cuda-webdoc-search

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

CUDA-X Documentation Search Guide

CUDA-X文档搜索指南

Overview

概述

Search and discover APIs across CUDA-X library documentation by querying Sphinx inventory files. Use this skill when you need to find specific CUDA library functions, check available APIs, or get documentation links.

通过查询Sphinx清单文件，搜索并发现CUDA-X库文档中的各类API。当你需要查找特定CUDA库函数、查看可用API或获取文档链接时，可使用该工具。

Setup

安装配置

bash

uv tool install "cuda-webdoc-search @ git+https://github.com/ultimatile/cuda-x-skills.git#subdirectory=skills/cuda-webdoc-search"

After installation,

cws

is available as a standalone command. For one-shot usage without installing:

bash

uvx --from "cuda-webdoc-search @ git+https://github.com/ultimatile/cuda-x-skills.git#subdirectory=skills/cuda-webdoc-search" cws search cublas --stats

bash

uv tool install "cuda-webdoc-search @ git+https://github.com/ultimatile/cuda-x-skills.git#subdirectory=skills/cuda-webdoc-search"

安装完成后，

cws

将作为独立命令可用。若无需安装即可一次性使用：

bash

uvx --from "cuda-webdoc-search @ git+https://github.com/ultimatile/cuda-x-skills.git#subdirectory=skills/cuda-webdoc-search" cws search cublas --stats

When to Use

使用场景

Finding CUDA library APIs (e.g., "what SVD functions does cuTensorNet have?")
Checking if a specific function exists in a library
Getting documentation URLs for CUDA APIs
Exploring available operations in a library

查找CUDA库API（例如：“cuTensorNet有哪些SVD函数？”）
检查某个特定函数是否存在于某库中
获取CUDA API的文档URL
探索库中的可用操作

Quick Reference

快速参考

Check Available Domains First

先查看可用领域

Before searching, check what domains (API types) are available:

bash

cws search <library> --stats

This returns domain counts:

cpp

(C++ APIs),

(C APIs),

py

(Python bindings),

std

(doc labels).

搜索前，先查看可用的领域（API类型）：

bash

cws search <library> --stats

该命令会返回领域数量：

cpp

（C++ API）、

（C API）、

py

（Python绑定）、

std

（文档标签）。

Search for APIs

搜索API

bash

cws search <library> --domains <domain> --keywords "<terms>" --fuzzy --limit 20

bash

cws search <library> --domains <domain> --keywords "<terms>" --fuzzy --limit 20

Keyword Syntax (fzf subset)

关键字语法（fzf子集）

Space-separated terms are AND (all must match):
```
--keywords "SVD batch"
```
Use
```
|
```
for OR (requires shell quoting):
```
--keywords "SVD | QR"
```
AND binds tighter than OR:
```
--keywords "a b | c"
```
= (a AND b) OR c
Note: Only AND/OR via
```
|
```
is supported. fzf operators
```
^
```
,
```
'
```
,
```
!
```
,
```
$
```
are not available.

空格分隔的术语表示AND（必须全部匹配）：
```
--keywords "SVD batch"
```
使用
```
|
```
表示OR（需要shell引号包裹）：
```
--keywords "SVD | QR"
```
AND的优先级高于OR：
```
--keywords "a b | c"
```
= (a AND b) OR c
注意：仅支持通过
```
|
```
实现AND/OR逻辑。fzf的
```
^
```
、
```
'
```
、
```
!
```
、
```
$
```
运算符暂不支持。

Available Libraries

可用库

See

registry.toml

for the full list. Common ones:

```
cuquantum
```
- cuTensorNet, cuStateVec (quantum/tensor network)
```
cublas
```
- BLAS operations
```
cusolver
```
- Dense/sparse solvers
```
cusparse
```
- Sparse matrix operations
```
cufft
```
- FFT operations
```
cudnn
```
- Deep learning primitives
```
cutlass
```
- GEMM templates
```
cuda_math
```
- Math functions

完整列表请查看

registry.toml

。常用库包括：

```
cuquantum
```
- cuTensorNet、cuStateVec（量子/张量网络）
```
cublas
```
- BLAS运算
```
cusolver
```
- 稠密/稀疏求解器
```
cusparse
```
- 稀疏矩阵运算
```
cufft
```
- FFT运算
```
cudnn
```
- 深度学习原语
```
cutlass
```
- GEMM模板
```
cuda_math
```
- 数学函数

Workflow

使用流程

Identify the library: Determine which CUDA library likely contains the API
Check domains: Run
```
--stats
```
to see available domain types
Search: Use
```
--keywords
```
with
```
--fuzzy
```
for flexible matching
Filter by domain: Use
```
--domains cpp
```
for C++ APIs,
```
--domains c
```
for C APIs
Extract details: Use
```
cws get
```
to get full documentation content

确定库：判断哪个CUDA库可能包含所需API
查看领域：运行
```
--stats
```
查看可用领域类型
搜索：结合
```
--keywords
```
和
```
--fuzzy
```
进行灵活匹配
按领域过滤：使用
```
--domains cpp
```
筛选C++ API，
```
--domains c
```
筛选C API
提取详情：使用
```
cws get
```
获取完整文档内容

Examples

使用示例

Find tensor decomposition APIs in cuTensorNet

在cuTensorNet中查找张量分解API

bash

cws search cuquantum --stats

bash

cws search cuquantum --stats

Shows: cpp:3179, py:1172, std:3378, c:4

输出：cpp:3179, py:1172, std:3378, c:4

cws search cuquantum --domains cpp --keywords "SVD" --fuzzy --limit 10

Returns: cutensornetTensorSVD, etc. (functions ranked above enumerators)

返回：cutensornetTensorSVD等（函数排序优先于枚举器）

cws search cuquantum --domains cpp --keywords "SVD | QR" --fuzzy

Returns: entries matching SVD OR QR

返回：匹配SVD或QR的条目

undefined

undefined

Search across multiple libraries

跨多个库搜索

bash

undefined

bash

undefined

"Where is SVD in CUDA-X?" — search cuSOLVER and cuDSS together

“SVD在CUDA-X的哪个库中？”——同时搜索cuSOLVER和cuDSS

cws search cusolver cudss --keywords "svd" --fuzzy --limit 10

Mix any number of sources

混合任意数量的源库

cws search cusolver cusparse cudss --keywords "solve" --fuzzy


Multi-source output uses `"sources"` (list) instead of `"source"` (string), and `"total_found"` becomes a per-source dict.

cws search cusolver cusparse cudss --keywords "solve" --fuzzy


多源输出会使用`sources`（列表）替代`source`（字符串），`total_found`会变为按源库统计的字典。

Find GEMM functions in cuBLAS

在cuBLAS中查找GEMM函数

bash

cws search cublas --stats

bash

cws search cublas --stats

Shows: std:36 (only doc labels, no cpp/c domain)

输出：std:36（仅包含文档标签，无cpp/c领域）

cws search cublas --domains std --keywords "gemm"

Returns: cublas-t-gemm, cublas-t-gemmex, etc. (doc section labels)

返回：cublas-t-gemm、cublas-t-gemmex等（文档章节标签）

undefined

undefined

Extract API documentation details

提取API文档详情

After finding a function URL, extract its full documentation:

bash

cws get <url> --section <function_name>

Example:

bash

cws get \
  "https://docs.nvidia.com/cuda/cuquantum/latest/cutensornet/api/functions.html" \
  --section "cutensornetTensorSVD"

Output is a brace-delimited text tree:

{
cutensornetTensorSVD
{
cutensornetStatus_t cutensornetTensorSVD {
const cutensornetHandle_t handle
const cutensornetTensorDescriptor_t descTensorIn
...
}
Performs SVD decomposition of a tensor. { ... }
Parameters { ... }
}
}

找到函数URL后，提取其完整文档：

bash

cws get <url> --section <function_name>

示例：

bash

cws get \
  "https://docs.nvidia.com/cuda/cuquantum/latest/cutensornet/api/functions.html" \
  --section "cutensornetTensorSVD"

输出为大括号分隔的文本树：

{
cutensornetTensorSVD
{
cutensornetStatus_t cutensornetTensorSVD {
const cutensornetHandle_t handle
const cutensornetTensorDescriptor_t descTensorIn
...
}
Performs SVD decomposition of a tensor. { ... }
Parameters { ... }
}
}

List available sources

列出可用源库

bash

cws list                     # table format
cws list --json              # JSON format

bash

cws list                     # 表格格式
cws list --json              # JSON格式

Audit registry health

检查注册表健康状态

bash

cws audit                    # audit all sources
cws audit --source cublas    # audit single source

bash

cws audit                    # 检查所有源库
cws audit --source cublas    # 检查单个源库

Output Format

输出格式

cws search

JSON output includes:

Single source:

```
source
```
: Library name (string)
```
total_found
```
: Total APIs in filtered domains
```
filtered_count
```
: APIs matching keywords
```
candidates
```
: List of matching APIs with
```
group
```
(name),
```
url
```
,
```
domain
```
,
```
role
```

Multi-source (

cws search A B

```
sources
```
: Library names (list)
```
total_found
```
: Per-source totals (dict)
```
filtered_count
```
: Total APIs matching keywords across all sources
```
candidates
```
: Merged list (k-way merge by score if
```
--fuzzy
```
, else concatenated)

JSON输出包含：

单源库：

```
source
```
: 库名称（字符串）
```
total_found
```
: 过滤后领域中的API总数
```
filtered_count
```
: 匹配关键字的API数量
```
candidates
```
: 匹配API列表，包含
```
group
```
（名称）、
```
url
```
、
```
domain
```
、
```
role
```

多源库（

cws search A B

）：

```
sources
```
: 库名称列表
```
total_found
```
: 按源库统计的总数（字典）
```
filtered_count
```
: 所有源库中匹配关键字的API总数
```
candidates
```
: 合并后的列表（若使用
```
--fuzzy
```
则按得分排序，否则直接拼接）

cws get

Brace-delimited text tree where:

```
{
```
```
}
```
denote hierarchy
Text content is preserved
Inline elements (code, links) are flattened to text

Options:

```
--section <id>
```
: Extract only a specific section
```
--main-only
```
: Extract only the main content area

大括号分隔的文本树，其中：

```
{
```
```
}
```
表示层级结构
文本内容保留原样
内联元素（代码、链接）转为纯文本

可选参数：

```
--section <id>
```
: 仅提取特定章节
```
--main-only
```
: 仅提取主内容区域

Files

文件说明

```
cli.py
```
- Unified CLI entry point (
```
cws
```
)
```
search.py
```
- Search APIs across library inventories
```
audit.py
```
- Audit registry entries for endpoint health
```
get.py
```
- Extract documentation content as text tree
```
fetchers.py
```
- Data fetching for Sphinx/Doxygen sources
```
scoring.py
```
- Search ranking and filtering logic
```
registry.py
```
- Registry loading utility
```
registry.toml
```
- Library metadata (URLs, doc types)

```
cli.py
```
- 统一CLI入口（
```
cws
```
）
```
search.py
```
- 跨库清单搜索API
```
audit.py
```
- 检查注册表条目的端点健康状态
```
get.py
```
- 提取文档内容为文本树
```
fetchers.py
```
- Sphinx/Doxygen源的数据获取
```
scoring.py
```
- 搜索排序与过滤逻辑
```
registry.py
```
- 注册表加载工具
```
registry.toml
```
- 库元数据（URL、文档类型）