 blockedClients
添加相应权限条目(例如 read/write/deny 等与其它字段一致的权限配置),并在校验规则示例(validation rules)中加入
blockedClients 的类型与约束说明(如 string[]、项格式或长度限制、必填/可选规则),确保字段名为 blockedClients
并与已有字段的示例格式保持一致以避免歧义。
---
Nitpick comments:
In `@src/app/docs/monitoring/logs/page.md`:
- Line 62: Update the Chinese wording in the markdown table cell that currently
reads "被复用请求可查看 Session 原始决策过程(v0.6.0+)" to a grammatically clearer form such as
"复用的请求可以查看 Session 原始决策过程(v0.6.0+)" (and likewise change the similar phrase at
the other occurrence "被复用的请求还可以" to "复用的请求还可以" or "可被复用的请求还可以") so both the
table entry containing "决策链溯源 | origin chain | 被复用请求可查看 Session 原始决策过程(v0.6.0+)"
and the later occurrence with "被复用的请求还可以" use the improved phrasing.
ℹ️ Review info
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Cache: Disabled due to Reviews > Disable Cache setting
📒 Files selected for processing (14)
src/app/docs/changelog/page.mdsrc/app/docs/monitoring/logs/page.mdsrc/app/docs/page.mdsrc/app/docs/providers/batch-operations/page.mdsrc/app/docs/providers/crud/page.mdsrc/app/docs/proxy/cache-ttl/page.mdsrc/app/docs/reference/env-variables/page.mdsrc/app/docs/reference/provider-types/page.mdsrc/app/docs/system/data-import-export/page.mdsrc/app/docs/system/langfuse/page.mdsrc/app/docs/system/webhook/page.mdsrc/app/docs/users/access-restrictions/page.mdsrc/app/docs/users/login-control/page.mdsrc/lib/navigation.ts
| | `group_tag` | 分组标签 | "production", "backup" | | ||
| | `active_time_start` | 定时启动时间(HH:mm 格式)| "08:00"(v0.6.0+) | | ||
| | `active_time_end` | 定时停止时间(HH:mm 格式)| "22:00"(v0.6.0+) | | ||
| {% /table %} |
There was a problem hiding this comment.
修复表格格式问题
静态分析工具检测到第 41 行存在表格格式问题:缺少前导和尾随管道符(|),且列数不匹配(预期 3 列,实际 1 列)。请检查该行是否为表格的一部分,如果是,请确保其格式与表格其他行一致。
🔧 可能的修复方案
如果第 41 行应该是表格的一部分,请确保格式如下:
| `active_time_end` | 定时停止时间(HH:mm 格式)| "22:00"(v0.6.0+) |
-{% /table %}
+{% /table %}如果第 41 行是表格结束标记,请确保上一行(第 40 行)的表格行格式正确。
🧰 Tools
🪛 markdownlint-cli2 (0.21.0)
[warning] 41-41: Table pipe style
Expected: leading_and_trailing; Actual: no_leading_or_trailing; Missing leading pipe
(MD055, table-pipe-style)
[warning] 41-41: Table pipe style
Expected: leading_and_trailing; Actual: no_leading_or_trailing; Missing trailing pipe
(MD055, table-pipe-style)
[warning] 41-41: Table column count
Expected: 3; Actual: 1; Too few cells, row will be missing data
(MD056, table-column-count)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/app/docs/providers/crud/page.md` at line 41, 当前文件中出现的标记 "{% /table %}"
表示此处可能是表格结束或表格行被误写;如果这是表格行,请把该行改为与表格其它行一致的 Markdown 表格格式——以竖线开头和结尾并包含 3 列(例如 "|
col1 | col2 | col3 |"),保证列数与表头分隔行一致;如果 "{% /table %}"
是表格结束标记,请恢复为正确的结束标记并检查前一行是否为合法的表格行(补上缺失的前导/尾随管道符和缺失的列分隔符)以修复列数不匹配问题。
| | 值 | 行为 | | ||
| |------|------| | ||
| | `false` | 仅支持 `claude`、`claude-auth`、`codex` | | ||
| | `true` | 额外支持 `gemini`、`gemini-cli`、`openai-compatible` | | ||
| | `false` | 禁用端点熔断器,所有启用的端点均可使用 | | ||
| | `true` | 启用端点熔断器,连续失败的端点会被临时屏蔽(默认 3 次失败后熔断 5 分钟) | | ||
|
|
||
| {% callout type="warning" title="实验性功能" %} | ||
| Gemini CLI、OpenAI Compatible 等类型功能仍在开发中,可能存在不稳定性。 | ||
| 生产环境暂不建议启用。 | ||
| {% callout type="note" %} | ||
| 此开关同时控制供应商类型级别和端点级别的熔断器。启用后,供应商类型和端点的熔断器都会生效。 | ||
| {% /callout %} |
There was a problem hiding this comment.
开关行为描述前后不完全一致,建议在值表中直接说明“双层生效范围”。
Line 480-481 仅描述“端点级”,但 Line 484-485 说明该开关同时影响“供应商类型级 + 端点级”。建议合并表达,减少误解。
✏️ 建议修改
-| `false` | 禁用端点熔断器,所有启用的端点均可使用 |
-| `true` | 启用端点熔断器,连续失败的端点会被临时屏蔽(默认 3 次失败后熔断 5 分钟) |
+| `false` | 禁用供应商类型级与端点级熔断器,所有启用端点按常规参与路由 |
+| `true` | 启用供应商类型级与端点级熔断器;连续失败的端点会被临时屏蔽(默认 3 次失败后熔断 5 分钟) |🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/app/docs/reference/env-variables/page.md` around lines 478 - 485, Update
the table entry for the `true` value so it explicitly states the switch affects
both provider-type-level and endpoint-level circuit breakers (e.g.,
"启用供应商类型级和端点级熔断器,连续失败的端点/供应商会被临时屏蔽(默认 3 次失败后熔断 5 分钟)") and keep the existing {%
callout type="note" %} text but remove redundant phrasing so both places
consistently say "双层生效:供应商类型级和端点级". Apply this change by editing the table row
containing `true` and the callout block to use the unified wording.
| ```typescript | ||
| { | ||
| allowedClients: string[], // 允许的客户端模式列表 | ||
| blockedClients: string[], // 拒绝的客户端模式列表(v0.6.0+) |
There was a problem hiding this comment.
请同步补全 blockedClients 的权限与校验示例,避免文档前后不一致。
当前已在字段定义里引入 blockedClients,但后文 USER_FIELD_PERMISSIONS 和验证规则示例仍未体现该字段,容易误导实施者。建议在这两个代码片段中补上 blockedClients。
✏️ 建议补充示例
export const USER_FIELD_PERMISSIONS = {
allowedClients: { requiredRole: "admin" },
+ blockedClients: { requiredRole: "admin" },
allowedModels: { requiredRole: "admin" },
// ...
}; allowedClients: z
.array(z.string().max(64, "客户端模式长度不能超过64个字符"))
.max(50, "客户端模式数量不能超过50个")
.optional()
.default([]),
+blockedClients: z
+ .array(z.string().max(64, "客户端模式长度不能超过64个字符"))
+ .max(50, "客户端模式数量不能超过50个")
+ .optional()
+ .default([]),🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@src/app/docs/users/access-restrictions/page.md` at line 113, 在文档中已新增字段
blockedClients,但未在 USER_FIELD_PERMISSIONS 和示例验证规则中同步补全,导致前后不一致;请在
USER_FIELD_PERMISSIONS 的权限映射中为 blockedClients 添加相应权限条目(例如 read/write/deny
等与其它字段一致的权限配置),并在校验规则示例(validation rules)中加入 blockedClients 的类型与约束说明(如
string[]、项格式或长度限制、必填/可选规则),确保字段名为 blockedClients 并与已有字段的示例格式保持一致以避免歧义。
reference/provider-fields/page.md: - 新增 Anthropic 参数覆写章节(anthropicMaxTokensPreference、anthropicThinkingBudgetPreference、anthropicAdaptiveThinking) - 新增 Gemini Google Search 偏好章节(geminiGoogleSearchPreference) - 新增供应商按组优先级(groupPriorities) - 新增定时启停配置(activeTimeStart/activeTimeEnd) - 新增客户端限制配置(allowedClients/blockedClients) - 新增交换缓存 TTL 计费(swapCacheTtlBilling) reference/env-variables/page.md: - 新增 API Key 安全配置节(ENABLE_API_KEY_VACUUM_FILTER、ENABLE_API_KEY_REDIS_CACHE、API_KEY_AUTH_CACHE_TTL_SECONDS) - 新增端点探测配置节(ENDPOINT_PROBE_*) - 新增高级配置变量(ENABLE_PROVIDER_CACHE、MESSAGE_REQUEST_WRITE_MODE、DB_POOL_MAX、MAX_RETRY_ATTEMPTS_DEFAULT) - 新增 STORE_SESSION_RESPONSE_BODY 变量 - 修正 STORE_SESSION_MESSAGES 描述(v0.5.1 行为变更:控制脱敏而非存储开关) - 更新概述和生产配置示例
There was a problem hiding this comment.
🧹 Nitpick comments (2)
src/app/docs/reference/provider-fields/page.md (2)
695-709: 交换缓存 TTL 计费文档简洁,可选补充示例。新增的交换缓存 TTL 计费功能(v0.6.0+)说明简洁明了,清楚解释了解决供应商计费不匹配问题的目的。如需保持与其他章节的一致性,可考虑补充 JSON 配置示例。
📝 可选:补充配置示例
启用后,在计费计算时将 1 小时和 5 分钟的缓存 TTL 互换,使计费与供应商的实际收费策略对齐。 + +### 配置示例 + +```json +{ + "swapCacheTtlBilling": true +} +```🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/app/docs/reference/provider-fields/page.md` around lines 695 - 709, Add a small JSON configuration example and consistent formatting to the "交换缓存 TTL 计费" section: update the description for the `swapCacheTtlBilling` field to include a one-line JSON snippet showing how to enable it (e.g., {"swapCacheTtlBilling": true}) and ensure the example is wrapped as a code block and placed immediately after the field table so it matches other sections' structure; reference the `swapCacheTtlBilling` field and the existing section header to find where to insert the example.
672-693: 客户端限制配置文档可补充示例。新增的客户端限制配置章节(v0.6.0 重构)包含了字段说明和匹配规则,但与其他章节相比缺少 JSON 配置示例。建议补充示例以保持文档风格一致性,方便管理员理解
allowedClients和blockedClients的具体配置格式。📝 建议补充的配置示例
系统支持自动检测 Claude Code CLI 子客户端(使用 4 信号检测系统),便于精细化控制不同 IDE 集成的访问权限。 + +### 配置示例 + +```json +{ + "allowedClients": ["vscode", "cursor"], + "blockedClients": ["windsurf"] +} +```🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/app/docs/reference/provider-fields/page.md` around lines 672 - 693, 在“客户端限制配置”章节中补充一个 JSON 配置示例来示范 allowedClients 和 blockedClients 的用法;在 page.md 的该节(涉及字段 allowedClients、blockedClients 及匹配规则)添加一个简短的代码块示例,展示 allowedClients 包含允许的客户端(如 "vscode","cursor")并同时包含 blockedClients 黑名单(如 "windsurf"),保持与文档中其他章节示例的格式和风格一致并放在字段说明或匹配规则之后以便阅读。
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@src/app/docs/reference/provider-fields/page.md`:
- Around line 695-709: Add a small JSON configuration example and consistent
formatting to the "交换缓存 TTL 计费" section: update the description for the
`swapCacheTtlBilling` field to include a one-line JSON snippet showing how to
enable it (e.g., {"swapCacheTtlBilling": true}) and ensure the example is
wrapped as a code block and placed immediately after the field table so it
matches other sections' structure; reference the `swapCacheTtlBilling` field and
the existing section header to find where to insert the example.
- Around line 672-693: 在“客户端限制配置”章节中补充一个 JSON 配置示例来示范 allowedClients 和
blockedClients 的用法;在 page.md 的该节(涉及字段 allowedClients、blockedClients
及匹配规则)添加一个简短的代码块示例,展示 allowedClients 包含允许的客户端(如 "vscode","cursor")并同时包含
blockedClients 黑名单(如 "windsurf"),保持与文档中其他章节示例的格式和风格一致并放在字段说明或匹配规则之后以便阅读。
ℹ️ Review info
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Cache: Disabled due to Reviews > Disable Cache setting
📒 Files selected for processing (2)
src/app/docs/reference/env-variables/page.mdsrc/app/docs/reference/provider-fields/page.md
- 统一 changelog v0.5.3-v0.5.8 标题格式为可点击链接格式 - 修复 v0.5.5 changelog 中 "Vacuum Filter has" 中英混杂问题 - 修复 langfuse/page.md 代码块缺少语言标识符 - 完善 ENABLE_ENDPOINT_CIRCUIT_BREAKER 表格描述,明确双层级影响 - 为客户端限制配置和交换缓存 TTL 计费添加 JSON 配置示例
There was a problem hiding this comment.
🧹 Nitpick comments (1)
src/app/docs/reference/env-variables/page.md (1)
547-554: 表格与 callout 内容有轻微重复,建议精简。Line 550 的表格行为描述已提到"供应商类型级别和端点级别的熔断器均生效",Line 552-554 的 callout 再次说明"此开关同时控制供应商类型级别和端点级别的熔断器",存在信息重复。
建议优化:要么在表格中简化描述,要么在 callout 中补充额外信息(如生效机制、使用建议等)。
♻️ 可选优化方案
方案 1:简化表格,让 callout 承载详细说明
| 值 | 行为 | |------|------| | `false` | 禁用端点熔断器,所有启用的端点均可使用 | -| `true` | 启用熔断器,供应商类型级别和端点级别的熔断器均生效,连续失败的端点会被临时屏蔽(默认 3 次失败后熔断 5 分钟) | +| `true` | 启用熔断器,连续失败的端点会被临时屏蔽(默认 3 次失败后熔断 5 分钟) | {% callout type="note" %} -此开关同时控制供应商类型级别和端点级别的熔断器。启用后,供应商类型和端点的熔断器都会生效。 +此开关同时控制**供应商类型级别**和**端点级别**的熔断器。启用后,系统会在两个层级同时监控失败率:供应商类型层级的熔断可快速屏蔽整类供应商,端点层级的熔断可精细隔离单个故障端点。 {% /callout %}方案 2:在 callout 中补充使用建议
{% callout type="note" %} -此开关同时控制供应商类型级别和端点级别的熔断器。启用后,供应商类型和端点的熔断器都会生效。 +此开关同时控制供应商类型级别和端点级别的熔断器。启用后,供应商类型和端点的熔断器都会生效。生产环境建议保持关闭(默认),仅在多端点部署且需要精细故障隔离时启用。 {% /callout %}🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/app/docs/reference/env-variables/page.md` around lines 547 - 554, 表格中对 `true` 的行为描述与后面的 {% callout %} 重复;请在表格的该单元(值为 `true` 的行为列)中简化为一句核心信息(例如“启用熔断器(默认:连续 3 次失败后熔断 5 分钟)”),并在 {% callout %} 块中保留/补充详细说明或使用建议(例如生效范围、建议场景或配置提示),以消除冗余并将概要与详述分别放在表格和 callout 中,修改目标为表格的 `true` 行和随后的 {% callout %} 块。
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@src/app/docs/reference/env-variables/page.md`:
- Around line 547-554: 表格中对 `true` 的行为描述与后面的 {% callout %} 重复;请在表格的该单元(值为 `true`
的行为列)中简化为一句核心信息(例如“启用熔断器(默认:连续 3 次失败后熔断 5 分钟)”),并在 {% callout %}
块中保留/补充详细说明或使用建议(例如生效范围、建议场景或配置提示),以消除冗余并将概要与详述分别放在表格和 callout 中,修改目标为表格的 `true`
行和随后的 {% callout %} 块。
ℹ️ Review info
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Cache: Disabled due to Reviews > Disable Cache setting
📒 Files selected for processing (4)
src/app/docs/changelog/page.mdsrc/app/docs/reference/env-variables/page.mdsrc/app/docs/reference/provider-fields/page.mdsrc/app/docs/system/langfuse/page.md
数据导出 API 实际使用 ?mode=full|excludeLogs|ledgerOnly 参数, 而非独立的 boolean 参数。同时补充 ledger-only 文件命名示例。
Summary
全量检查 v0.4.2 ~ v0.6.1 所有版本变更,更新文档覆盖。
Commit 1: 初始更新(v0.5.3 ~ v0.6.1)
新增文档
system/langfuse/page.md)更新内容(14 个文件)
Commit 2: 全量补充(v0.4.2 ~ v0.6.1 覆盖检查)
reference/provider-fields/page.md(+192 行)
reference/env-variables/page.md(+197 行)
修复预存问题
Test plan
bun run build验证构建无误Greptile Summary
This PR updates documentation coverage from v0.4.2 to v0.6.1, adding comprehensive documentation for new features and correcting outdated references.
Major additions:
Key corrections:
ENABLE_MULTI_PROVIDER_TYPES(no longer required)STORE_SESSION_MESSAGESbehavior description to reflect v0.5.1 changesexcludeLogsboolean tomodeenum withledgerOnlyoptionVersion coverage:
The documentation is well-structured, comprehensive, and accurately reflects the application's evolution across these versions.
Confidence Score: 5/5
Important Files Changed
Last reviewed commit: dd0aa58