public static List <String> AnlayzeNER(string xmlfilename)
{
//由于结果是多个XML构成的
//1.掉所有的<?xml version="1.0" encoding="utf-8" ?>
//2.加入<sentence></sentence> root节点
var NerList = new List <String>();
if (!File.Exists(xmlfilename))
{
return(NerList);
}
var sr = new StreamReader(xmlfilename);
List <struWordNER> wl = null;
var pl = new List <List <struWordNER> >();
var ner = String.Empty;
while (!sr.EndOfStream)
{
var line = sr.ReadLine().Trim();
if (line.StartsWith("<sent"))
{
if (wl != null)
{
pl.Add(wl);
}
//一个新的句子
wl = new List <struWordNER>();
}
if (line.StartsWith("<word"))
{
var word = new struWordNER(line);
wl.Add(word);
switch (word.ne)
{
case "B-Ni":
ner = word.cont;
break;
case "I-Ni":
ner += word.cont;
break;
case "E-Ni":
ner += word.cont;
NerList.Add(ner);
break;
}
}
}
if (wl != null)
{
pl.Add(wl);
}
sr.Close();
return(NerList);
}
public static List <List <struWordNER> > GetParagraghList(string xmlfilename)
{
var paragrapghList = new List <List <struWordNER> >();
//由于结果是多个XML构成的
//1.掉所有的<?xml version="1.0" encoding="utf-8" ?>
//2.加入<sentence></sentence> root节点
if (!File.Exists(xmlfilename))
{
return(paragrapghList);
}
List <struWordNER> wordList = null;
var sr = new StreamReader(xmlfilename);
while (!sr.EndOfStream)
{
var line = sr.ReadLine().Trim();
if (line.StartsWith("<sent"))
{
if (wordList != null)
{
paragrapghList.Add(wordList);
}
//一个新的句子
wordList = new List <struWordNER>();
}
if (line.StartsWith("<word"))
{
var word = new struWordNER(line);
wordList.Add(word);
}
}
if (wordList != null)
{
paragrapghList.Add(wordList);
}
sr.Close();
return(paragrapghList);
}
/// <summary>
/// 使用NerInfo抽取
/// </summary>
public static void GetCompanyNameByNerInfo(List <List <struWordNER> > paragragh)
{
var Rule1 = new NerExtract.NerExtractRule();
Rule1.MaxWordLength = 10;
Rule1.StartWord = new List <struWordNER>();
//首词NER属性
var word = new struWordNER();
word.pos = LTPTrainingNER.地名; //只设定类型
Rule1.StartWord.Add(word);
//结束词NER属性
Rule1.EndWord = new List <struWordNER>();
word = new struWordNER();
word.cont = "有限公司"; //只设定词语
Rule1.EndWord.Add(word);
var Rule2 = new NerExtract.NerExtractRule();
Rule2.MaxWordLength = 10;
Rule2.StartWord = new List <struWordNER>();
//首词NER属性
word = new struWordNER();
word.pos = LTPTrainingNER.地名; //只设定类型
Rule2.StartWord.Add(word);
//结束词NER属性
Rule2.EndWord = new List <struWordNER>();
word = new struWordNER();
word.cont = "有限"; //只设定词语
Rule2.EndWord.Add(word);
word = new struWordNER();
word.cont = "责任"; //只设定词语
Rule2.EndWord.Add(word);
word = new struWordNER();
word.cont = "公司"; //只设定词语
Rule2.EndWord.Add(word);
var Rule3 = new NerExtract.NerExtractRule();
Rule3.MaxWordLength = 10;
Rule3.StartWord = new List <struWordNER>();
//首词NER属性
word = new struWordNER();
word.pos = LTPTrainingNER.地名; //只设定类型
Rule3.StartWord.Add(word);
//结束词NER属性
Rule3.EndWord = new List <struWordNER>();
word = new struWordNER();
word.cont = "（"; //只设定词语
Rule3.EndWord.Add(word);
word = new struWordNER();
word.cont = "有限"; //只设定词语
Rule3.EndWord.Add(word);
word = new struWordNER();
word.cont = "合伙"; //只设定词语
Rule3.EndWord.Add(word);
word = new struWordNER();
word.cont = "）"; //只设定词语
Rule3.EndWord.Add(word);
var company1 = NerExtract.Extract(Rule1, paragragh);
var company2 = NerExtract.Extract(Rule2, paragragh);
var company3 = NerExtract.Extract(Rule3, paragragh);
}
public static List <String> GetProjectNameByNer(AnnouceDocument doc)
{
//由于结果是多个XML构成的
//1.掉所有的<?xml version="1.0" encoding="utf-8" ?>
//2.加入<sentence></sentence> root节点
var ProjList = new List <String>();
if (!File.Exists(doc.NerXMLFileName))
{
return(ProjList);
}
var sr = new StreamReader(doc.NerXMLFileName);
List <struWordNER> wl = null;
var pl = new List <List <struWordNER> >();
var ner = String.Empty;
while (!sr.EndOfStream)
{
var line = sr.ReadLine().Trim();
if (line.StartsWith("<sent"))
{
if (wl != null)
{
pl.Add(wl);
}
//一个新的句子
wl = new List <struWordNER>();
}
if (line.StartsWith("<word"))
{
var word = new struWordNER(line);
wl.Add(word);
}
}
if (wl != null)
{
pl.Add(wl);
}
sr.Close();
var proj = String.Empty;
foreach (var p in pl)
{
for (int ScanIdx = 0; ScanIdx < p.Count; ScanIdx++)
{
var word = p[ScanIdx];
if (word.ne == "B-Ns" || word.ne == "S-Ns" ||
word.cont == "新建")
{
//遇到地名开始或者单独地名，加入到项目字符中
if (!string.IsNullOrEmpty(proj) && proj.StartsWith("新建"))
{
proj += word.cont;
}
else
{
proj = word.cont;
}
}
else
{
if (word.cont.Equals("项目") ||
word.cont.Equals("工程") ||
word.cont.Equals("标段") ||
word.cont.Equals("采购"))
{
if (!String.IsNullOrEmpty(proj))
{
proj += word.cont;
var FurtherTo = Math.Min(p.Count, ScanIdx + 5);
var ShardProj = proj;
//标段的后检索
if (word.cont == "标段")
{
//检查之后3个词汇的距离是否存在项目，工程，承包
for (int TrailingScanIdx = ScanIdx + 1;
TrailingScanIdx < FurtherTo;
TrailingScanIdx++)
{
ShardProj += p[TrailingScanIdx].cont;
if (p[TrailingScanIdx].cont == "项目" ||
p[TrailingScanIdx].cont == "工程" ||
p[TrailingScanIdx].cont == "承包")
{
proj = ShardProj;
break;
}
}
}
//工程
if (word.cont == "工程" || word.cont == "项目")
{
//检查之后3个词汇的距离是否存在项目，工程，承包
var isContranBrack = false;
for (int TrailingScanIdx = ScanIdx + 1; TrailingScanIdx < FurtherTo;
TrailingScanIdx++)
{
ShardProj += p[TrailingScanIdx].cont;
if (p[TrailingScanIdx].cont.Trim() == "（")
{
isContranBrack = true;
}
if (p[TrailingScanIdx].cont.Trim() == "）")
{
isContranBrack = false;
}
if (p[TrailingScanIdx].cont == "标段")
{
ScanIdx = TrailingScanIdx;
if (isContranBrack)
{
ShardProj += "）";
ScanIdx++;
}
proj = ShardProj;
break;
}
}
}
//整体的再检查，是否下面一个单词还是工程，项目，标段
if (ScanIdx + 1 <= p.Count - 1)
{
if (p[ScanIdx + 1].cont == "工程" || p[ScanIdx + 1].cont == "项目" ||
p[ScanIdx + 1].cont == "标段" || p[ScanIdx + 1].cont == "活动")
{
proj += p[ScanIdx + 1].cont;
ScanIdx++;
}
}
var isOK = true;
if (proj.Contains("重大工程"))
{
isOK = false;
}
if (proj.Length > 50)
{
isOK = false;
}
if (proj.Contains("；"))
{
isOK = false;
}
if (proj.Contains(""))
{
isOK = false;
}
if (isOK)
{
Console.WriteLine(doc.Id + " NER 发现工程：" + proj);
ProjList.Add(proj);
}
proj = string.Empty;
}
}
else
{
if (!String.IsNullOrEmpty(proj))
{
proj += word.cont;
}
}
}
}
}
return(ProjList.Distinct().ToList());
}
