# deeplake - Doramagic AI Context Pack

> 定位：安装前体验与判断资产。它帮助宿主 AI 有一个好的开始，但不代表已经安装、执行或验证目标项目。

## 充分原则

- **充分原则，不是压缩原则**：AI Context Pack 应该充分到让宿主 AI 在开工前理解项目价值、能力边界、使用入口、风险和证据来源；它可以分层组织，但不以最短摘要为目标。
- **压缩策略**：只压缩噪声和重复内容，不压缩会影响判断和开工质量的上下文。

## 给宿主 AI 的使用方式

你正在读取 Doramagic 为 deeplake 编译的 AI Context Pack。请把它当作开工前上下文：帮助用户理解适合谁、能做什么、如何开始、哪些必须安装后验证、风险在哪里。不要声称你已经安装、运行或执行了目标项目。

## Claim 消费规则

- **事实来源**：Repo Evidence + Claim/Evidence Graph；Human Wiki 只提供显著性、术语和叙事结构。
- **事实最低状态**：`supported`
- `supported`：可以作为项目事实使用，但回答中必须引用 claim_id 和证据路径。
- `weak`：只能作为低置信度线索，必须要求用户继续核实。
- `inferred`：只能用于风险提示或待确认问题，不能包装成项目事实。
- `unverified`：不得作为事实使用，应明确说证据不足。
- `contradicted`：必须展示冲突来源，不得替用户强行选择一个版本。

## 它最适合谁

- **AI 研究者或研究型 Agent 构建者**：README 明确围绕研究、实验或论文工作流展开。 证据：`README.md` Claim：`clm_0002` supported 0.86

## 它能做什么

- **命令行启动或安装流程**（需要安装后验证）：项目文档中存在可执行命令，真实使用需要在本地或宿主环境中运行这些命令。 证据：`README.md` Claim：`clm_0001` supported 0.86

## 怎么开始

- `pip install deeplake` 证据：`README.md` Claim：`clm_0003` supported 0.86

## 继续前判断卡

- **当前建议**：先做角色匹配试用
- **为什么**：这个项目更像角色库，核心风险是选错角色或把角色文案当执行能力；先用 Prompt Preview 试角色匹配，再决定是否沙盒导入。

### 30 秒判断

- **现在怎么做**：先做角色匹配试用
- **最小安全下一步**：先用 Prompt Preview 试角色匹配；满意后再隔离导入
- **先别相信**：角色质量和任务匹配不能直接相信。
- **继续会触碰**：角色选择偏差、命令执行、本地环境或项目文件

### 现在可以相信

- **适合人群线索：AI 研究者或研究型 Agent 构建者**（supported）：有 supported claim 或项目证据支撑，但仍不等于真实安装效果。 证据：`README.md` Claim：`clm_0002` supported 0.86
- **能力存在：命令行启动或安装流程**（supported）：可以相信项目包含这类能力线索；是否适合你的具体任务仍要试用或安装后验证。 证据：`README.md` Claim：`clm_0001` supported 0.86
- **存在 Quick Start / 安装命令线索**（supported）：可以相信项目文档出现过启动或安装入口；不要因此直接在主力环境运行。 证据：`README.md` Claim：`clm_0003` supported 0.86

### 现在还不能相信

- **角色质量和任务匹配不能直接相信。**（unverified）：角色库证明有很多角色，不证明每个角色都适合你的具体任务，也不证明角色能产生高质量结果。
- **不能把角色文案当成真实执行能力。**（unverified）：安装前只能判断角色描述和任务画像是否匹配，不能证明它能在宿主 AI 里完成任务。
- **真实输出质量不能在安装前相信。**（unverified）：Prompt Preview 只能展示引导方式，不能证明真实项目中的结果质量。
- **宿主 AI 版本兼容性不能在安装前相信。**（unverified）：Claude、Cursor、Codex、Gemini 等宿主加载规则和版本差异必须在真实环境验证。
- **不会污染现有宿主 AI 行为，不能直接相信。**（inferred）：Skill、plugin、AGENTS/CLAUDE/GEMINI 指令可能改变宿主 AI 的默认行为。
- **可安全回滚不能默认相信。**（unverified）：除非项目明确提供卸载和恢复说明，否则必须先在隔离环境验证。
- **真实安装后是否与用户当前宿主 AI 版本兼容？**（unverified）：兼容性只能通过实际宿主环境验证。
- **项目输出质量是否满足用户具体任务？**（unverified）：安装前预览只能展示流程和边界，不能替代真实评测。

### 继续会触碰什么

- **角色选择偏差**：用户对任务应该由哪个专家角色处理的判断。 原因：选错角色会让 AI 从错误专业视角回答，浪费时间或误导决策。
- **命令执行**：包管理器、网络下载、本地插件目录、项目配置或用户主目录。 原因：运行第一条命令就可能产生环境改动；必须先判断是否值得跑。 证据：`README.md`
- **本地环境或项目文件**：安装结果、插件缓存、项目配置或本地依赖目录。 原因：安装前无法证明写入范围和回滚方式，需要隔离验证。 证据：`README.md`
- **宿主 AI 上下文**：AI Context Pack、Prompt Preview、Skill 路由、风险规则和项目事实。 原因：导入上下文会影响宿主 AI 后续判断，必须避免把未验证项包装成事实。

### 最小安全下一步

- **先跑 Prompt Preview**：先用交互式试用验证任务画像和角色匹配，不要先导入整套角色库。（适用：任何项目都适用，尤其是输出质量未知时。）
- **只在隔离目录或测试账号试装**：避免安装命令污染主力宿主 AI、真实项目或用户主目录。（适用：存在命令执行、插件配置或本地写入线索时。）
- **安装后只验证一个最小任务**：先验证加载、兼容、输出质量和回滚，再决定是否深用。（适用：准备从试用进入真实工作流时。）

### 退出方式

- **保留安装前状态**：记录原始宿主配置和项目状态，后续才能判断是否可恢复。
- **保留原始角色选择记录**：如果输出偏题，可以回到任务画像阶段重新选择角色，而不是继续沿着错误角色推进。
- **记录安装命令和写入路径**：没有明确卸载说明时，至少要知道哪些目录或配置需要手动清理。
- **如果没有回滚路径，不进入主力环境**：不可回滚是继续前阻断项，不应靠信任或运气继续。

## 哪些只能预览

- 解释项目适合谁和能做什么
- 基于项目文档演示典型对话流程
- 帮助用户判断是否值得安装或继续研究

## 哪些必须安装后验证

- 真实安装 Skill、插件或 CLI
- 执行脚本、修改本地文件或访问外部服务
- 验证真实输出质量、性能和兼容性

## 边界与风险判断卡

- **把安装前预览误认为真实运行**：用户可能高估项目已经完成的配置、权限和兼容性验证。 处理方式：明确区分 prompt_preview_can_do 与 runtime_required。 Claim：`clm_0004` inferred 0.45
- **命令执行会修改本地环境**：安装命令可能写入用户主目录、宿主插件目录或项目配置。 处理方式：先在隔离环境或测试账号中运行。 证据：`README.md` Claim：`clm_0005` supported 0.86
- **待确认**：真实安装后是否与用户当前宿主 AI 版本兼容？。原因：兼容性只能通过实际宿主环境验证。
- **待确认**：项目输出质量是否满足用户具体任务？。原因：安装前预览只能展示流程和边界，不能替代真实评测。
- **待确认**：安装命令是否需要网络、权限或全局写入？。原因：这影响企业环境和个人环境的安装风险。

## 开工前工作上下文

### 加载顺序

- 先读取 how_to_use.host_ai_instruction，建立安装前判断资产的边界。
- 读取 claim_graph_summary，确认事实来自 Claim/Evidence Graph，而不是 Human Wiki 叙事。
- 再读取 intended_users、capabilities 和 quick_start_candidates，判断用户是否匹配。
- 需要执行具体任务时，优先查 role_skill_index，再查 evidence_index。
- 遇到真实安装、文件修改、网络访问、性能或兼容性问题时，转入 risk_card 和 boundaries.runtime_required。

### 任务路由

- **命令行启动或安装流程**：先说明这是安装后验证能力，再给出安装前检查清单。 边界：必须真实安装或运行后验证。 证据：`README.md` Claim：`clm_0001` supported 0.86

### 上下文规模

- 文件总数：870
- 重要文件覆盖：40/870
- 证据索引条目：78
- 角色 / Skill 条目：48

### 证据不足时的处理

- **missing_evidence**：说明证据不足，要求用户提供目标文件、README 段落或安装后验证记录；不要补全事实。
- **out_of_scope_request**：说明该任务超出当前 AI Context Pack 证据范围，并建议用户先查看 Human Manual 或真实安装后验证。
- **runtime_request**：给出安装前检查清单和命令来源，但不要替用户执行命令或声称已执行。
- **source_conflict**：同时展示冲突来源，标记为待核实，不要强行选择一个版本。

## Prompt Recipes

### 适配判断

- 目标：判断这个项目是否适合用户当前任务。
- 预期输出：适配结论、关键理由、证据引用、安装前可预览内容、必须安装后验证内容、下一步建议。

```text
请基于 deeplake 的 AI Context Pack，先问我 3 个必要问题，然后判断它是否适合我的任务。回答必须包含：适合谁、能做什么、不能做什么、是否值得安装、证据来自哪里。所有项目事实必须引用 evidence_refs、source_paths 或 claim_id。
```

### 安装前体验

- 目标：让用户在安装前感受核心工作流，同时避免把预览包装成真实能力或营销承诺。
- 预期输出：一段带边界标签的体验剧本、安装后验证清单和谨慎建议；不含真实运行承诺或强营销表述。

```text
请把 deeplake 当作安装前体验资产，而不是已安装工具或真实运行环境。

请严格输出四段：
1. 先问我 3 个必要问题。
2. 给出一段“体验剧本”：用 [安装前可预览]、[必须安装后验证]、[证据不足] 三种标签展示它可能如何引导工作流。
3. 给出安装后验证清单：列出哪些能力只有真实安装、真实宿主加载、真实项目运行后才能确认。
4. 给出谨慎建议：只能说“值得继续研究/试装”“先补充信息后再判断”或“不建议继续”，不得替项目背书。

硬性边界：
- 不要声称已经安装、运行、执行测试、修改文件或产生真实结果。
- 不要写“自动适配”“确保通过”“完美适配”“强烈建议安装”等承诺性表达。
- 如果描述安装后的工作方式，必须使用“如果安装成功且宿主正确加载 Skill，它可能会……”这种条件句。
- 体验剧本只能写成“示例台词/假设流程”：使用“可能会询问/可能会建议/可能会展示”，不要写“已写入、已生成、已通过、正在运行、正在生成”。
- Prompt Preview 不负责给安装命令；如用户准备试装，只能提示先阅读 Quick Start 和 Risk Card，并在隔离环境验证。
- 所有项目事实必须来自 supported claim、evidence_refs 或 source_paths；inferred/unverified 只能作风险或待确认项。

```

### 角色 / Skill 选择

- 目标：从项目里的角色或 Skill 中挑选最匹配的资产。
- 预期输出：候选角色或 Skill 列表，每项包含适用场景、证据路径、风险边界和是否需要安装后验证。

```text
请读取 role_skill_index，根据我的目标任务推荐 3-5 个最相关的角色或 Skill。每个推荐都要说明适用场景、可能输出、风险边界和 evidence_refs。
```

### 风险预检

- 目标：安装或引入前识别环境、权限、规则冲突和质量风险。
- 预期输出：环境、权限、依赖、许可、宿主冲突、质量风险和未知项的检查清单。

```text
请基于 risk_card、boundaries 和 quick_start_candidates，给我一份安装前风险预检清单。不要替我执行命令，只说明我应该检查什么、为什么检查、失败会有什么影响。
```

### 宿主 AI 开工指令

- 目标：把项目上下文转成一次对话开始前的宿主 AI 指令。
- 预期输出：一段边界明确、证据引用明确、适合复制给宿主 AI 的开工前指令。

```text
请基于 deeplake 的 AI Context Pack，生成一段我可以粘贴给宿主 AI 的开工前指令。这段指令必须遵守 not_runtime=true，不能声称项目已经安装、运行或产生真实结果。
```

## 角色 / Skill 索引

- 共索引 48 个角色 / Skill / 项目文档条目。

- **Deep Lake Documentation**（project_doc）：This directory contains the documentation for the Deep Lake project. The documentation is written in markdown and built using . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/README.md`
- **Readme**（project_doc）：Developer Documentation dev-docs.md Supported SQL Queries syntax-support.md Known Limitations & Missing Features known-limitations.md Basic Usage basic-usage.md 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/docs/README.md`
- **What is Deep Lake?**（project_doc）：Docs &bull; Get Started &bull; API Reference &bull; LangChain & VectorDBs Course &bull; Blog &bull; Whitepaper &bull; Slack &bull; Twitter 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`README.md`
- **pg deeplake**（project_doc）：PostgreSQL extension for vector similarity search, full-text search, and hybrid search using DeepLake. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`postgres/README.md`
- **OPENJPEG Library and Applications**（project_doc）：OpenJPEG is an open-source JPEG 2000 codec written in C language. It has been developed in order to promote the use of JPEG 2000 http://www.jpeg.org/jpeg2000 , a still-image compression standard from the Joint Photographic Experts Group JPEG http://www.jpeg.org . Since April 2015, it is officially recognized by ISO/IEC and ITU-T as a JPEG 2000 Reference Software http://www.itu.int/rec/T-REC-T.804-201504-I!Amd2 . 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/openjpeg/README.md`
- **Usage**（project_doc）：C++ SQL Parser ========================= ! Build Status https://app.travis-ci.com/hyrise/sql-parser.svg?branch=master https://app.travis-ci.com/github/hyrise/sql-parser 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/README.md`
- **Benchmark**（project_doc）：This directory contains the scripts to execute benchmarks of the parser. We use Google Benchmark https://github.com/google/benchmark to define and run benchmarks. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/benchmark/README.md`
- **VCPKG custom ports**（project_doc）：This document is added to describe the issues with the official ports and why these are needed. It might also be possible that an offical port was not available at the moment we started using it. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/vcpkg-ports/README.md`
- **Contributing Standards**（project_doc）：We use the black https://pypi.org/project/black/ python linter. You can have your code auto-formatted by running pip install black , then black . inside the directory you want to format. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`CONTRIBUTING.md`
- **Changelog**（project_doc）：v2.4.0 https://github.com/uclouvain/openjpeg/releases/v2.4.0 2020-12-28 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/openjpeg/CHANGELOG.md`
- **Example Code**（project_doc）：Using the Library ======================= 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/docs/basic-usage.md`
- **Basic Requirements**（project_doc）：Developer Documentation ======================= 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/docs/dev-docs.md`
- **Completely Missing Statement Types**（project_doc）：Known Limitations & Missing Features ==================================== 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/docs/known-limitations.md`
- **Select Statements**（project_doc）：Supported SQL Queries ===================== 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`cpp/3rd_party/sql-parser/docs/syntax-support.md`
- **Best Practices for Deep Lake Usage**（project_doc）：Learn how to tune your Deep Lake usage for optimal performance and cost 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/advanced/best-practices.md`
- **Dataset Copying and Synchronization**（project_doc）：Learn how to copy and synchronize Deep Lake datasets across different storage locations and cloud providers 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/advanced/sync.md`
- **TQL Syntax**（project_doc）：Up to 10x More Efficient Data Retrieval with TQL, Deep Lake's Custom Query Language Query Datasets, Join Cross-Cloud Datasets, Manipulate Multi-Modal Datasets. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/advanced/tql.md`
- **Dataset Visualization**（project_doc）：Learn how to visualize Deep Lake datasets using the Deep Lake web interface. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/advanced/visualize.md`
- **Column Classes**（project_doc）：Access Deep Lake Documentation For Complete Setup, API Reference, Guides On Efficient Multi-Modal AI Search, Dataset Management, Cost-Efficient Training, And Retrieval-Augmented Generation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/column.md`
- **Dataset Classes**（project_doc）：Access Deep Lake Documentation For Complete Setup, API Reference, Guides On Efficient Multi-Modal AI Search, Dataset Management, Cost-Efficient Training, And Retrieval-Augmented Generation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/dataset.md`
- **API Reference**（project_doc）：Complete API Reference For Deep Lake, Including Dataset Creation, Multi-Modal Query Engine, Vector Search, And ML Framework Integrations. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/index.md`
- **Metadata**（project_doc）：Access Deep Lake Documentation For Metadata. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/metadata.md`
- **Query**（project_doc）：Deep Lake Query Engine Documentation and Examples. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/query.md`
- **Schemas**（project_doc）：Deeplake builtin schemas to create dataset. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/schemas.md`
- **Types**（project_doc）：Definition and Examples of types used in Deeplake Datasets and Columns 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/types.md`
- **Version Control**（project_doc）：Access Deep Lake Documentation For Version Control. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/api/version_control.md`
- **User Authentication**（project_doc）：Set Up Authentication Protocols For Secure Access To Deep Lake Datasets, Ensuring Data Integrity And Privacy In Collaborative Machine Learning Environments. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/authentication.md`
- **Quickstart Guide**（project_doc）：Access Deep Lake Documentation For Complete Setup, API Reference, Guides On Efficient Multi-Modal AI Search, Dataset Management, Cost-Efficient Training, And Retrieval-Augmented Generation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/quickstart.md`
- **Storage and Credentials**（project_doc）：Comprehensive Guide To Configuring Storage And Credentials In Deep Lake 4.0, Including Cloud, Local, And Hybrid Storage Options For Efficient AI Model Data Handling. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/index.md`
- **Enabling CORS in S3**（project_doc）：Follow This Guide To Configure CORS For AWS With Deep Lake 4.0, Ensuring Secure Cross-Origin Resource Sharing And Compliance For Machine Learning Training Data. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/aws/cors.md`
- **Provisioning Role-Based Access**（project_doc）：Provision AWS Resources To Store And Access Deep Lake Datasets, Supporting Scalable Machine Learning Training With Optimized Cloud Storage Setup. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/aws/provisioning.md`
- **Enabling CORS in Azure**（project_doc）：Set Up Azure CORS With Deep Lake To Enable Secure Cross-Origin Requests, Providing Efficient Access To Training Data For Machine Learning Projects. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/azure/cors.md`
- **Provisioning Federated Credentials**（project_doc）：Step-By-Step Guide For Provisioning Azure Storage With Deep Lake, Ensuring Secure And Scalable Data Access To Support Machine Learning Workflows. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/azure/provisioning.md`
- **Azure Workload Identities**（project_doc）：Manage Azure Workload Identities For Role-Based Access To Deep Lake, Providing Secure Data Access And Ensuring Safety Across Machine Learning Workflows. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/azure/workload-identities.md`
- **Enabling CORS in GCP**（project_doc）：Set Up CORS For Google Cloud Platform With Deep Lake To Securely Handle Cross-Origin Requests, Ensuring Compliance And Easy Access To Machine Learning Data. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/gcp/cors.md`
- **Provisioning Federated Credentials**（project_doc）：Provision Google Cloud Resources For Use With Deep Lake, Enabling Secure, Scalable Storage Solutions For Managing Large Multi-Modal AI Datasets. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/gcp/provisioning.md`
- **Setting up Deep Lake in Your Cloud**（project_doc）：Set Up Managed Credentials To Enable Secure Access To Cloud Storage Resources In Deep Lake, Essential For Managing Large-Scale AI And Machine Learning Data. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/index.md`
- **Storage Options**（project_doc）：Explore Various Storage Options In Deep Lake, To Store and Search Across Data Efficiently 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/getting-started/storage-and-creds/storage-options.md`
- **Labelbox Integration**（project_doc）：Labelbox integration for annotations. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/annotations/labelbox.md`
- **Async Data Loader**（project_doc）：Use The Async Data Loader To Parallelize Data Ingestion, Enhancing Speed And Efficiency When Loading Multi-Modal Data For ML Model Training or Fast AI Search. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/deep-learning/async-data-loader.md`
- **Training Object Detection Models with Deep Lake and MMDetection**（project_doc）：Deeplake Dataset training for Object detection using MMDetection framework. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/deep-learning/mmdet.md`
- **Semantic Segmentation with Deep Lake and MMSegmentation**（project_doc）：Deeplake Dataset training for semantic segmentation using MMSegmentation framework. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/deep-learning/mmseg.md`
- **Advancing Search Capabilities: From Lexical to Multi-Modal with Deep Lake**（project_doc）：Up to 10x More Efficient Data Retrieval with TQL, learn how to build RAG applications with deeplake. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/rag.md`
- **Migrating to Deep Lake v4**（project_doc）：Step-By-Step Guide For Migrating Deep Lake 3.xx Datasets To 4.xx, Enabling Advanced Vector Search, RAG Applications, And Efficient ML Training. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/v3-conversion.md`
- **Using Deep Lake as a Vector Store in LangChain**（project_doc）：Up to 10x More Efficient Data Retrieval with TQL, learn how to build RAG applications with deeplake. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/guide/vectorstore.md`
- **🌊 Deep Lake: Multi-Modal AI Database**（project_doc）：Access Deep Lake Documentation For Complete Setup, API Reference, Guides On Efficient Multi-Modal AI Search, Dataset Management, Cost-Efficient Training, And Retrieval-Augmented Generation. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/index.md`
- **Academic Paper**（project_doc）：Read the Deep Lake academic paper published on arXiv covering lakehouse architecture for deep learning, tensor storage, streaming for PyTorch/TensorFlow/JAX, and MLOps integration for NLP, computer vision, and audio processing. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/resources/academic-paper.md`
- **White Paper**（project_doc）：Download the Deep Lake white paper covering database architecture, storage strategies, query optimization for ML workloads, ML framework integration, and performance benchmarks for AI applications. 激活提示：当用户需要理解项目结构、安装方式或边界时参考。 证据：`docs/docs/resources/whitepaper.md`

## 证据索引

- 共索引 78 条证据。

- **Deep Lake Documentation**（documentation）：This directory contains the documentation for the Deep Lake project. The documentation is written in markdown and built using . 证据：`docs/README.md`
- **Readme**（documentation）：Developer Documentation dev-docs.md Supported SQL Queries syntax-support.md Known Limitations & Missing Features known-limitations.md Basic Usage basic-usage.md 证据：`cpp/3rd_party/sql-parser/docs/README.md`
- **What is Deep Lake?**（documentation）：Docs &bull; Get Started &bull; API Reference &bull; LangChain & VectorDBs Course &bull; Blog &bull; Whitepaper &bull; Slack &bull; Twitter 证据：`README.md`
- **pg deeplake**（documentation）：PostgreSQL extension for vector similarity search, full-text search, and hybrid search using DeepLake. 证据：`postgres/README.md`
- **OPENJPEG Library and Applications**（documentation）：OpenJPEG is an open-source JPEG 2000 codec written in C language. It has been developed in order to promote the use of JPEG 2000 http://www.jpeg.org/jpeg2000 , a still-image compression standard from the Joint Photographic Experts Group JPEG http://www.jpeg.org . Since April 2015, it is officially recognized by ISO/IEC and ITU-T as a JPEG 2000 Reference Software http://www.itu.int/rec/T-REC-T.804-201504-I!Amd2 . 证据：`cpp/3rd_party/openjpeg/README.md`
- **Usage**（documentation）：C++ SQL Parser ========================= ! Build Status https://app.travis-ci.com/hyrise/sql-parser.svg?branch=master https://app.travis-ci.com/github/hyrise/sql-parser 证据：`cpp/3rd_party/sql-parser/README.md`
- **Benchmark**（documentation）：This directory contains the scripts to execute benchmarks of the parser. We use Google Benchmark https://github.com/google/benchmark to define and run benchmarks. 证据：`cpp/3rd_party/sql-parser/benchmark/README.md`
- **VCPKG custom ports**（documentation）：This document is added to describe the issues with the official ports and why these are needed. It might also be possible that an offical port was not available at the moment we started using it. 证据：`cpp/vcpkg-ports/README.md`
- **Contributing Standards**（documentation）：We use the black https://pypi.org/project/black/ python linter. You can have your code auto-formatted by running pip install black , then black . inside the directory you want to format. 证据：`CONTRIBUTING.md`
- **License**（source_file）：Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ 证据：`LICENSE`
- **License**（source_file）：/ The copyright in this software is being made available under the 2-clauses BSD License, included below. This software may be subject to other third party and contributor rights, including patent rights, and no such rights are granted under this license. Copyright c 2002-2014, Universite catholique de Louvain UCL , Belgium Copyright c 2002-2014, Professor Benoit Macq Copyright c 2003-2014, Antonin Descampe Copyright c 2003-2009, Francois-Olivier Devaux Copyright c 2005, Herve Drolon, FreeImage Team Copyright c 2002-2003, Yannick Verschueren Copyright c 2001-2003, David Janssens Copyright c 2011-2012, Centre National d'Etudes Spatiales CNES , France Copyright c 2012, CS Systemes d'Informati… 证据：`cpp/3rd_party/openjpeg/LICENSE`
- **License**（source_file）：Copyright c 2012-2017 Hasso-Plattner-Institut 证据：`cpp/3rd_party/sql-parser/LICENSE`
- **Changelog**（documentation）：v2.4.0 https://github.com/uclouvain/openjpeg/releases/v2.4.0 2020-12-28 证据：`cpp/3rd_party/openjpeg/CHANGELOG.md`
- **Example Code**（documentation）：Using the Library ======================= 证据：`cpp/3rd_party/sql-parser/docs/basic-usage.md`
- **Basic Requirements**（documentation）：Developer Documentation ======================= 证据：`cpp/3rd_party/sql-parser/docs/dev-docs.md`
- **Completely Missing Statement Types**（documentation）：Known Limitations & Missing Features ==================================== 证据：`cpp/3rd_party/sql-parser/docs/known-limitations.md`
- **Select Statements**（documentation）：Supported SQL Queries ===================== 证据：`cpp/3rd_party/sql-parser/docs/syntax-support.md`
- **Best Practices for Deep Lake Usage**（documentation）：Deep Lake supports wide range of data types and powerful tools to ingest, load and query your data. This page provides tips for optimizing your usage of Deep Lake for best experience and performance. 证据：`docs/docs/advanced/best-practices.md`
- **Dataset Copying and Synchronization**（documentation）：<!-- test-context python import deeplake import numpy as np def get builtin signature func : name = func. name doc = func. doc or '' sig = doc.split '\n' 0 .strip return f"{name}{sig}" 证据：`docs/docs/advanced/sync.md`
- **TQL Syntax**（documentation）：Deep Lake offers a performant SQL-based query engine called "TQL" Tensor Query Language optimized for machine learning and AI workloads. TQL combines familiar SQL syntax with powerful tensor operations, enabling efficient querying of embeddings, images, and other multi-modal data. 证据：`docs/docs/advanced/tql.md`
- **Dataset Visualization**（documentation）：Deep Lake has a web interface for visualizing, versioning, and querying machine learning datasets. It utilizes the Deep Lake format under-the-hood, and it can be connected to datasets stored in all Deep Lake storage locations ../../getting-started/storage-and-creds/ . 证据：`docs/docs/advanced/visualize.md`
- **Column Classes**（documentation）：Deep Lake provides two column classes for different access levels: 证据：`docs/docs/api/column.md`
- **Dataset Classes**（documentation）：Deep Lake provides three dataset classes with different access levels: 证据：`docs/docs/api/dataset.md`
- **API Reference**（documentation）：This reference documents the Python API of Deep Lake. 证据：`docs/docs/api/index.md`
- **Metadata**（documentation）：Metadata provides key-value storage for datasets and columns. 证据：`docs/docs/api/metadata.md`
- **Query**（documentation）：Deep Lake provides powerful query capabilities through its Tensor Query Language TQL , with special focus on vector similarity search, text search, and operations on multidimensional arrays. 证据：`docs/docs/api/query.md`
- **Schemas**（documentation）：Deep Lake provides pre-built schema templates for common data structures. 证据：`docs/docs/api/schemas.md`
- **Types**（documentation）：Deep Lake provides a comprehensive type system designed for efficient data storage and retrieval. The type system includes basic numeric types as well as specialized types optimized for common data formats like images, embeddings, and text. 证据：`docs/docs/api/types.md`
- **Version Control**（documentation）：<!-- test-context python import numpy as np import deeplake from deeplake import types 证据：`docs/docs/api/version_control.md`
- **User Authentication**（documentation）：How to Register and Authenticate in Deep Lake 证据：`docs/docs/getting-started/authentication.md`
- **Quickstart Guide**（documentation）：Get started with Deep Lake by following these examples. 证据：`docs/docs/getting-started/quickstart.md`
- **Storage and Credentials**（documentation）：How to access datasets in other clouds and manage their credentials. 证据：`docs/docs/getting-started/storage-and-creds/index.md`
- **Enabling CORS in S3**（documentation）：In order to visualize Deep Lake datasets stored in your own S3 buckets in the Deep Lake app https://app.activeloop.ai/ , please enable Cross-Origin Resource Sharing CORS https://en.wikipedia.org/wiki/Cross-origin\ resource\ sharing in the buckets containing the Deep Lake dataset and any linked data, by inserting the snippet below in the CORS section of the Permissions tab for the bucket: 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/aws/cors.md`
- **Provisioning Role-Based Access**（documentation）：Setting up Role-Based Access for AWS S3 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/aws/provisioning.md`
- **Enabling CORS in Azure**（documentation）：Cross-Origin Resource Sharing CORS is typically enabled by default in Azure. If that's not the case in your Azure account, please enable CORS https://en.wikipedia.org/wiki/Cross-origin\ resource\ sharing in order to use the Deep Lake app https://app.activeloop.ai/ to visualize Deep Lake datasets stored in your own Azure storage. CORS https://en.wikipedia.org/wiki/Cross-origin\ resource\ sharing should be enabled in the storage account containing the Deep Lake dataset and any linked data. 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/azure/cors.md`
- **Provisioning Federated Credentials**（documentation）：Setting up Federated Credentials in Microsoft Azure 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/azure/provisioning.md`
- **Azure Workload Identities**（documentation）：How to authenticate using workload identities instead of user credentials. 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/azure/workload-identities.md`
- **Enabling CORS in GCP**（documentation）：In order to visualize Deep Lake datasets stored in your own GCP buckets in the Deep Lake app https://app.activeloop.ai/ , please enable Cross-Origin Resource Sharing CORS https://en.wikipedia.org/wiki/Cross-origin\ resource\ sharing in the buckets containing the Deep Lake dataset and any linked data, by inserting the snippet below in the CORS section of the Permissions tab for the bucket: 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/gcp/cors.md`
- **Provisioning Federated Credentials**（documentation）：Setting up Federated Credentials in Google Cloud Platform 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/gcp/provisioning.md`
- **Setting up Deep Lake in Your Cloud**（documentation）：Connecting Data From Your Cloud Using Deep Lake Managed Credentials 证据：`docs/docs/getting-started/storage-and-creds/managed-credentials/index.md`
- **Storage Options**（documentation）：How to authenticate using Activeloop storage, AWS S3, and Google Cloud Storage. 证据：`docs/docs/getting-started/storage-and-creds/storage-options.md`
- **Labelbox Integration**（documentation）：Labelbox Integration This document describes how to create Deep Lake datasets from Labelbox https://labelbox.com/ annotations. The API also allows you to update the dataset with new annotations. 证据：`docs/docs/guide/annotations/labelbox.md`
- **Async Data Loader**（documentation）：This document describes the implementation of a custom DataLoader for handling data retrieval using deeplake.Dataset with PyTorch . The DataLoader supports both sequential and asynchronous data fetching, with the asynchronous approach being optimized for performance and speed. 证据：`docs/docs/guide/deep-learning/async-data-loader.md`
- **Training Object Detection Models with Deep Lake and MMDetection**（documentation）：Training Object Detection Models with Deep Lake and MMDetection 证据：`docs/docs/guide/deep-learning/mmdet.md`
- **Semantic Segmentation with Deep Lake and MMSegmentation**（documentation）：Semantic Segmentation with Deep Lake and MMSegmentation 证据：`docs/docs/guide/deep-learning/mmseg.md`
- **Advancing Search Capabilities: From Lexical to Multi-Modal with Deep Lake**（documentation）：Advancing Search Capabilities: From Lexical to Multi-Modal with Deep Lake 证据：`docs/docs/guide/rag.md`
- **Migrating to Deep Lake v4**（documentation）：Deep Lake 4.0 introduces major improvements for ML and AI applications: 证据：`docs/docs/guide/v3-conversion.md`
- **Using Deep Lake as a Vector Store in LangChain**（documentation）：Using Deep Lake as a Vector Store in LangChain 证据：`docs/docs/guide/vectorstore.md`
- **🌊 Deep Lake: Multi-Modal AI Database**（documentation）：🌊 Deep Lake: Multi-Modal AI Database 证据：`docs/docs/index.md`
- **Academic Paper**（documentation）：Deep Lake: a Lakehouse for Deep Learning 证据：`docs/docs/resources/academic-paper.md`
- **White Paper**（documentation）：The Deep Lake whitepaper provides a comprehensive overview of the Deep Lake database architecture, its core features, and use cases for AI and machine learning workflows. 证据：`docs/docs/resources/whitepaper.md`
- **Vcpkg Configuration**（structured_config）：{ "default-registry": { "kind": "git", "baseline": "6f29f12e82a8293156836ad81cc9bf5af41fe836", "repository": "https://github.com/microsoft/vcpkg" }, "registries": { "kind": "artifact", "location": "https://github.com/microsoft/vcpkg-ce-catalog/archive/refs/heads/main.zip", "name": "microsoft" } , "overlay-ports": "./vcpkg-ports" , "overlay-triplets": "./vcpkg-triplets" } 证据：`cpp/vcpkg-configuration.json`
- **Vcpkg**（structured_config）：{ "name": "deeplake", "version": "0.0.1", "dependencies": "abseil", "cpr", "gtest", "curl", "eigen3", "fmt", "nlohmann-json", "nlohmann-fifo-map", "boost-algorithm", "boost-container", "boost-iterator", "boost-histogram", "boost-mpl", "boost-json", "boost-format", "boost-tti", "openssl", { "name": "opentelemetry-cpp", "platform": "!emscripten", "features": "otlp-http" }, "indicators", "rapidjson", "simdjson", "roaring", "zlib", "minimp3" } 证据：`cpp/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "abseil", "version": "20240116.1", "port-version": 1, "description": "Abseil is an open-source collection of C++ library code designed to augment the C++ standard library. The Abseil library code is collected from Google's own C++ code base, has been extensively tested and used in production, and is the same code we depend on in our daily coding lives.", "In some cases, Abseil provides pieces missing from the C++ standard; in others, Abseil provides alternatives to the standard for special needs we've found through usage in the Google code base. We denote those cases clearly within the library code we provide you.", "Abseil is not meant to be a competitor to the standard library;… 证据：`cpp/vcpkg-ports/abseil/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "crc32c", "version": "1.1.2", "port-version": 2, "description": "CRC32C implementation with support for CPU-specific acceleration instructions.", "homepage": "https://github.com/google/crc32c", "license": "BSD-3-Clause", "dependencies": { "name": "vcpkg-cmake", "host": true }, { "name": "vcpkg-cmake-config", "host": true } } 证据：`cpp/vcpkg-ports/crc32c/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "curl", "version": "8.11.1", "description": "A library for transferring data with URLs", "homepage": "https://curl.se/", "license": "curl AND ISC AND BSD-3-Clause", "dependencies": { "name": "vcpkg-cmake", "host": true }, { "name": "vcpkg-cmake-config", "host": true }, "zlib" , "default-features": "non-http", "ssl" , "features": { "brotli": { "description": "brotli support brotli ", "dependencies": "brotli" }, "c-ares": { "description": "c-ares support", "dependencies": "c-ares" }, "gnutls": { "description": "SSL support gnutls ", "dependencies": { "name": "libgnutls", "platform": "!windows mingw" }, { "name": "shiftmedia-libgnutls", "platform": "windows & !mingw" } }, "gsasl": {… 证据：`cpp/vcpkg-ports/curl/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "gflags", "version": "2.2.2", "port-version": 9, "description": "A C++ library that implements commandline flags processing", "homepage": "https://github.com/gflags/gflags", "license": "BSD-3-Clause", "supports": "!uwp", "dependencies": { "name": "vcpkg-cmake", "host": true }, { "name": "vcpkg-cmake-config", "host": true } } 证据：`cpp/vcpkg-ports/gflags/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "google-cloud-cpp", "version": "2.33.0", "description": "C++ Client Libraries for Google Cloud Platform APIs.", "homepage": "https://github.com/googleapis/google-cloud-cpp", "license": "Apache-2.0", "supports": "!uwp", "dependencies": "abseil", { "name": "openssl", "platform": "!windows" }, { "name": "vcpkg-cmake", "host": true }, { "name": "vcpkg-cmake-config", "host": true } , "default-features": "storage" , "features": { "storage": { "description": "The Google Cloud Storage C++ client library", "dependencies": "crc32c", "openssl", { "name": "curl", "features": "ssl" }, "nlohmann-json" } } } 证据：`cpp/vcpkg-ports/google-cloud-cpp/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "libevent", "version": "2.1.12+20230128", "port-version": 1, "description": "An event notification library", "homepage": "https://github.com/libevent/libevent", "license": "BSD-3-Clause", "dependencies": { "name": "vcpkg-cmake", "host": true }, { "name": "vcpkg-cmake-config", "host": true } , "default-features": "thread" , "features": { "openssl": { "description": "Support for openssl", "dependencies": { "name": "libevent", "features": "thread" }, "openssl" }, "thread": { "description": "Support for thread" } } } 证据：`cpp/vcpkg-ports/libevent/vcpkg.json`
- **Vcpkg**（structured_config）：{ "name": "nlohmann-fifo-map", "version-string": "2018.05.07", "port-version": 3, "description": "a FIFO-ordered associative container for C++", "homepage": "https://github.com/nlohmann/fifo map", "dependencies": { "name": "vcpkg-cmake", "host": true }, { "name": "vcpkg-cmake-config", "host": true } } 证据：`cpp/vcpkg-ports/nlohmann-fifo-map/vcpkg.json`
- 其余 18 条证据见 `AI_CONTEXT_PACK.json` 或 `EVIDENCE_INDEX.json`。

## 宿主 AI 必须遵守的规则

- **把本资产当作开工前上下文，而不是运行环境。**：AI Context Pack 只包含证据化项目理解，不包含目标项目的可执行状态。 证据：`docs/README.md`, `cpp/3rd_party/sql-parser/docs/README.md`, `README.md`
- **回答用户时区分可预览内容与必须安装后才能验证的内容。**：安装前体验的消费者价值来自降低误装和误判，而不是伪装成真实运行。 证据：`docs/README.md`, `cpp/3rd_party/sql-parser/docs/README.md`, `README.md`

## 用户开工前应该回答的问题

- 你准备在哪个宿主 AI 或本地环境中使用它？
- 你只是想先体验工作流，还是准备真实安装？
- 你最在意的是安装成本、输出质量、还是和现有规则的冲突？

## 验收标准

- 所有能力声明都能回指到 evidence_refs 中的文件路径。
- AI_CONTEXT_PACK.md 没有把预览包装成真实运行。
- 用户能在 3 分钟内看懂适合谁、能做什么、如何开始和风险边界。

---

## Doramagic Context Augmentation

下面内容用于强化 Repomix/AI Context Pack 主体。Human Manual 只提供阅读骨架；踩坑日志会被转成宿主 AI 必须遵守的工作约束。

## Human Manual 骨架

使用规则：这里只是项目阅读路线和显著性信号，不是事实权威。具体事实仍必须回到 repo evidence / Claim Graph。

宿主 AI 硬性规则：
- 不得把页标题、章节顺序、摘要或 importance 当作项目事实证据。
- 解释 Human Manual 骨架时，必须明确说它只是阅读路线/显著性信号。
- 能力、安装、兼容性、运行状态和风险判断必须引用 repo evidence、source path 或 Claim Graph。

- **Deep Lake 概述与版本演进**：importance `high`
  - source_paths: README.md, DEEPLAKE_API_VERSION
- **系统架构总览**：importance `high`
  - source_paths: cpp/CMakeLists.txt, cpp/deeplake_api/deeplake_api.hpp, cpp/deeplake_core/deeplake_core/chunk_strategy.hpp, cpp/nd/nd.hpp, cpp/async/async.hpp
- **核心数据类型与数据文件格式**：importance `high`
  - source_paths: cpp/deeplake_core/image_type.hpp, cpp/deeplake_core/video_type.hpp, cpp/deeplake_core/audio_type.hpp, cpp/deeplake_core/text_type.hpp, cpp/deeplake_core/mesh_type.hpp
- **视图系统与 Heimdall 延迟求值**：importance `medium`
  - source_paths: cpp/heimdall/dataset.hpp, cpp/heimdall/column.hpp, cpp/heimdall_common/chained_column_view.hpp, cpp/heimdall_common/filtered_column.hpp, cpp/heimdall_common/merged_column.hpp
- **TQL 张量查询语言与查询引擎**：importance `high`
  - source_paths: cpp/tql/tql.hpp, cpp/tql/executor.hpp, cpp/tql/functions_registry.hpp, cpp/query_core/expr.hpp, cpp/query_core/order_statement.hpp
- **PostgreSQL 集成与扩展**：importance `medium`
  - source_paths: cpp/deeplake_pg/table_am.hpp, cpp/deeplake_pg/table_am.cpp, cpp/deeplake_pg/duckdb_deeplake_convert.cpp, cpp/deeplake_pg/pg_to_duckdb_translator.cpp, cpp/deeplake_pg/hybrid_query_merge.hpp
- **Python API 与第三方集成生态**：importance `high`
  - source_paths: python/deeplake/__init__.py, python/deeplake/core.py, python/deeplake/storage.py, python/deeplake/tql.py, python/deeplake/types.py
- **性能优化、依赖管理与已知问题**：importance `high`
  - source_paths: cpp/vcpkg.json, cpp/vcpkg-configuration.json, cpp/3rd_party/CMakeLists.txt, cpp/format/format.hpp, cpp/nd/impl/dynamic_array.hpp

## Repo Inspection Evidence / 源码检查证据

- repo_clone_verified: true
- repo_inspection_verified: true
- repo_commit: `f432041fabbf4a1fc1d342aeb550e0a7de41b0da`
- inspected_files: `README.md`, `docs/README.md`, `docs/deeplake.tsconfig.json`, `docs/docker-compose.yaml`, `docs/docs/advanced/best-practices.md`, `docs/docs/advanced/sync.md`, `docs/docs/advanced/tql.md`, `docs/docs/advanced/visualize.md`, `docs/docs/api/column.md`, `docs/docs/api/dataset.md`, `docs/docs/api/index.md`, `docs/docs/api/metadata.md`, `docs/docs/api/query.md`, `docs/docs/api/schemas.md`, `docs/docs/api/types.md`, `docs/docs/api/version_control.md`, `docs/docs/getting-started/authentication.md`, `docs/docs/getting-started/quickstart.md`, `docs/docs/getting-started/storage-and-creds/index.md`, `docs/docs/getting-started/storage-and-creds/managed-credentials/aws/cors.md`

宿主 AI 硬性规则：
- 没有 repo_clone_verified=true 时，不得声称已经读过源码。
- 没有 repo_inspection_verified=true 时，不得把 README/docs/package 文件判断写成事实。
- 没有 quick_start_verified=true 时，不得声称 Quick Start 已跑通。

## Doramagic Pitfall Constraints / 踩坑约束

这些规则来自 Doramagic 发现、验证或编译过程中的项目专属坑点。宿主 AI 必须把它们当作工作约束，而不是普通说明文字。

### Constraint 1: 来源证据：deeplake 3.9.52 incompatible with NumPy 2.x (NEP 50): TypeError in get_incompatible_dtype

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：deeplake 3.9.52 incompatible with NumPy 2.x (NEP 50): TypeError in get_incompatible_dtype
- Why it matters: 可能影响升级、迁移或版本选择。
- Evidence: community_evidence:github | https://github.com/activeloopai/deeplake/issues/3144 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 2: 来源证据：[BUG] deeplake v4.5.6 produces a segmentation fault with WHERE (NOT ((f9['e1'] IS NOT NULL)))

- Trigger: GitHub 社区证据显示该项目存在一个安装相关的待验证问题：[BUG] deeplake v4.5.6 produces a segmentation fault with WHERE (NOT ((f9['e1'] IS NOT NULL)))
- Host AI rule: 来源显示可能已有修复、规避或版本变化，说明书中必须标注适用版本。
- Why it matters: 可能影响升级、迁移或版本选择。
- Evidence: community_evidence:github | https://github.com/activeloopai/deeplake/issues/3145 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 3: 来源证据：[BUG] deeplake 4.5.10 raises Dtype is unknown error for int * JSON, but not for JSON * JSON

- Trigger: GitHub 社区证据显示该项目存在一个配置相关的待验证问题：[BUG] deeplake 4.5.10 raises Dtype is unknown error for int * JSON, but not for JSON * JSON
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/activeloopai/deeplake/issues/3149 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 4: 能力判断依赖假设

- Trigger: README/documentation is current enough for a first validation pass.
- Host AI rule: 将假设转成下游验证清单。
- Why it matters: 假设不成立时，用户拿不到承诺的能力。
- Evidence: capability.assumptions | https://github.com/activeloopai/deeplake | README/documentation is current enough for a first validation pass.
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 5: 来源证据：[BUG] deeplake v4.5.6 raises 'deeplake._deeplake.InvalidType: Dtype is unknown.' for filter ((-104454 * f3[12]) != 0),…

- Trigger: GitHub 社区证据显示该项目存在一个运行相关的待验证问题：[BUG] deeplake v4.5.6 raises 'deeplake._deeplake.InvalidType: Dtype is unknown.' for filter ((-104454 * f3[12]) != 0), but no error for ((f3[12] * -104454) !=…
- Host AI rule: 来源显示可能已有修复、规避或版本变化，说明书中必须标注适用版本。
- Why it matters: 可能增加新用户试用和生产接入成本。
- Evidence: community_evidence:github | https://github.com/activeloopai/deeplake/issues/3147 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 6: 来源证据：[BUG] deeplake v4.5.6 returns inconsistent query results after delete() when dataset contains Zero Vectors (NaN poisoni…

- Trigger: GitHub 社区证据显示该项目存在一个维护/版本相关的待验证问题：[BUG] deeplake v4.5.6 returns inconsistent query results after delete() when dataset contains Zero Vectors (NaN poisoning breaks ORDER BY)
- Host AI rule: 来源显示可能已有修复、规避或版本变化，说明书中必须标注适用版本。
- Why it matters: 可能影响升级、迁移或版本选择。
- Evidence: community_evidence:github | https://github.com/activeloopai/deeplake/issues/3146 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 7: 来源证据：[BUG] deeplake v4.5.8 returns empty set for the same query executed after deleting some data

- Trigger: GitHub 社区证据显示该项目存在一个维护/版本相关的待验证问题：[BUG] deeplake v4.5.8 returns empty set for the same query executed after deleting some data
- Host AI rule: 来源显示可能已有修复、规避或版本变化，说明书中必须标注适用版本。
- Why it matters: 可能影响升级、迁移或版本选择。
- Evidence: community_evidence:github | https://github.com/activeloopai/deeplake/issues/3148 | 来源讨论提到 python 相关条件，需在安装/试用前复核。
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 8: 维护活跃度未知

- Trigger: 未记录 last_activity_observed。
- Host AI rule: 补 GitHub 最近 commit、release、issue/PR 响应信号。
- Why it matters: 新项目、停更项目和活跃项目会被混在一起，推荐信任度下降。
- Evidence: evidence.maintainer_signals | https://github.com/activeloopai/deeplake | last_activity_observed missing
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

- Trigger: no_demo
- Evidence: downstream_validation.risk_items | https://github.com/activeloopai/deeplake | no_demo; severity=medium
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。

### Constraint 10: 存在评分风险

- Trigger: no_demo
- Why it matters: 风险会影响是否适合普通用户安装。
- Evidence: risks.scoring_risks | https://github.com/activeloopai/deeplake | no_demo; severity=medium
- Hard boundary: 不要把这个坑点包装成已解决、已验证或可忽略，除非后续验证证据明确证明它已经关闭。
