xiongchenlong
/
compose-analysis
forked from liuzongren/compose-analysis
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

1.优化程序分析逻辑，修改文件级别分析逻辑

Browse Source
master
liuzongren 7 months ago
parent 09a0ffa44f
commit e62e8824fa
5 changed files with 53 additions and 191 deletions
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Show Stats Download Patch File Download Diff File
  1. 41
      src/main/java/com/keyware/composeanalysis/task/CodeBlockAnalysisTask.java
  2. 104
      src/main/java/com/keyware/composeanalysis/task/FileAnalysisTask.java
  3. 42
      src/main/java/com/keyware/composeanalysis/task/FunctionAnalysisTask.java
  4. 43
      src/main/java/com/keyware/composeanalysis/task/LineAnalysisTask.java
  5. 14
      src/main/resources/application.yaml

41
src/main/java/com/keyware/composeanalysis/task/CodeBlockAnalysisTask.java
Unescape View File

@ -168,8 +168,6 @@ public class CodeBlockAnalysisTask extends IAnalysisTask {
private void doAnalysis(SolrDocumentList matcheOpenSourceFiles, String sourceFileBaseCoreName, CodeFile fileAnalysisRes) {
if (CollectionUtil.isEmpty(matcheOpenSourceFiles)) {
//因为代码块的特征库较少，这里补充一个对比逻辑,如果当前文件解析失败，或者没有通过代码块匹配到数据，则直接通过文件的md5 再次查询一次solr库
checkByOriginalFileMd5(sourceFileBaseCoreName, analysisFile.getMd5());
return;
}
@ -313,45 +311,6 @@ public class CodeBlockAnalysisTask extends IAnalysisTask {
}
/**
* 防止代码块特征库不全，再次根据文件MD5查询开源文件信息, 做二次校验
*
* @param originalFileMd5
* @param versionIdCoreName
*/
private void checkByOriginalFileMd5(String versionIdCoreName, String originalFileMd5) {
//根据文件的MD5,查询特征库，看当前文件是否在开源代码库中
SolrDocument versionIdAndPath = solrUtils.queryOne(versionIdCoreName, "sourceFileMd5:" + originalFileMd5, "versionId,fullPath,sourceFileMd5");
if (versionIdAndPath != null) {
//根据版本ID查询版本的详细信息
VersionTree versionInfo = solrUtils.queryVersionInfoByVersionId((String) versionIdAndPath.get("versionId"));
if (versionInfo != null) {
//当前开源文件的开源项目信息
MatchOpenFile matchOpenFileInfo = new MatchOpenFile();
matchOpenFileInfo.setPId(versionInfo.getProId())
.setPName(versionInfo.getProName())
.setSourceUrl(versionInfo.getDownUrl())
.setFeatureSimilarity(100.00f)
.setOpenRate(100.00f)
.setAnalyzeType(AnalysisLevelEnum.FILE_LEVEL.getCode());
//保存当前文件的开源信息到mongo库中
MatchOpenFileMongoDto matchOpenFileMongo = new MatchOpenFileMongoDto();
matchOpenFileMongo.setId(IdGenerator.uuid32())
.setFilePath(analysisFile.getFileUrl())
.setFileName(analysisFile.getName())
.setOpenRate(100.00f)
.setOpenType(analysisFile.getOpenType())
.setMatchOpenFile(Arrays.asList(matchOpenFileInfo));
mongoTemplate.save(matchOpenFileMongo);
}
}
}
/**
* 获取当前文件的代码块特征值
*

104
src/main/java/com/keyware/composeanalysis/task/FileAnalysisTask.java
Unescape View File

@ -4,7 +4,6 @@ import cn.hutool.core.lang.Pair;
import com.keyware.common.constant.enums.AnalysisStatusEnum;
import com.keyware.composeanalysis.constant.FixedValue;
import com.keyware.composeanalysis.constant.RedisConst;
import com.keyware.composeanalysis.constant.SolrDBConst;
import com.keyware.composeanalysis.constant.enums.AnalysisLevelEnum;
import com.keyware.composeanalysis.constant.enums.FileAnalysisStatusEnum;
import com.keyware.composeanalysis.entity.AnalysisTask;
@ -18,8 +17,6 @@ import com.keyware.keyswan.anaysis.AnalysisFactory;
import com.keyware.keyswan.common.CodeFile;
import com.keyware.utils.IdGenerator;
import lombok.extern.log4j.Log4j2;
import org.apache.commons.collections.CollectionUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.solr.common.SolrDocument;
import org.apache.solr.common.SolrDocumentList;
import org.springframework.data.mongodb.core.MongoTemplate;
@ -81,42 +78,31 @@ public class FileAnalysisTask extends IAnalysisTask {
//获取当前文件名称
String fileName = analysisFile.getName();
AnalysisLogUtil.insert(mongoTemplate, "【文件级分析】正在分析" + fileName);
try {
//只有主流语言的才能解析
//非32种主流语言的不能提取文件特征,在文件级MD5匹配的时候,已经做过匹配
if (StringUtils.isNotEmpty(analysisFile.getSuffix()) && FixedValue.SUFFIX_SOLR_VERSION.containsKey(analysisFile.getSuffix())) {
//根据文件后缀 查询 *_CutFileInfo库名称
String featureCoreName = FixedValue.SUFFIX_SOLR_FILE.get(analysisFile.getSuffix());
//根据文件名称，获取文件解析器
Analysis analysis = AnalysisFactory.getAnalysis(fileName);
//如果 analysis 返回值为null 说明还未支持这种语言的特征提取 可以直接通过文件的MD5值去solr库中匹配
if (analysis != null) {
//如果文件大小超过3M,则不进行文件级行级特征提取
Integer fileSize = analysisFile.getFileSize();
if (fileSize < (3 * 1024 * 1024)) {
CodeFile codeFile = analysis.analysisFile(analysisFile.getFileUrl(), "1", "0");
//根据文件的特征值,去相应文件文件后缀的特征库中进行查询
if (codeFile != null) {
String querySb = "sourceMd5:" + codeFile.getSourceMd5() + " OR cutFileMd5:" + codeFile.getCutFileMd5() + " OR traitFileMd5:" + codeFile.getTraitFileMd5();
SolrDocumentList openSourceFileList = solrUtils.query(featureCoreName, querySb, "sourceMd5");
//如果当前文件在源码库中，匹配到了数据,则统计当前文件的开源率
if (CollectionUtils.isNotEmpty(openSourceFileList)) {
ananlyzeFileOpenRate(openSourceFileList,codeFile);
}
}
}
}
}
//根据文件后缀 查询 *_CutFileInfo库名称
String featureCoreName = FixedValue.SUFFIX_SOLR_FILE.get(analysisFile.getSuffix());
//根据文件名称，获取文件解析器
Analysis analysis = AnalysisFactory.getAnalysis(fileName);
//如果文件大小超过3M,则不进行文件级行级特征提取
CodeFile codeFile = analysis.analysisFile(analysisFile.getFileUrl(), "1", "0");
//根据文件的特征值,去相应文件文件后缀的特征库中进行查询
SolrDocumentList openSourceFileList = getFeatureSimilarityFromSolr(featureCoreName, codeFile);
//则统计当前文件的开源率
ananlyzeFileOpenRate(openSourceFileList, codeFile);
//更新文件级分析结果
analysisFile.setFileAnalysisStatus(FileAnalysisStatusEnum.ANALYSIS_DONE.getCode());
mongoTemplate.update(FileDataMongoDto.class)
.matching(where("_id").is(analysisFile.getId()))
.replaceWith(analysisFile)
.findAndReplace();
AnalysisLogUtil.insert(mongoTemplate, "【文件级分析】成功" + fileName);
} catch (Exception e) {
AnalysisLogUtil.insertErrorInfo(mongoTemplate, "【文件级】提取失败" + fileName, e);
log.error("文件:" + fileName + "文件级别特征提取失败!", e);
AnalysisLogUtil.insertErrorInfo(mongoTemplate, "【文件级】分析失败" + fileName, e);
log.error("文件:" + fileName + "文件级别分析失败!", e);
//将当前文件的分析状态变更为失败
analysisFile.setFileAnalysisStatus(FileAnalysisStatusEnum.FAILED_ANALYSIS.getCode());
//更新文件级分析结果
@ -130,40 +116,39 @@ public class FileAnalysisTask extends IAnalysisTask {
}
/**
* 根据 特征值 从特征库中检索 具有特征相似的
*
* @param featureCoreName 检索的solr 库名称
* @param codeFile 源文件解析结果
* @return
*/
private SolrDocumentList getFeatureSimilarityFromSolr(String featureCoreName, CodeFile codeFile) {
if (codeFile == null) {
log.error("特征为空,无法查询:{}", analysisFile.getName());
return new SolrDocumentList();
}
String queryStr = "sourceMd5:" + codeFile.getSourceMd5() + " OR cutFileMd5:" + codeFile.getCutFileMd5() + " OR traitFileMd5:" + codeFile.getTraitFileMd5();
return solrUtils.query(featureCoreName, queryStr, "sourceMd5");
}
/**
* 分析文件的开源率
*
* @param openSourceFileList 匹配的开源文件信息
* @throws IOException
*/
private void ananlyzeFileOpenRate(SolrDocumentList openSourceFileList,CodeFile fileAnalysisRes) throws IOException {
//根据匹配的开源文件的md5 获取版本ID
Set<String> sourceFileMd5 = openSourceFileList.stream().map(solrDocument -> (String) solrDocument.get("sourceMd5")).collect(Collectors.toSet());
String sourceCoreName = FixedValue.SUFFIX_SOLR_VERSION.get(analysisFile.getSuffix());
Map<String, SolrDocument> md5VersionObjMap = solrUtils.batchQueryVersionIdFromSourceFileBaseBySourceMd5(sourceCoreName, sourceFileMd5);
//根据版本ID获取版本信息
Set<String> versionIds = md5VersionObjMap.values().stream().map(solrDocument -> (String) solrDocument.get("versionId")).collect(Collectors.toSet());
List<VersionTree> treeInfoList = solrUtils.queryBatchVersionInfoByVersionIds(versionIds);
Map<String, VersionTree> versionIdMap = treeInfoList.stream().collect(Collectors.toMap(VersionTree::getVersionId, Function.identity()));
//获取被测件文本内容
String fileContent = new String(Files.readAllBytes(Paths.get(analysisFile.getFileUrl())), "utf-8").replaceAll(" ", "");
//将被测件的文本内容拆分成行信息，用于匹配开源信息
List<String> fileLines = SimilarityUtil.getSplitWords(fileContent);
private void ananlyzeFileOpenRate(SolrDocumentList openSourceFileList, CodeFile fileAnalysisRes) {
HashSet<Integer> openLineNum = new HashSet<>();
List<MatchOpenFile> matchOpenFilesRes = calculateOpenRate(openSourceFileList,fileAnalysisRes,sourceCoreName,openLineNum);
//统计被测件的总体开源率
//计算每个文件的开源率和特征相似度
List<MatchOpenFile> matchOpenFilesRes = calculateOpenRate(openSourceFileList, fileAnalysisRes, openLineNum);
//获取开源率阈值，判断当前文件是否开源
Integer openRateThreshold = analysisTask.getOpenRateThreshold();
BigDecimal openRate = new BigDecimal(openLineNum.size()).divide(new BigDecimal(fileLines.size()), 4, RoundingMode.HALF_UP).multiply(new BigDecimal(100)).setScale(2);
BigDecimal openRate = new BigDecimal(openLineNum.size()).divide(fileAnalysisRes.getCodeRowNum(), 4, RoundingMode.HALF_UP).multiply(new BigDecimal(100)).setScale(2);
//超过阈值，则认为当前文件是开源文件
if (openRate.compareTo(new BigDecimal(openRateThreshold)) > 0) {
@ -192,20 +177,21 @@ public class FileAnalysisTask extends IAnalysisTask {
}
/**
* 计算当前文件的特征相似度 和 开源率
*
* @param matchOpenFiles 通过MD5 匹配到的所有开源文件
* @param sourceFileBaseCoreName 当前文件特征文件的 solr coreName
* @param matchLineRowsNum 所有开源文件匹配到的开源行号列表
* @param matchOpenFiles 通过MD5 匹配到的所有开源文件
* @param matchLineRowsNum 所有开源文件匹配到的开源行号列表
* @return 匹配的开源文件解析后的结果集
*/
private List<MatchOpenFile> calculateOpenRate(SolrDocumentList matchOpenFiles, CodeFile fileAnalysisRes, String sourceFileBaseCoreName, Set<Integer> matchLineRowsNum) {
private List<MatchOpenFile> calculateOpenRate(SolrDocumentList matchOpenFiles, CodeFile fileAnalysisRes, Set<Integer> matchLineRowsNum) {
//匹配的开源文件列表
List<MatchOpenFile> matchOpenFilesRes = new ArrayList<>();
//根据匹配的开源文件的md5 获取版本ID
String sourceFileBaseCoreName = FixedValue.SUFFIX_SOLR_VERSION.get(analysisFile.getSuffix());
//首先根据文件的MD5查询开源文件的版本ID,和路径信息
Set<String> openSourceFileMd5s = matchOpenFiles.stream().map(doc -> (String) doc.get("sourceMd5")).collect(Collectors.toSet());
Map<String, SolrDocument> md5VersionInfoMap = solrUtils.batchQueryVersionIdFromSourceFileBaseBySourceMd5(sourceFileBaseCoreName, openSourceFileMd5s);
@ -230,7 +216,7 @@ public class FileAnalysisTask extends IAnalysisTask {
SolrDocument openEntries = md5VersionInfoMap.get(openSourceFileMd5);
VersionTree versionInfo = versionIdVersionInfoMap.get(openEntries.get("versionId"));
if (versionInfo == null){
if (versionInfo == null) {
log.error("找不到开源文件版本信息,versionId:{}", openEntries.get("versionId"));
}
@ -250,6 +236,4 @@ public class FileAnalysisTask extends IAnalysisTask {
}
}

42
src/main/java/com/keyware/composeanalysis/task/FunctionAnalysisTask.java
Unescape View File

@ -145,8 +145,6 @@ public class FunctionAnalysisTask extends IAnalysisTask {
private void doAnalysis(SolrDocumentList matchOpenFiles, String sourceFileBaseCoreName, CodeFile fileAnalysisRes) {
if (CollectionUtil.isEmpty(matchOpenFiles)) {
//因为函数的特征库较少，这里补充一个对比逻辑,如果当前文件解析失败，或者没有通过函数匹配到数据，则直接通过文件的md5 再次查询一次solr库
checkByOriginalFileMd5(sourceFileBaseCoreName, analysisFile.getMd5());
return;
}
@ -290,46 +288,6 @@ public class FunctionAnalysisTask extends IAnalysisTask {
return matchOpenFilesRes;
}
/**
* 防止函数特征库不全，再次根据文件MD5查询开源文件信息, 做二次校验
*
* @param originalFileMd5
* @param versionIdCoreName
*/
private void checkByOriginalFileMd5(String versionIdCoreName, String originalFileMd5) {
//根据文件的MD5,查询特征库，看当前文件是否在开源代码库中
SolrDocument versionIdAndPath = solrUtils.queryOne(versionIdCoreName, "sourceFileMd5:" + originalFileMd5, "versionId,fullPath,sourceFileMd5");
if (versionIdAndPath != null) {
//根据版本ID查询版本的详细信息
VersionTree versionInfo = solrUtils.queryVersionInfoByVersionId((String) versionIdAndPath.get("versionId"));
if (versionInfo != null) {
//当前开源文件的开源项目信息
MatchOpenFile matchOpenFileInfo = new MatchOpenFile();
matchOpenFileInfo.setPId(versionInfo.getProId())
.setPName(versionInfo.getProName())
.setSourceUrl(versionInfo.getDownUrl())
.setFeatureSimilarity(100.00f)
.setOpenRate(100.00f)
.setAnalyzeType(AnalysisLevelEnum.FILE_LEVEL.getCode());
//保存当前文件的开源信息到mongo库中
MatchOpenFileMongoDto matchOpenFileMongo = new MatchOpenFileMongoDto();
matchOpenFileMongo.setId(IdGenerator.uuid32())
.setFilePath(analysisFile.getFileUrl())
.setFileName(analysisFile.getName())
.setOpenRate(100.00f)
.setOpenType(analysisFile.getOpenType())
.setMatchOpenFile(Arrays.asList(matchOpenFileInfo));
mongoTemplate.save(matchOpenFileMongo);
}
}
}
/**
* 根据 特征值 从特征库中检索 具有特征相似的
*

43
src/main/java/com/keyware/composeanalysis/task/LineAnalysisTask.java
Unescape View File

@ -138,8 +138,6 @@ public class LineAnalysisTask extends IAnalysisTask {
String versionIdCoreName = FixedValue.SUFFIX_SOLR_VERSION.get(analysisFile.getSuffix());
if (CollectionUtil.isEmpty(matcheOpenSourceFiles)) {
//因为行的特征库较少，这里补充一个对比逻辑,如果当前文件解析失败，或者没有通过代码块匹配到数据，则直接通过文件的md5 再次查询一次solr库
checkByOriginalFileMd5(versionIdCoreName, analysisFile.getMd5());
return;
}
@ -271,47 +269,6 @@ public class LineAnalysisTask extends IAnalysisTask {
return matchOpenFilesRes;
}
/**
* 防止代码块特征库不全，再次根据文件MD5查询开源文件信息, 做二次校验
*
* @param originalFileMd5
* @param versionIdCoreName
*/
private void checkByOriginalFileMd5(String versionIdCoreName, String originalFileMd5) {
//根据文件的MD5,查询特征库，看当前文件是否在开源代码库中
SolrDocument versionIdAndPath = solrUtils.queryOne(versionIdCoreName, "sourceFileMd5:" + originalFileMd5, "versionId,fullPath,sourceFileMd5");
if (versionIdAndPath != null) {
//根据版本ID查询版本的详细信息
VersionTree versionInfo = solrUtils.queryVersionInfoByVersionId((String) versionIdAndPath.get("versionId"));
if (versionInfo != null) {
//当前开源文件的开源项目信息
MatchOpenFile matchOpenFileInfo = new MatchOpenFile();
matchOpenFileInfo.setPId(versionInfo.getProId())
.setPName(versionInfo.getProName())
.setSourceUrl(versionInfo.getDownUrl())
.setFeatureSimilarity(100.00f)
.setOpenRate(100.00f)
.setAnalyzeType(AnalysisLevelEnum.FILE_LEVEL.getCode());
//保存当前文件的开源信息到mongo库中
MatchOpenFileMongoDto matchOpenFileMongo = new MatchOpenFileMongoDto();
matchOpenFileMongo.setId(IdGenerator.uuid32())
.setFilePath(analysisFile.getFileUrl())
.setFileName(analysisFile.getName())
.setOpenRate(100.00f)
.setOpenType(analysisFile.getOpenType())
.setMatchOpenFile(Arrays.asList(matchOpenFileInfo));
mongoTemplate.save(matchOpenFileMongo);
}
}
}
/**
* 将特征值插入到mongo库中
* @param features 特征集合

14
src/main/resources/application.yaml
Unescape View File

@ -5,14 +5,18 @@ spring:
application:
name: compose-analysis-service
cloud:
inetutils:
preferred-networks:
#优先使用下列网段的IP进行网络通信
- 172.16
nacos:
discovery:
server-addr: 172.16.36.100:8848
namespace: 7f9bb282-8ee3-4948-8182-24b7dcadcd5a
server-addr: 172.16.36.7:8848
namespace: 2fad0ca9-bc32-4afd-9f2e-ebc133d5e781
config:
server-addr: 172.16.36.100:8848
namespace: 7f9bb282-8ee3-4948-8182-24b7dcadcd5a
group: dev_group
server-addr: 172.16.36.7:8848
namespace: 2fad0ca9-bc32-4afd-9f2e-ebc133d5e781
file-extension: yaml
config:
import: nacos:compose-analysis-dev.yaml

Write Preview
Loading…
Cancel
Save