<!-- HTML header for doxygen 1.8.7--> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/> <meta http-equiv="X-UA-Compatible" content="IE=9"/> <meta name="generator" content="Doxygen 1.8.14"/> <title>RapidJSON: Schema</title> <link href="tabs.css" rel="stylesheet" type="text/css"/> <script type="text/javascript" src="jquery.js"></script> <script type="text/javascript" src="dynsections.js"></script> <link href="navtree.css" rel="stylesheet" type="text/css"/> <script type="text/javascript" src="resize.js"></script> <script type="text/javascript" src="navtreedata.js"></script> <script type="text/javascript" src="navtree.js"></script> <script type="text/javascript"> /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ $(document).ready(initResizable); /* @license-end */</script> <link href="search/search.css" rel="stylesheet" type="text/css"/> <script type="text/javascript" src="search/searchdata.js"></script> <script type="text/javascript" src="search/search.js"></script> <script type="text/javascript"> /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ $(document).ready(function() { init_search(); }); /* @license-end */ </script> <link href="doxygen.css" rel="stylesheet" type="text/css" /> <link href="doxygenextra.css" rel="stylesheet" type="text/css"/> </head> <body> <div id="top"><!-- do not remove this div, it is closed by doxygen! --> <div id="topbanner"><a href="https://github.com/miloyip/rapidjson" title="RapidJSON GitHub"><i class="githublogo"></i></a></div> <div id="MSearchBox" class="MSearchBoxInactive"> <span class="left"> <img id="MSearchSelect" src="search/mag_sel.png" onmouseover="return searchBox.OnSearchSelectShow()" onmouseout="return searchBox.OnSearchSelectHide()" alt=""/> <input type="text" id="MSearchField" value="搜索" accesskey="S" onfocus="searchBox.OnSearchFieldFocus(true)" onblur="searchBox.OnSearchFieldFocus(false)" onkeyup="searchBox.OnSearchFieldChange(event)"/> </span><span class="right"> <a id="MSearchClose" href="javascript:searchBox.CloseResultsWindow()"><img id="MSearchCloseImg" border="0" src="search/close.png" alt=""/></a> </span> </div> <!-- end header part --> <!-- 制作者 Doxygen 1.8.14 --> <script type="text/javascript"> /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ var searchBox = new SearchBox("searchBox", "search",false,'搜索'); /* @license-end */ </script> </div><!-- top --> <div id="side-nav" class="ui-resizable side-nav-resizable"> <div id="nav-tree"> <div id="nav-tree-contents"> <div id="nav-sync" class="sync"></div> </div> </div> <div id="splitbar" style="-moz-user-select:none;" class="ui-resizable-handle"> </div> </div> <script type="text/javascript"> /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ $(document).ready(function(){initNavTree('md_doc_schema_8zh-cn.html','');}); /* @license-end */ </script> <div id="doc-content"> <!-- window showing the filter options --> <div id="MSearchSelectWindow" onmouseover="return searchBox.OnSearchSelectShow()" onmouseout="return searchBox.OnSearchSelectHide()" onkeydown="return searchBox.OnSearchSelectKey(event)"> </div> <!-- iframe showing the search results (closed by default) --> <div id="MSearchResultsWindow"> <iframe src="javascript:void(0)" frameborder="0" name="MSearchResults" id="MSearchResults"> </iframe> </div> <div class="header"> <div class="headertitle"> <div class="title">Schema </div> </div> </div><!--header--> <div class="contents"> <div class="textblock"><p>(本功能于 v1.1.0 发布)</p> <p>JSON Schema 是描述 JSON 格式的一个标准草案。一个 schema 本身也是一个 JSON。使用 JSON Schema 去校验 JSON,可以让你的代码安全地访问 DOM,而无须检查类型或键值是否存在等。这也能确保输出的 JSON 是符合指定的 schema。</p> <p>RapidJSON 实现了一个 <a href="http://json-schema.org/documentation.html">JSON Schema Draft v4</a> 的校验器。若你不熟悉 JSON Schema,可以参考 <a href="http://spacetelescope.github.io/understanding-json-schema/">Understanding JSON Schema</a>。</p> <h2>基本用法</h2> <p>首先,你要把 JSON Schema 解析成 <code>Document</code>,再把它编译成一个 <code>SchemaDocument</code>。</p> <p>然后,利用该 <code>SchemaDocument</code> 创建一个 <code>SchemaValidator</code>。它与 <code>Writer</code> 相似,都是能够处理 SAX 事件的。因此,你可以用 <code>document.Accept(validator)</code> 去校验一个 JSON,然后再获取校验结果。</p> <div class="fragment"><div class="line"><span class="preprocessor">#include "rapidjson/schema.h"</span></div><div class="line"></div><div class="line"><span class="comment">// ...</span></div><div class="line"></div><div class="line"><a class="code" href="namespacerapidjson.html#ace11b5b575baf1cccd5ba5f8586dcdc8">Document</a> sd;</div><div class="line"><span class="keywordflow">if</span> (!sd.Parse(schemaJson).HasParseError()) {</div><div class="line"> <span class="comment">// 此 schema 不是合法的 JSON</span></div><div class="line"> <span class="comment">// ... </span></div><div class="line">}</div><div class="line"><a class="code" href="namespacerapidjson.html#a52bbb5d64d1319495089e1713a0653cf">SchemaDocument</a> schema(sd); <span class="comment">// 把一个 Document 编译至 SchemaDocument</span></div><div class="line"><span class="comment">// 之后不再需要 sd</span></div><div class="line"></div><div class="line"><a class="code" href="namespacerapidjson.html#ace11b5b575baf1cccd5ba5f8586dcdc8">Document</a> d;</div><div class="line"><span class="keywordflow">if</span> (!d.Parse(inputJson).HasParseError()) {</div><div class="line"> <span class="comment">// 输入不是一个合法的 JSON</span></div><div class="line"> <span class="comment">// ... </span></div><div class="line">}</div><div class="line"></div><div class="line">SchemaValidator validator(schema);</div><div class="line"><span class="keywordflow">if</span> (!d.Accept(validator)) {</div><div class="line"> <span class="comment">// 输入的 JSON 不合乎 schema</span></div><div class="line"> <span class="comment">// 打印诊断信息</span></div><div class="line"> <a class="code" href="namespacerapidjson.html#ac0765ea91f41539645c4b78689d03f21">StringBuffer</a> sb;</div><div class="line"> validator.GetInvalidSchemaPointer().StringifyUriFragment(sb);</div><div class="line"> printf(<span class="stringliteral">"Invalid schema: %s\n"</span>, sb.GetString());</div><div class="line"> printf(<span class="stringliteral">"Invalid keyword: %s\n"</span>, validator.GetInvalidSchemaKeyword());</div><div class="line"> sb.Clear();</div><div class="line"> validator.GetInvalidDocumentPointer().StringifyUriFragment(sb);</div><div class="line"> printf(<span class="stringliteral">"Invalid document: %s\n"</span>, sb.GetString());</div><div class="line">}</div></div><!-- fragment --><p>一些注意点:</p> <ul> <li>一个 <code>SchemaDocment</code> 能被多个 <code>SchemaValidator</code> 引用。它不会被 <code>SchemaValidator</code> 修改。</li> <li>可以重复使用一个 <code>SchemaValidator</code> 来校验多个文件。在校验其他文件前,须先调用 <code>validator.Reset()</code>。</li> </ul> <h2>在解析/生成时进行校验</h2> <p>与大部分 JSON Schema 校验器有所不同,RapidJSON 提供了一个基于 SAX 的 schema 校验器实现。因此,你可以在输入流解析 JSON 的同时进行校验。若校验器遇到一个与 schema 不符的值,就会立即终止解析。这设计对于解析大型 JSON 文件时特别有用。</p> <h3>DOM 解析</h3> <p>在使用 DOM 进行解析时,<code>Document</code> 除了接收 SAX 事件外,还需做一些准备及结束工作,因此,为了连接 <code>Reader</code>、<code>SchemaValidator</code> 和 <code>Document</code> 要做多一点事情。<code>SchemaValidatingReader</code> 是一个辅助类去做那些工作。</p> <div class="fragment"><div class="line"><span class="preprocessor">#include "rapidjson/filereadstream.h"</span></div><div class="line"></div><div class="line"><span class="comment">// ...</span></div><div class="line"><a class="code" href="namespacerapidjson.html#a52bbb5d64d1319495089e1713a0653cf">SchemaDocument</a> schema(sd); <span class="comment">// 把一个 Document 编译至 SchemaDocument</span></div><div class="line"></div><div class="line"><span class="comment">// 使用 reader 解析 JSON</span></div><div class="line">FILE* fp = fopen(<span class="stringliteral">"big.json"</span>, <span class="stringliteral">"r"</span>);</div><div class="line">FileReadStream is(fp, buffer, <span class="keyword">sizeof</span>(buffer));</div><div class="line"></div><div class="line"><span class="comment">// 用 reader 解析 JSON,校验它的 SAX 事件,并存储至 d</span></div><div class="line"><a class="code" href="namespacerapidjson.html#ace11b5b575baf1cccd5ba5f8586dcdc8">Document</a> d;</div><div class="line">SchemaValidatingReader<kParseDefaultFlags, FileReadStream, UTF8<> > reader(is, schema);</div><div class="line">d.Populate(reader);</div><div class="line"></div><div class="line"><span class="keywordflow">if</span> (!reader.GetParseResult()) {</div><div class="line"> <span class="comment">// 不是一个合法的 JSON</span></div><div class="line"> <span class="comment">// 当 reader.GetParseResult().Code() == kParseErrorTermination,</span></div><div class="line"> <span class="comment">// 它可能是被以下原因中止:</span></div><div class="line"> <span class="comment">// (1) 校验器发现 JSON 不合乎 schema;或</span></div><div class="line"> <span class="comment">// (2) 输入流有 I/O 错误。</span></div><div class="line"></div><div class="line"> <span class="comment">// 检查校验结果</span></div><div class="line"> <span class="keywordflow">if</span> (!reader.IsValid()) {</div><div class="line"> <span class="comment">// 输入的 JSON 不合乎 schema</span></div><div class="line"> <span class="comment">// 打印诊断信息</span></div><div class="line"> <a class="code" href="namespacerapidjson.html#ac0765ea91f41539645c4b78689d03f21">StringBuffer</a> sb;</div><div class="line"> reader.GetInvalidSchemaPointer().StringifyUriFragment(sb);</div><div class="line"> printf(<span class="stringliteral">"Invalid schema: %s\n"</span>, sb.GetString());</div><div class="line"> printf(<span class="stringliteral">"Invalid keyword: %s\n"</span>, reader.GetInvalidSchemaKeyword());</div><div class="line"> sb.Clear();</div><div class="line"> reader.GetInvalidDocumentPointer().StringifyUriFragment(sb);</div><div class="line"> printf(<span class="stringliteral">"Invalid document: %s\n"</span>, sb.GetString());</div><div class="line"> }</div><div class="line">}</div></div><!-- fragment --><h3>SAX 解析</h3> <p>使用 SAX 解析时,情况就简单得多。若只需要校验 JSON 而无需进一步处理,那么仅需要:</p> <div class="fragment"><div class="line">SchemaValidator validator(schema);</div><div class="line">Reader reader;</div><div class="line">if (!reader.Parse(stream, validator)) {</div><div class="line"> if (!validator.IsValid()) {</div><div class="line"> // ... </div><div class="line"> }</div><div class="line">}</div></div><!-- fragment --><p>这种方式和 <a href="example/schemavalidator/schemavalidator.cpp">schemavalidator</a> 例子完全相同。这带来的独特优势是,无论 JSON 多巨大,永远维持低内存用量(内存用量只与 Schema 的复杂度相关)。</p> <p>若你需要进一步处理 SAX 事件,便可使用模板类 <code>GenericSchemaValidator</code> 去设置校验器的输出 <code>Handler</code>:</p> <div class="fragment"><div class="line">MyHandler handler;</div><div class="line">GenericSchemaValidator<SchemaDocument, MyHandler> validator(schema, handler);</div><div class="line">Reader reader;</div><div class="line">if (!reader.Parse(ss, validator)) {</div><div class="line"> if (!validator.IsValid()) {</div><div class="line"> // ... </div><div class="line"> }</div><div class="line">}</div></div><!-- fragment --><h3>生成</h3> <p>我们也可以在生成(serialization)的时候进行校验。这能确保输出的 JSON 符合一个 JSON Schema。</p> <div class="fragment"><div class="line">StringBuffer sb;</div><div class="line">Writer<StringBuffer> writer(sb);</div><div class="line">GenericSchemaValidator<SchemaDocument, Writer<StringBuffer> > validator(s, writer);</div><div class="line">if (!d.Accept(validator)) {</div><div class="line"> // Some problem during Accept(), it may be validation or encoding issues.</div><div class="line"> if (!validator.IsValid()) {</div><div class="line"> // ...</div><div class="line"> }</div><div class="line">}</div></div><!-- fragment --><p>当然,如果你的应用仅需要 SAX 风格的生成,那么只需要把 SAX 事件由原来发送到 <code>Writer</code>,改为发送到 <code>SchemaValidator</code>。</p> <h2>远程 Schema</h2> <p>JSON Schema 支持 <a href="http://spacetelescope.github.io/understanding-json-schema/structuring.html"><code>$ref</code> 关键字</a>,它是一个 <a class="el" href="md_doc_pointer_8zh-cn.html">JSON pointer</a> 引用至一个本地(local)或远程(remote) schema。本地指针的首字符是 <code>#</code>,而远程指针是一个相对或绝对 URI。例如:</p> <div class="fragment"><div class="line">{ "$ref": "definitions.json#/address" }</div></div><!-- fragment --><p>由于 <code>SchemaDocument</code> 并不知道如何处理那些 URI,它需要使用者提供一个 <code>IRemoteSchemaDocumentProvider</code> 的实例去处理。</p> <div class="fragment"><div class="line">class MyRemoteSchemaDocumentProvider : public IRemoteSchemaDocumentProvider {</div><div class="line">public:</div><div class="line"> virtual const SchemaDocument* GetRemoteDocument(const char* uri, SizeTyp length) {</div><div class="line"> // Resolve the uri and returns a pointer to that schema.</div><div class="line"> }</div><div class="line">};</div><div class="line"></div><div class="line">// ...</div><div class="line"></div><div class="line">MyRemoteSchemaDocumentProvider provider;</div><div class="line">SchemaDocument schema(sd, &provider);</div></div><!-- fragment --><h2>标准的符合程度</h2> <p>RapidJSON 通过了 <a href="https://github.com/json-schema/JSON-Schema-Test-Suite">JSON Schema Test Suite</a> (Json Schema draft 4) 中 263 个测试的 262 个。</p> <p>没通过的测试是 <code>refRemote.json</code> 中的 "change resolution scope" - "changed scope ref invalid"。这是由于未实现 <code>id</code> schema 关键字及 URI 合并功能。</p> <p>除此以外,关于字符串类型的 <code>format</code> schema 关键字也会被忽略,因为标准中并没需求必须实现。</p> <h3>正则表达式</h3> <p><code>pattern</code> 及 <code>patternProperties</code> 这两个 schema 关键字使用了正则表达式去匹配所需的模式。</p> <p>RapidJSON 实现了一个简单的 NFA 正则表达式引擎,并预设使用。它支持以下语法。</p> <table class="markdownTable"> <tr class="markdownTableHead"> <th class="markdownTableHeadNone">语法 </th><th class="markdownTableHeadNone">描述 </th></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone"><code>ab</code> </td><td class="markdownTableBodyNone">串联 </td></tr> </table> <p>|<code>a|b</code> | 交替 | |<code>a?</code> | 零或一次 | |<code>a*</code> | 零或多次 | |<code>a+</code> | 一或多次 | |<code>a{3}</code> | 刚好 3 次 | |<code>a{3,}</code> | 至少 3 次 | |<code>a{3,5}</code>| 3 至 5 次 | |<code>(ab)</code> | 分组 | |<code>^a</code> | 在开始处 | |<code>a$</code> | 在结束处 | |<code>.</code> | 任何字符 | |<code>[abc]</code> | 字符组 | |<code>[a-c]</code> | 字符组范围 | |<code>[a-z0-9_]</code> | 字符组组合 | |<code>[^abc]</code> | 字符组取反 | |<code>[^a-c]</code> | 字符组范围取反 | |<code>[\b]</code> | 退格符 (U+0008) | |<code>\|</code>, <code>\\</code>, ... | 转义字符 | |<code>\f</code> | 馈页 (U+000C) | |<code>\n</code> | 馈行 (U+000A) | |<code>\r</code> | 回车 (U+000D) | |<code>\t</code> | 制表 (U+0009) | |<code>\v</code> | 垂直制表 (U+000B) |</p> <p>对于使用 C++11 编译器的使用者,也可使用 <code>std::regex</code>,只需定义 <code>RAPIDJSON_SCHEMA_USE_INTERNALREGEX=0</code> 及 <code>RAPIDJSON_SCHEMA_USE_STDREGEX=1</code>。若你的 schema 无需使用 <code>pattern</code> 或 <code>patternProperties</code>,可以把两个宏都设为零,以禁用此功能,这样做可节省一些代码体积。</p> <h2>性能</h2> <p>大部分 C++ JSON 库都未支持 JSON Schema。因此我们尝试按照 <a href="https://github.com/ebdrup/json-schema-benchmark">json-schema-benchmark</a> 去评估 RapidJSON 的 JSON Schema 校验器。该评测测试了 11 个运行在 node.js 上的 JavaScript 库。</p> <p>该评测校验 <a href="https://github.com/json-schema/JSON-Schema-Test-Suite">JSON Schema Test Suite</a> 中的测试,当中排除了一些测试套件及个别测试。我们在 <a href="test/perftest/schematest.cpp"><code>schematest.cpp</code></a> 实现了相同的评测。</p> <p>在 MacBook Pro (2.8 GHz Intel Core i7) 上收集到以下结果。</p> <table class="markdownTable"> <tr class="markdownTableHead"> <th class="markdownTableHeadNone">校验器 </th><th class="markdownTableHeadCenter">相对速度 </th><th class="markdownTableHeadCenter">每秒执行的测试数目 </th></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone">RapidJSON </td><td class="markdownTableBodyCenter">155% </td><td class="markdownTableBodyCenter">30682 </td></tr> <tr class="markdownTableBody" class="markdownTableRowEven"> <td class="markdownTableBodyNone"><a href="https://github.com/epoberezkin/ajv"><code>ajv</code></a> </td><td class="markdownTableBodyCenter">100% </td><td class="markdownTableBodyCenter">19770 (± 1.31%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone"><a href="https://github.com/mafintosh/is-my-json-valid"><code>is-my-json-valid</code></a> </td><td class="markdownTableBodyCenter">70% </td><td class="markdownTableBodyCenter">13835 (± 2.84%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowEven"> <td class="markdownTableBodyNone"><a href="https://github.com/bugventure/jsen"><code>jsen</code></a> </td><td class="markdownTableBodyCenter">57.7% </td><td class="markdownTableBodyCenter">11411 (± 1.27%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone"><a href="https://github.com/AlexeyGrishin/schemasaurus"><code>schemasaurus</code></a> </td><td class="markdownTableBodyCenter">26% </td><td class="markdownTableBodyCenter">5145 (± 1.62%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowEven"> <td class="markdownTableBodyNone"><a href="https://github.com/playlyfe/themis"><code>themis</code></a> </td><td class="markdownTableBodyCenter">19.9% </td><td class="markdownTableBodyCenter">3935 (± 2.69%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone"><a href="https://github.com/zaggino/z-schema"><code>z-schema</code></a> </td><td class="markdownTableBodyCenter">7% </td><td class="markdownTableBodyCenter">1388 (± 0.84%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowEven"> <td class="markdownTableBodyNone"><a href="https://github.com/pandastrike/jsck#readme"><code>jsck</code></a> </td><td class="markdownTableBodyCenter">3.1% </td><td class="markdownTableBodyCenter">606 (± 2.84%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone"><a href="https://github.com/tdegrunt/jsonschema#readme"><code>jsonschema</code></a> </td><td class="markdownTableBodyCenter">0.9% </td><td class="markdownTableBodyCenter">185 (± 1.01%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowEven"> <td class="markdownTableBodyNone"><a href="https://github.com/Prestaul/skeemas#readme"><code>skeemas</code></a> </td><td class="markdownTableBodyCenter">0.8% </td><td class="markdownTableBodyCenter">154 (± 0.79%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowOdd"> <td class="markdownTableBodyNone">tv4 </td><td class="markdownTableBodyCenter">0.5% </td><td class="markdownTableBodyCenter">93 (± 0.94%) </td></tr> <tr class="markdownTableBody" class="markdownTableRowEven"> <td class="markdownTableBodyNone"><a href="https://github.com/natesilva/jayschema"><code>jayschema</code></a> </td><td class="markdownTableBodyCenter">0.1% </td><td class="markdownTableBodyCenter">21 (± 1.14%) </td></tr> </table> <p>换言之,RapidJSON 比最快的 JavaScript 库(ajv)快约 1.5x。比最慢的快 1400x。 </p> </div></div><!-- contents --> </div><!-- doc-content --> <!-- HTML footer for doxygen 1.8.7--> <!-- start footer part --> <div id="nav-path" class="navpath"><!-- id is needed for treeview function! --> <ul> </ul> </div> </body> </html>