如果您的代理使用语音识别(也称为音频输入或语音转文字 (STT)),您的正则表达式在匹配字母和数字时需要特殊处理。
最终用户所说话语经过语音识别程序处理后才能进行实体匹配。
当话语中包含一系列字母或数字时,识别程序会用空格填充各个字符。
此外,识别程序会用字词形式解读数字。
举例来说,最终用户说出的“My ID is 123”可能会被识别为以下任意一种形式:
[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-04-03。"],[[["Regexp entities allow matching patterns like national IDs or license plates using regular expressions."],["Multiple regular expressions for a single entity are combined into a compound regular expression using the alternation operator (`|`)."],["The order of regular expressions within a compound expression matters, as matching stops after the first valid match is found."],["Speech recognition requires special handling for regexp entities because the speech recognizer might add spaces between characters or interpret digits as words."],["There are limitations to using regexp entities, including no fuzzy matching, a maximum of 50 regexp entities per agent, and a 2000-character limit for compound regular expressions."]]],[]]