{"id":84614,"date":"2020-05-15T21:00:21","date_gmt":"2020-05-15T13:00:21","guid":{"rendered":"http:\/\/4563.org\/?p=84614"},"modified":"2020-05-15T21:00:21","modified_gmt":"2020-05-15T13:00:21","slug":"serverless-%e5%ae%9e%e6%88%98%ef%bc%9a%e5%a6%82%e4%bd%95%e7%bb%93%e5%90%88-nlp-%e5%ae%9e%e7%8e%b0%e6%96%87%e6%9c%ac%e6%91%98%e8%a6%81%e5%92%8c%e5%85%b3%e9%94%ae%e8%af%8d%e6%8f%90%e5%8f%96%ef%bc%9f","status":"publish","type":"post","link":"http:\/\/4563.org\/?p=84614","title":{"rendered":"Serverless \u5b9e\u6218\uff1a\u5982\u4f55\u7ed3\u5408 NLP \u5b9e\u73b0\u6587\u672c\u6458\u8981\u548c\u5173\u952e\u8bcd\u63d0\u53d6\uff1f"},"content":{"rendered":"<div>\n<div>\n<div>\n<h1>                  Serverless \u5b9e\u6218\uff1a\u5982\u4f55\u7ed3\u5408 NLP \u5b9e\u73b0\u6587\u672c\u6458\u8981\u548c\u5173\u952e\u8bcd\u63d0\u53d6\uff1f               <\/h1>\n<p> <\/p>\n<div>\n<div> <span>\u8cc7\u6df1\u5927\u4f6c : scf10cent <\/span>  <span><i><\/i> 7<\/span> <\/div>\n<div> <\/div>\n<\/p><\/div>\n<\/p><\/div>\n<\/p><\/div>\n<div isfirst=\"1\"> <\/p>\n<p>\u5bf9\u6587\u672c\u8fdb\u884c\u81ea\u52a8\u6458\u8981\u7684\u63d0\u53d6\u548c\u5173\u952e\u8bcd\u7684\u63d0\u53d6\uff0c\u5c5e\u4e8e\u81ea\u7136\u8bed\u8a00\u5904\u7406\u7684\u8303\u7574\u3002\u63d0\u53d6\u6458\u8981\u7684\u4e00\u4e2a\u597d\u5904\u662f\u53ef\u4ee5\u8ba9\u9605\u8bfb\u8005\u901a\u8fc7\u6700\u5c11\u7684\u4fe1\u606f\u5224\u65ad\u51fa\u8fd9\u4e2a\u6587\u7ae0\u5bf9\u81ea\u5df1\u662f\u5426\u6709\u610f\u4e49\u6216\u8005\u4ef7\u503c\uff0c\u662f\u5426\u9700\u8981\u8fdb\u884c\u66f4\u52a0\u8be6\u7ec6\u7684\u9605\u8bfb\uff1b\u800c\u63d0\u53d6\u5173\u952e\u8bcd\u7684\u597d\u5904\u662f\u53ef\u4ee5\u8ba9\u6587\u7ae0\u4e0e\u6587\u7ae0\u4e4b\u95f4\u4ea7\u751f\u5173\u8054\uff0c\u540c\u65f6\u4e5f\u53ef\u4ee5\u8ba9\u8bfb\u8005\u901a\u8fc7\u5173\u952e\u8bcd\u5feb\u901f\u5b9a\u4f4d\u5230\u548c\u8be5\u5173\u952e\u8bcd\u76f8\u5173\u7684\u6587\u7ae0\u5185\u5bb9\u3002<\/p>\n<p>\u6587\u672c\u6458\u8981\u548c\u5173\u952e\u8bcd\u63d0\u53d6\u90fd\u53ef\u4ee5\u548c\u4f20\u7edf\u7684 CMS \u8fdb\u884c\u7ed3\u5408\uff0c\u901a\u8fc7\u5bf9\u6587\u7ae0 \/ \u65b0\u95fb\u7b49\u53d1\u5e03\u529f\u80fd\u8fdb\u884c\u6539\u9020\uff0c\u540c\u6b65\u63d0\u53d6\u5173\u952e\u8bcd\u548c\u6458\u8981\uff0c\u653e\u5230 HTML \u9875\u9762\u4e2d\u4f5c\u4e3a Description \u548c Keyworks \u3002\u8fd9\u6837\u505a\u5728\u4e00\u5b9a\u7a0b\u5ea6\u4e0a\u6709\u5229\u4e8e\u641c\u7d22\u5f15\u64ce\u6536\u5f55\uff0c\u5c5e\u4e8e SEO \u4f18\u5316\u7684\u8303\u7574\u3002<\/p>\n<h2>\u5173\u952e\u8bcd\u63d0\u53d6<\/h2>\n<p>\u5173\u952e\u8bcd\u63d0\u53d6\u7684\u65b9\u6cd5\u5f88\u591a\uff0c\u4f46\u662f\u6700\u5e38\u89c1\u7684\u5e94\u8be5\u5c31\u662f<code>tf-idf<\/code>\u4e86\u3002<\/p>\n<p>\u901a\u8fc7<code>jieba<\/code>\u5b9e\u73b0\u57fa\u4e8e<code>tf-idf<\/code>\u5173\u952e\u8bcd\u63d0\u53d6\u7684\u65b9\u6cd5\uff1a<\/p>\n<pre><code>jieba.analyse.extract_tags(text, topK=5, withWeight=False, allowPOS=('n', 'vn', 'v')) <\/code><\/pre>\n<h2>\u6587\u672c\u6458\u8981<\/h2>\n<p>\u6587\u672c\u6458\u8981\u7684\u65b9\u6cd5\u4e5f\u6709\u5f88\u591a\uff0c\u5982\u679c\u4ece\u5e7f\u4e49\u4e0a\u6765\u5212\u5206\uff0c\u5305\u62ec\u63d0\u53d6\u5f0f\u548c\u751f\u6210\u5f0f\u3002\u5176\u4e2d\u63d0\u53d6\u5f0f\u5c31\u662f\u5728\u6587\u7ae0\u4e2d\u901a\u8fc7<code>TextRank<\/code>\u7b49\u7b97\u6cd5\uff0c\u627e\u51fa\u5173\u952e\u53e5\u7136\u540e\u8fdb\u884c\u62fc\u88c5\uff0c\u5f62\u6210\u6458\u8981\uff0c\u8fd9\u79cd\u65b9\u6cd5\u76f8\u5bf9\u6765\u8bf4\u6bd4\u8f83\u7b80\u5355\uff0c\u4f46\u662f\u5f88\u96be\u63d0\u53d6\u51fa\u771f\u5b9e\u7684\u8bed\u4e49\u7b49\uff1b\u53e6\u4e00\u79cd\u65b9\u6cd5\u662f\u751f\u6210\u5f0f\uff0c\u901a\u8fc7\u6df1\u5ea6\u5b66\u4e60\u7b49\u65b9\u6cd5\uff0c\u5bf9\u6587\u672c\u8bed\u4e49\u8fdb\u884c\u63d0\u53d6\u518d\u751f\u6210\u6458\u8981\u3002<\/p>\n<p>\u5982\u679c\u7b80\u5355\u7406\u89e3\uff0c\u63d0\u53d6\u5f0f\u65b9\u5f0f\u751f\u6210\u7684\u6458\u8981\uff0c\u6240\u6709\u53e5\u5b50\u6765\u81ea\u539f\u6587\uff0c\u800c\u751f\u6210\u5f0f\u65b9\u6cd5\u5219\u662f\u72ec\u7acb\u751f\u6210\u7684\u3002<\/p>\n<p>\u4e3a\u4e86\u7b80\u5316\u96be\u5ea6\uff0c\u672c\u6587\u5c06\u91c7\u7528\u63d0\u53d6\u5f0f\u6765\u5b9e\u73b0\u6587\u672c\u6458\u8981\u529f\u80fd\uff0c\u901a\u8fc7 SnowNLP \u7b2c\u4e09\u65b9\u5e93\uff0c\u5b9e\u73b0\u57fa\u4e8e<code>TextRank<\/code>\u7684\u6587\u672c\u6458\u8981\u529f\u80fd\u3002\u6211\u4eec\u4ee5\u300a\u6d77\u5e95\u4e24\u4e07\u91cc\u300b\u90e8\u5206\u5185\u5bb9\u4f5c\u4e3a\u539f\u6587\uff0c\u8fdb\u884c\u6458\u8981\u751f\u6210\uff1a<\/p>\n<p>\u539f\u6587\uff1a<\/p>\n<blockquote>\n<p>\u8fd9\u4e9b\u4e8b\u4ef6\u53d1\u751f\u65f6\uff0c\u6211\u521a\u4ece\u7f8e\u56fd\u5185\u5e03\u62c9\u65af\u52a0\u5dde\u7684\u8d2b\u7620\u5730\u533a\u505a\u5b8c\u4e00\u9879\u79d1\u8003\u5de5\u4f5c\u56de\u6765\u3002\u6211\u5f53\u65f6\u662f\u5df4\u9ece\u81ea\u7136\u53f2\u535a\u7269\u9986\u7684\u5ba2\u5ea7\u6559\u6388\uff0c\u6cd5\u56fd\u653f\u5e9c\u6d3e\u6211\u53c2\u52a0\u8fd9\u6b21\u8003\u5bdf\u6d3b\u52a8\u3002\u6211\u5728\u5185\u5e03\u62c9\u65af\u52a0\u5dde\u5ea6\u8fc7\u4e86\u534a\u5e74\u65f6\u95f4\uff0c\u6536\u96c6\u4e86\u8bb8\u591a\u73cd\u8d35\u8d44\u6599\uff0c\u6ee1\u8f7d\u800c\u5f52\uff0c3 \u6708\u5e95\u62b5\u8fbe\u7ebd\u7ea6\u3002\u6211\u51b3\u5b9a 5 \u6708\u521d\u52a8\u8eab\u56de\u6cd5\u56fd\u3002\u4e8e\u662f\uff0c\u6211\u5c31\u6293\u7d27\u8fd9\u6bb5\u5019\u8239\u9017\u7559\u65f6\u95f4\uff0c\u628a\u6536\u96c6\u5230\u7684\u77ff\u7269\u548c\u52a8\u690d\u7269\u6807\u672c\u8fdb\u884c\u5206\u7c7b\u6574\u7406\uff0c\u53ef\u5c31\u5728\u8fd9\u65f6\uff0c\u65af\u79d1\u820d\u53f7\u51fa\u4e8b\u4e86\u3002 \u6211\u5bf9\u5f53\u65f6\u7684\u8857\u8c08\u5df7\u8bae\u81ea\u7136\u4e86\u5982\u6307\u638c\uff0c\u518d\u8bf4\u4e86\uff0c\u6211\u600e\u80fd\u542c\u800c\u4e0d\u95fb\u3001\u65e0\u52a8\u4e8e\u8877\u5462\uff1f\u6211\u628a\u7f8e\u56fd\u548c\u6b27\u6d32\u7684\u5404\u79cd\u62a5\u520a\u8bfb\u4e86\u53c8\u8bfb\uff0c\u4f46\u672a\u80fd\u6df1\u5165\u4e86\u89e3\u771f\u76f8\u3002\u795e\u79d8\u83ab\u6d4b\uff0c\u767e\u601d\u4e0d\u5f97\u5176\u89e3\u3002\u6211\u5de6\u601d\u53f3\u60f3\uff0c\u6447\u6446\u4e8e\u4e24\u4e2a\u6781\u7aef\u4e4b\u95f4\uff0c\u59cb\u7ec8\u5f62\u4e0d\u6210\u4e00\u79cd\u89c1\u89e3\u3002\u5176\u4e2d\u80af\u5b9a\u6709\u540d\u5802\uff0c\u8fd9\u662f\u4e0d\u5bb9\u7f6e\u7591\u7684\uff0c\u5982\u679c\u6709\u4eba\u8868\u793a\u6000\u7591\uff0c\u5c31\u8bf7\u4ed6\u4eec\u53bb\u6478\u4e00\u6478\u65af\u79d1\u820d\u53f7\u7684\u4f24\u53e3\u597d\u4e86\u3002 \u6211\u5230\u7ebd\u7ea6\u65f6\uff0c\u8fd9\u4e2a\u95ee\u9898\u6b63\u7092\u5f97\u6cb8\u53cd\u76c8\u5929\u3002\u67d0\u4e9b\u4e0d\u5b66\u65e0\u672f\u4e4b\u5f92\u63d0\u51fa\u8bbe\u60f3\uff0c\u6709\u8bf4\u662f\u6d6e\u52a8\u7684\u5c0f\u5c9b\uff0c\u4e5f\u6709\u8bf4\u662f\u4e0d\u53ef\u6349\u6478\u7684\u6697\u7901\uff0c\u4e0d\u8fc7\uff0c\u8fd9\u4e9b\u4e2a\u5047\u8bbe\u901a\u901a\u90fd\u88ab\u63a8\u7ffb\u4e86\u3002\u5f88\u663e\u7136\uff0c\u9664\u975e\u8fd9\u6697\u7901\u8179\u90e8\u88c5\u6709\u673a\u5668\uff0c\u4e0d\u7136\u7684\u8bdd\uff0c\u5b83\u600e\u80fd\u5982\u6b64\u5feb\u901f\u5730\u8f6c\u79fb\u5462\uff1f \u540c\u6837\u7684\u9053\u7406\uff0c\u8bf4\u5b83\u662f\u4e00\u5757\u6d6e\u52a8\u7684\u8239\u4f53\u6216\u662f\u4e00\u5806\u5927\u8239\u6b8b\u7247\uff0c\u8fd9\u79cd\u5047\u8bbe\u4e5f\u4e0d\u80fd\u6210\u7acb\uff0c\u7406\u7531\u4ecd\u7136\u662f\u79fb\u52a8\u901f\u5ea6\u592a\u5feb\u3002 \u90a3\u4e48\uff0c\u95ee\u9898\u53ea\u80fd\u6709\u4e24\u79cd\u89e3\u91ca\uff0c\u4eba\u4eec\u5404\u6301\u5df1\u89c1\uff0c\u81ea\u7136\u5c31\u5206\u6210\u89c2\u70b9\u622a\u7136\u4e0d\u540c\u7684\u4e24\u6d3e\uff1a\u4e00\u6d3e\u8bf4\u8fd9\u662f\u4e00\u4e2a\u529b\u5927\u65e0\u6bd4\u7684\u602a\u7269\uff0c\u53e6\u4e00\u6d3e\u8bf4\u8fd9\u662f\u4e00\u8258\u52a8\u529b\u6781\u5f3a\u7684\u201c\u6f5c\u6c34\u8239\u201d\u3002 \u54e6\uff0c\u6700\u540e\u90a3\u79cd\u5047\u8bbe\u56fa\u7136\u53ef\u4ee5\u63a5\u53d7\uff0c\u4f46\u5230\u6b27\u7f8e\u5404\u56fd\u8c03\u67e5\u4e4b\u540e\uff0c\u4e5f\u5c31\u96be\u4ee5\u81ea\u5706\u5176\u8bf4\u4e86\u3002\u6709\u54ea\u4e2a\u666e\u901a\u4eba\u4f1a\u62e5\u6709\u5982\u6b64\u5f3a\u5927\u52a8\u529b\u7684\u673a\u68b0\uff1f\u8fd9\u662f\u4e0d\u53ef\u80fd\u7684\u3002\u4ed6\u5728\u4f55\u5730\u4f55\u65f6\u53eb\u4f55\u4eba\u5236\u9020\u4e86\u8fd9\u4e48\u4e2a\u5e9e\u7136\u5927\u7269\uff0c\u800c\u4e14\u5982\u4f55\u80fd\u5728\u5efa\u9020\u4e2d\u505a\u5230\u98ce\u58f0\u4e0d\u8d70\u6f0f\u5462\uff1f \u770b\u6765\uff0c\u53ea\u6709\u653f\u5e9c\u624d\u6709\u53ef\u80fd\u62e5\u6709\u8fd9\u79cd\u7834\u574f\u6027\u7684\u673a\u5668\uff0c\u5728\u8fd9\u4e2a\u707e\u96be\u6df1\u91cd\u7684\u65f6\u4ee3\uff0c\u4eba\u4eec\u5343\u65b9\u767e\u8ba1\u8981\u589e\u5f3a\u6218\u4e89\u6b66\u5668\u5a01\u529b\uff0c\u90a3\u5c31\u6709\u8fd9\u79cd\u53ef\u80fd\uff0c\u4e00\u4e2a\u56fd\u5bb6\u7792\u7740\u5176\u4ed6\u56fd\u5bb6\u5728\u8bd5\u5236\u8fd9\u7c7b\u9a87\u4eba\u542c\u95fb\u7684\u6b66\u5668\u3002\u7ee7\u590f\u65af\u52c3\u6b65\u67aa\u4e4b\u540e\u6709\u6c34\u96f7\uff0c\u6c34\u96f7\u4e4b\u540e\u6709\u6c34\u4e0b\u649e\u9524\uff0c\u7136\u540e\u9b54\u9053\u6500\u5347\u53cd\u5e94\uff0c\u4e8b\u6001\u6108\u6f14\u6108\u70c8\u3002\u81f3\u5c11\uff0c\u6211\u662f\u8fd9\u6837\u60f3\u7684\u3002<\/p>\n<\/blockquote>\n<p>\u901a\u8fc7 SnowNLP \u63d0\u4f9b\u7684\u7b97\u6cd5\uff1a<\/p>\n<pre><code>from snownlp import SnowNLP   text = \" \u4e0a\u9762\u7684\u539f\u6587\u5185\u5bb9\uff0c\u6b64\u5904\u7701\u7565 \" s = SnowNLP(text) print(\"\u3002\".join(s.summary(5))) <\/code><\/pre>\n<p>\u8f93\u51fa\u7ed3\u679c\uff1a<\/p>\n<pre><code>\u81ea\u7136\u5c31\u5206\u6210\u89c2\u70b9\u622a\u7136\u4e0d\u540c\u7684\u4e24\u6d3e\uff1a\u4e00\u6d3e\u8bf4\u8fd9\u662f\u4e00\u4e2a\u529b\u5927\u65e0\u6bd4\u7684\u602a\u7269\u3002\u8fd9\u79cd\u5047\u8bbe\u4e5f\u4e0d\u80fd\u6210\u7acb\u3002\u6211\u5230\u7ebd\u7ea6\u65f6\u3002\u8bf4\u5b83\u662f\u4e00\u5757\u6d6e\u52a8\u7684\u8239\u4f53\u6216\u662f\u4e00\u5806\u5927\u8239\u6b8b\u7247\u3002\u53e6\u4e00\u6d3e\u8bf4\u8fd9\u662f\u4e00\u8258\u52a8\u529b\u6781\u5f3a\u7684\u201c\u6f5c\u6c34\u8239\u201d <\/code><\/pre>\n<p>\u521d\u6b65\u6765\u770b\uff0c\u6548\u679c\u5e76\u4e0d\u662f\u5f88\u597d\uff0c\u63a5\u4e0b\u6765\u6211\u4eec\u81ea\u5df1\u8ba1\u7b97\u53e5\u5b50\u6743\u91cd\uff0c\u5b9e\u73b0\u4e00\u4e2a\u7b80\u5355\u7684\u6458\u8981\u529f\u80fd\uff0c\u8fd9\u4e2a\u5c31\u9700\u8981<code>jieba<\/code>\uff1a<\/p>\n<pre><code>import re import jieba.analyse import jieba.posseg     class TextSummary:     def __init__(self, text):         self.text = text       def splitSentence(self):         sectionNum = 0         self.sentences = []         for eveSection in self.text.split(\"n\"):             if eveSection:                 sentenceNum = 0                 for eveSentence in re.split(\"!|\u3002|\uff1f\", eveSection):                     if eveSentence:                         mark = []                         if sectionNum == 0:                             mark.append(\"FIRSTSECTION\")                         if sentenceNum == 0:                             mark.append(\"FIRSTSENTENCE\")                         self.sentences.append({                             \"text\": eveSentence,                             \"pos\": {                                 \"x\": sectionNum,                                 \"y\": sentenceNum,                                 \"mark\": mark                             }                         })                         sentenceNum = sentenceNum + 1                 sectionNum = sectionNum + 1                 self.sentences[-1][\"pos\"][\"mark\"].append(\"LASTSENTENCE\")         for i in range(0, len(self.sentences)):             if self.sentences[i][\"pos\"][\"x\"] == self.sentences[-1][\"pos\"][\"x\"]:                 self.sentences[i][\"pos\"][\"mark\"].append(\"LASTSECTION\")       def getKeywords(self):         self.keywords = jieba.analyse.extract_tags(self.text, topK=20, withWeight=False, allowPOS=('n', 'vn', 'v'))       def sentenceWeight(self):         # \u8ba1\u7b97\u53e5\u5b50\u7684\u4f4d\u7f6e\u6743\u91cd         for sentence in self.sentences:             mark = sentence[\"pos\"][\"mark\"]             weightPos = 0             if \"FIRSTSECTION\" in mark:                 weightPos = weightPos + 2             if \"FIRSTSENTENCE\" in mark:                 weightPos = weightPos + 2             if \"LASTSENTENCE\" in mark:                 weightPos = weightPos + 1             if \"LASTSECTION\" in mark:                 weightPos = weightPos + 1             sentence[\"weightPos\"] = weightPos           # \u8ba1\u7b97\u53e5\u5b50\u7684\u7ebf\u7d22\u8bcd\u6743\u91cd         index = [\" \u603b\u4e4b \", \" \u603b\u800c\u8a00\u4e4b \"]         for sentence in self.sentences:             sentence[\"weightCueWords\"] = 0             sentence[\"weightKeywords\"] = 0         for i in index:             for sentence in self.sentences:                 if sentence[\"text\"].find(i) &gt;= 0:                     sentence[\"weightCueWords\"] = 1           for keyword in self.keywords:             for sentence in self.sentences:                 if sentence[\"text\"].find(keyword) &gt;= 0:                     sentence[\"weightKeywords\"] = sentence[\"weightKeywords\"] + 1           for sentence in self.sentences:             sentence[\"weight\"] = sentence[\"weightPos\"] + 2 * sentence[\"weightCueWords\"] + sentence[\"weightKeywords\"]       def getSummary(self, ratio=0.1):         self.keywords = list()         self.sentences = list()         self.summary = list()           # \u8c03\u7528\u65b9\u6cd5\uff0c\u5206\u522b\u8ba1\u7b97\u5173\u952e\u8bcd\u3001\u5206\u53e5\uff0c\u8ba1\u7b97\u6743\u91cd         self.getKeywords()         self.splitSentence()         self.sentenceWeight()           # \u5bf9\u53e5\u5b50\u7684\u6743\u91cd\u503c\u8fdb\u884c\u6392\u5e8f         self.sentences = sorted(self.sentences, key=lambda k: k['weight'], reverse=True)           # \u6839\u636e\u6392\u5e8f\u7ed3\u679c\uff0c\u53d6\u6392\u540d\u5360\u524d ratio% \u7684\u53e5\u5b50\u4f5c\u4e3a\u6458\u8981         for i in range(len(self.sentences)):             if i &lt; ratio * len(self.sentences):                 sentence = self.sentences[i]                 self.summary.append(sentence[\"text\"])           return self.summary   <\/code><\/pre>\n<p>\u8fd9\u6bb5\u4ee3\u7801\u4e3b\u8981\u662f\u901a\u8fc7<code>tf-idf<\/code>\u5b9e\u73b0\u5173\u952e\u8bcd\u63d0\u53d6\uff0c\u7136\u540e\u901a\u8fc7\u5173\u952e\u8bcd\u63d0\u53d6\u5bf9\u53e5\u5b50\u5c3d\u5fc3\u6743\u91cd\u8d4b\u4e88\uff0c\u6700\u540e\u83b7\u5f97\u5230\u6574\u4f53\u7684\u7ed3\u679c\uff0c\u8fd0\u884c\uff1a<\/p>\n<pre><code>testSummary = TextSummary(text) print(\"\u3002\".join(testSummary.getSummary())) <\/code><\/pre>\n<p>\u53ef\u4ee5\u5f97\u5230\u7ed3\u679c\uff1a<\/p>\n<pre><code>Building prefix dict from the default dictionary ... Loading model from cache \/var\/folders\/yb\/wvy_7wm91mzd7cjg4444gvdjsglgs8\/T\/jieba.cache Loading model cost 0.721 seconds. Prefix dict has been built successfully. \u770b\u6765\uff0c\u53ea\u6709\u653f\u5e9c\u624d\u6709\u53ef\u80fd\u62e5\u6709\u8fd9\u79cd\u7834\u574f\u6027\u7684\u673a\u5668\uff0c\u5728\u8fd9\u4e2a\u707e\u96be\u6df1\u91cd\u7684\u65f6\u4ee3\uff0c\u4eba\u4eec\u5343\u65b9\u767e\u8ba1\u8981\u589e\u5f3a\u6218\u4e89\u6b66\u5668\u5a01\u529b\uff0c\u90a3\u5c31\u6709\u8fd9\u79cd\u53ef\u80fd\uff0c\u4e00\u4e2a\u56fd\u5bb6\u7792\u7740\u5176\u4ed6\u56fd\u5bb6\u5728\u8bd5\u5236\u8fd9\u7c7b\u9a87\u4eba\u542c\u95fb\u7684\u6b66\u5668\u3002\u4e8e\u662f\uff0c\u6211\u5c31\u6293\u7d27\u8fd9\u6bb5\u5019\u8239\u9017\u7559\u65f6\u95f4\uff0c\u628a\u6536\u96c6\u5230\u7684\u77ff\u7269\u548c\u52a8\u690d\u7269\u6807\u672c\u8fdb\u884c\u5206\u7c7b\u6574\u7406\uff0c\u53ef\u5c31\u5728\u8fd9\u65f6\uff0c\u65af\u79d1\u820d\u53f7\u51fa\u4e8b\u4e86\u3002\u540c\u6837\u7684\u9053\u7406\uff0c\u8bf4\u5b83\u662f\u4e00\u5757\u6d6e\u52a8\u7684\u8239\u4f53\u6216\u662f\u4e00\u5806\u5927\u8239\u6b8b\u7247\uff0c\u8fd9\u79cd\u5047\u8bbe\u4e5f\u4e0d\u80fd\u6210\u7acb\uff0c\u7406\u7531\u4ecd\u7136\u662f\u79fb\u52a8\u901f\u5ea6\u592a\u5feb <\/code><\/pre>\n<p>\u6211\u4eec\u53ef\u4ee5\u770b\u5230\uff0c\u6574\u4f53\u6548\u679c\u8981\u6bd4\u521a\u624d\u7684\u597d\u4e00\u4e9b\u3002<\/p>\n<h2>\u53d1\u5e03 API<\/h2>\n<p>\u901a\u8fc7 Serverless \u67b6\u6784\uff0c\u5c06\u4e0a\u9762\u4ee3\u7801\u8fdb\u884c\u6574\u7406\uff0c\u5e76\u53d1\u5e03\u3002<\/p>\n<p>\u4ee3\u7801\u6574\u7406\u7ed3\u679c\uff1a<\/p>\n<pre><code>import re, json import jieba.analyse import jieba.posseg     class NLPAttr:     def __init__(self, text):         self.text = text       def splitSentence(self):         sectionNum = 0         self.sentences = []         for eveSection in self.text.split(\"n\"):             if eveSection:                 sentenceNum = 0                 for eveSentence in re.split(\"!|\u3002|\uff1f\", eveSection):                     if eveSentence:                         mark = []                         if sectionNum == 0:                             mark.append(\"FIRSTSECTION\")                         if sentenceNum == 0:                             mark.append(\"FIRSTSENTENCE\")                         self.sentences.append({                             \"text\": eveSentence,                             \"pos\": {                                 \"x\": sectionNum,                                 \"y\": sentenceNum,                                 \"mark\": mark                             }                         })                         sentenceNum = sentenceNum + 1                 sectionNum = sectionNum + 1                 self.sentences[-1][\"pos\"][\"mark\"].append(\"LASTSENTENCE\")         for i in range(0, len(self.sentences)):             if self.sentences[i][\"pos\"][\"x\"] == self.sentences[-1][\"pos\"][\"x\"]:                 self.sentences[i][\"pos\"][\"mark\"].append(\"LASTSECTION\")       def getKeywords(self):         self.keywords = jieba.analyse.extract_tags(self.text, topK=20, withWeight=False, allowPOS=('n', 'vn', 'v'))         return self.keywords       def sentenceWeight(self):         # \u8ba1\u7b97\u53e5\u5b50\u7684\u4f4d\u7f6e\u6743\u91cd         for sentence in self.sentences:             mark = sentence[\"pos\"][\"mark\"]             weightPos = 0             if \"FIRSTSECTION\" in mark:                 weightPos = weightPos + 2             if \"FIRSTSENTENCE\" in mark:                 weightPos = weightPos + 2             if \"LASTSENTENCE\" in mark:                 weightPos = weightPos + 1             if \"LASTSECTION\" in mark:                 weightPos = weightPos + 1             sentence[\"weightPos\"] = weightPos           # \u8ba1\u7b97\u53e5\u5b50\u7684\u7ebf\u7d22\u8bcd\u6743\u91cd         index = [\" \u603b\u4e4b \", \" \u603b\u800c\u8a00\u4e4b \"]         for sentence in self.sentences:             sentence[\"weightCueWords\"] = 0             sentence[\"weightKeywords\"] = 0         for i in index:             for sentence in self.sentences:                 if sentence[\"text\"].find(i) &gt;= 0:                     sentence[\"weightCueWords\"] = 1           for keyword in self.keywords:             for sentence in self.sentences:                 if sentence[\"text\"].find(keyword) &gt;= 0:                     sentence[\"weightKeywords\"] = sentence[\"weightKeywords\"] + 1           for sentence in self.sentences:             sentence[\"weight\"] = sentence[\"weightPos\"] + 2 * sentence[\"weightCueWords\"] + sentence[\"weightKeywords\"]       def getSummary(self, ratio=0.1):         self.keywords = list()         self.sentences = list()         self.summary = list()           # \u8c03\u7528\u65b9\u6cd5\uff0c\u5206\u522b\u8ba1\u7b97\u5173\u952e\u8bcd\u3001\u5206\u53e5\uff0c\u8ba1\u7b97\u6743\u91cd         self.getKeywords()         self.splitSentence()         self.sentenceWeight()           # \u5bf9\u53e5\u5b50\u7684\u6743\u91cd\u503c\u8fdb\u884c\u6392\u5e8f         self.sentences = sorted(self.sentences, key=lambda k: k['weight'], reverse=True)           # \u6839\u636e\u6392\u5e8f\u7ed3\u679c\uff0c\u53d6\u6392\u540d\u5360\u524d ratio% \u7684\u53e5\u5b50\u4f5c\u4e3a\u6458\u8981         for i in range(len(self.sentences)):             if i &lt; ratio * len(self.sentences):                 sentence = self.sentences[i]                 self.summary.append(sentence[\"text\"])           return self.summary     def main_handler(event, context):     nlp = NLPAttr(json.loads(event['body'])['text'])     return {         \"keywords\": nlp.getKeywords(),         \"summary\": \"\u3002\".join(nlp.getSummary())     } <\/code><\/pre>\n<p>\u7f16\u5199\u9879\u76ee<code>serverless.yaml<\/code>\u6587\u4ef6\uff1a<\/p>\n<pre><code>nlpDemo:   component: \"@serverless\/tencent-scf\"   inputs:     name: nlpDemo     codeUri: .\/     handler: index.main_handler     runtime: Python3.6     region: ap-guangzhou     description: \u6587\u672c\u6458\u8981 \/ \u5173\u952e\u8bcd\u529f\u80fd     memorySize: 256     timeout: 10     events:       - apigw:           name: nlpDemo_apigw_service           parameters:             protocols:               - http             serviceName: serverless             description: \u6587\u672c\u6458\u8981 \/ \u5173\u952e\u8bcd\u529f\u80fd             environment: release             endpoints:               - path: \/nlp                 method: ANY <\/code><\/pre>\n<p>\u7531\u4e8e\u9879\u76ee\u4e2d\u4f7f\u7528\u4e86<code>jieba<\/code>\uff0c\u6240\u4ee5\u5728\u5b89\u88c5\u7684\u65f6\u5019\u63a8\u8350\u5728 CentOS \u7cfb\u7edf\u4e0b\u4e0e\u5bf9\u5e94\u7684 Python \u7248\u672c\u4e0b\u5b89\u88c5\uff0c\u4e5f\u53ef\u4ee5\u4f7f\u7528\u6211\u4e4b\u524d\u4e3a\u4e86\u65b9\u4fbf\u505a\u7684\u4e00\u4e2a\u4f9d\u8d56\u5de5\u5177\uff1a<\/p>\n<p><img decoding=\"async\" src=\"http:\/\/4563.org\/wp-content\/uploads\/2020\/05\/20200516_5ebf8f8ed3b64.png\" alt=\"Serverless \u5b9e\u6218\uff1a\u5982\u4f55\u7ed3\u5408 NLP \u5b9e\u73b0\u6587\u672c\u6458\u8981\u548c\u5173\u952e\u8bcd\u63d0\u53d6\uff1f\" \/><\/p>\n<p>\u901a\u8fc7<code>sls --debug<\/code>\u8fdb\u884c\u90e8\u7f72\uff1a<\/p>\n<p><img decoding=\"async\" src=\"http:\/\/4563.org\/wp-content\/uploads\/2020\/05\/20200516_5ebf8f9953696.png\" alt=\"Serverless \u5b9e\u6218\uff1a\u5982\u4f55\u7ed3\u5408 NLP \u5b9e\u73b0\u6587\u672c\u6458\u8981\u548c\u5173\u952e\u8bcd\u63d0\u53d6\uff1f\" \/><\/p>\n<p>\u90e8\u7f72\u5b8c\u6210\uff0c\u53ef\u4ee5\u901a\u8fc7 PostMan \u8fdb\u884c\u7b80\u5355\u7684\u6d4b\u8bd5\uff1a<\/p>\n<p><img decoding=\"async\" src=\"http:\/\/4563.org\/wp-content\/uploads\/2020\/05\/20200516_5ebf8f9f4e48b.png\" alt=\"Serverless \u5b9e\u6218\uff1a\u5982\u4f55\u7ed3\u5408 NLP \u5b9e\u73b0\u6587\u672c\u6458\u8981\u548c\u5173\u952e\u8bcd\u63d0\u53d6\uff1f\" \/><\/p>\n<p>\u4ece\u4e0a\u56fe\u53ef\u4ee5\u770b\u5230\uff0c\u6211\u4eec\u5df2\u7ecf\u6309\u7167\u9884\u671f\u8f93\u51fa\u4e86\u76ee\u6807\u7ed3\u679c\u3002\u81f3\u6b64\uff0c\u6587\u672c\u6458\u8981 \/ \u5173\u952e\u8bcd\u63d0\u53d6\u7684 API \u5df2\u7ecf\u90e8\u7f72\u5b8c\u6210\u3002<\/p>\n<h2>\u603b\u7ed3<\/h2>\n<p>\u76f8\u5bf9\u6765\u8bf4\uff0c\u901a\u8fc7 Serveless \u67b6\u6784\u505a API \u662f\u975e\u5e38\u5bb9\u6613\u548c\u65b9\u4fbf\u7684\uff0c\u53ef\u5b9e\u73b0 API \u7684\u63d2\u62d4\u884c\uff0c\u7ec4\u4ef6\u5316\uff0c\u5e0c\u671b\u672c\u6587\u80fd\u591f\u7ed9\u8bfb\u8005\u66f4\u591a\u7684\u601d\u8def\u548c\u542f\u53d1\u3002<\/p>\n<p>\u6b22\u8fce\u8bbf\u95ee\uff1aServerless \u4e2d\u6587\u793e\u533a<\/p>\n<\/p><\/div>\n<div> <b>\u5927\u4f6c\u6709\u8a71\u8aaa<\/b> (<span>0<\/span>)        <\/div>\n<div> <\/div>\n<\/p><\/div>\n<\/p><\/div>\n<ul>\n<li>\n","protected":false},"excerpt":{"rendered":"<p>Serverless \u5b9e\u6218\uff1a\u5982\u4f55\u7ed3&hellip;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[],"tags":[],"_links":{"self":[{"href":"http:\/\/4563.org\/index.php?rest_route=\/wp\/v2\/posts\/84614"}],"collection":[{"href":"http:\/\/4563.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/4563.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/4563.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/4563.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=84614"}],"version-history":[{"count":0,"href":"http:\/\/4563.org\/index.php?rest_route=\/wp\/v2\/posts\/84614\/revisions"}],"wp:attachment":[{"href":"http:\/\/4563.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=84614"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/4563.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=84614"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/4563.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=84614"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}