1st homework

class1/week1/homework/houlisha/data.txt

@@ -0,0 +1,30 @@
+刘三 Liu, San
++86 15912348765
+[email protected]
+--------------------------
+李四
+北清大数据产业联合会
+电话：010-34355675
+邮箱：[email protected]
+地址：北京市海淀区北清大学东楼201室
+--------------------------
+John Smith
+Data and Web Science Group
+University of Mannheim, Germany
+
+http://dws.informatik.uni-mannheim.de/~johnsmith
+Tel: +49 621 123 4567
+--------------------------
+王五
+CSDN-全球最大中文IT技术社区（www.csdn.net）
+电话:010-51661202-257
+手机:13934567890
+E-mail:[email protected]
+QQ、微信：34534563
+地址：北京市朝阳区广顺北大街33号院一号楼福码大厦B座12层
+--------------------------
+张三
+北京市张三律师事务所|Beijing Zhangsan Law Firm
+北京市海淀区中关村有条街1号，邮编：100080
+No. 1 Youtiao Street , ZhongGuanCun West, Haidian District, Beijing 100080
+Mobile: 15023345465|Email: [email protected]

class1/week1/homework/houlisha/week1.py

@@ -0,0 +1,91 @@
+#-*- coding:utf-8 -*-
+
+__author__='houlisha'
+
+import os
+import re
+import codecs
+import jieba
+import jieba.posseg as pseg
+from nltk.tag import StanfordNERTagger
+
+os.environ['CLASSPATH'] = '/Users/maidoudou/Library/stanfordNLPLibrary/stanford-ner-2015-12-09'
+os.environ['STANFORD_MODELS'] = '/Users/maidoudou/Library/stanfordNLPLibrary/stanford-ner-2015-12-09/classifiers'
+
+stch = StanfordNERTagger('chinese.misc.distsim.crf.ser.gz')
+sten = StanfordNERTagger('english.all.3class.distsim.crf.ser.gz')
+
+divider_pattern = re.compile('--*')
+
+tel_pattern = re.compile(r'(\+\d{2})?[ -]?\d{3,4}[ -]?\d{3,4}[ -]?\d{4,5}[-]?\d*')
+def tel_match(data):
+    iterator = tel_pattern.finditer(data)
+    tels = [i.group() for i in iterator if iterator != None]
+    return tels
+
+mail_pattern = re.compile(r'\w+([-+.]\w+)*@\w+([-.]\w+)*\.\w+([-.]\w+)*')
+def mail_match(data):
+    iterator = mail_pattern.finditer(data)
+    mails = [i.group() for i in iterator if iterator != None]
+    return mails
+
+# 中文姓名用jieba词性标注匹配
+# 英文名字用nltk实体命名匹配
+def name_match(data):
+    ch_name = [w.word for w in pseg.cut(data) if w.flag == 'nr']
+    en_name = [w[0] for w in sten.tag(data.split()) if w[1] == 'PERSON']
+    return ch_name + en_name
+
+# nltk效果不太好
+def com_match(data):
+    coms = []
+    items = data.split('\n')
+    for item in items:
+        words = list(jieba.cut(item))
+	ch_tags = stch.tag(words)
+        ch_com = [w[0] for w in ch_tags if w[1] == 'ORG']
+	if ch_com == []:
+	    en_com = [w[0] for w in sten.tag(item.split()) if w[1] == 'ORGANIZATION']
+	    if en_com == []:
+	        continue
+	coms.append(item)
+    return coms
+
+def signature_extract(datas):
+    signatures = []
+    for data in datas:
+        signature = {}
+        signature['name'] = name_match(data)
+	signature['tel'] = tel_match(data)
+	signature['email'] = mail_match(data)
+	signature['com'] = com_match(data)
+
+        signatures.append(signature)
+    return signatures
+
+def main():
+    datafile = codecs.open('data.txt')
+    datas = []
+    data = ''
+    line = datafile.readline()
+    while line:
+        if divider_pattern.match(line) != None: #another signature
+	    datas.append(data)
+	    data = ''
+	else:
+	    data = data + "\n" + line.strip()
+	line = datafile.readline()
+    if data != '':
+        datas.append(data)
+
+    signatures = signature_extract(datas)
+
+    for signature in signatures:
+        print '---------------------'
+        print '姓名:' , " ".join(set(signature['name']))
+        print '单位:' , " ".join(signature['com'])
+        print '电话号码:' , signature['tel']
+        print '电子邮箱:' , signature['email']
+
+if __name__=='__main__':
+    main()