如何在Scala上做增量覆盖率(incremental coverage)计算?

在Scala上写单测

下面是之前写的英文版简述,直接粘贴过来了。 你还可以参考:Spark-Scala单元测试实践 - 码农教程 (manongjc.com)

Unit Tests

Unit testing is a powerful tool for software quality -- and has been for decades. Unit tests provide a fundamental check that an application meets its software design specifications and behaves as intended.

What tools are used?

UnitTest Tool: ScalaTest

Mock Tool: Mockito Scala

HTTP Mock Tool: WireMock

Code Coverage Tool: Scoverage

How to write tests?

First, you need to know that writing a single test is easy. More information on ScalaTest

Tips:

  1. You can mock functions by Mockito Scala
  2. You can mock http service by WireMock
  3. When there are public functions, make full use of before, after and withFixture.

重点来了,如何做增量覆盖率测试?

我们用 Scoverage 做覆盖率测试,但是它不支持增量覆盖率测试,所以我们需要手动改造一下。 如果你用 jacoco做的覆盖率测试,可以试试 jacoco: jacoco二开,支持增量代码覆盖率 (gitee.com) 这篇文章。

大概的思路是:

  1. Calculate the number of files and lines with changes through git diff
  2. Using the feature of the scoverage, comments are added before and after the number of lines that have changed.
    see: https://github.com/scoverage/scalac-scoverage-plugin

贴上python脚本: HandleIncrementalCoverage.py

#!/usr/bin/env python

"""
Handle code for Incremental Coverage
Tips: only active with scoverage
Principle:
  1. Calculate the number of files and lines with changes through git diff
  2. Using the feature of the scoverage, comments are added before and after the number of lines that have changed.
     see: https://github.com/scoverage/scalac-scoverage-plugin
"""
import os
import re
import subprocess
import sys


def getChangedLineInfoFromDiffLines(lines):
    """
    args : Lines, the description of a file output by the git-diff command
    returns : List[(changedLineStart, changedLineEnd)], Front-closed and back-open interval
    """
    changedLineInfo = []
    # Get line change information according to "@"
    # Matching [0]: "," + the number of rows deleted from here;
    # [1]: the number of rows added here;
    # [2]: "," + the number of rows added from here
    reg = re.compile("^@@ -[0-9]+(,[0-9]+)? \+([0-9]+)(,[0-9]+)? @@")
    for line in lines:
        r = reg.findall(line)
        if len(r) > 0:
            changedLineStart = int(r[0][1])
            caughtLineCountStr = r[0][2]
            if len(caughtLineCountStr) > 0:
                changedLineCount = int(caughtLineCountStr[1:])
            else:
                changedLineCount = 1
            changedLineInfo.append((changedLineStart, changedLineStart + changedLineCount))
    return changedLineInfo


def getDiffLines(baseBranch='HEAD~1', newBranch='HEAD', dir="./"):
    """get diff lines from two branches."""
    gitCmd = f"git diff --unified=0 --diff-filter=d {baseBranch} {newBranch} {dir}"
    print("Git Cmd: ", gitCmd)
    gitDiffOutputRaw = subprocess.check_output(gitCmd.split(" "))
    outputStr = gitDiffOutputRaw.decode('utf-8')
    diffOutputLines = outputStr.splitlines()

    map = {}
    separateLineReg = re.compile("^diff --git a/\S+ b/(\S+)")
    currentCheckFileName = ""
    diffLinesForCurrentCheckFile = []
    for i in range(len(diffOutputLines)):
        l = diffOutputLines[i]
        separateLineMatchResult = separateLineReg.findall(l)
        if len(separateLineMatchResult) > 0:
            if len(diffLinesForCurrentCheckFile) > 0:
                a = getChangedLineInfoFromDiffLines(diffLinesForCurrentCheckFile)
                map[currentCheckFileName] = a
                diffLinesForCurrentCheckFile.clear()
            # filter submodule name
            currentCheckFileName = '/'.join((separateLineMatchResult[0]).split('/')[1:])
        else:
            diffLinesForCurrentCheckFile.append(l)
        if i == len(diffOutputLines) - 1:
            a = getChangedLineInfoFromDiffLines(diffLinesForCurrentCheckFile)
            map[currentCheckFileName] = a
    print("Git Diff Output: ", map)
    return map

def findAllFile(base):
    for root, ds, fs in os.walk(base):
        for f in fs:
            if f.endswith('.scala'):
                fullname = os.path.join(root, f).replace('\\', '/')
                yield fullname

def preHandleIncrementalCoverage(diffDict):
    """PreHandle Incremental Coverage"""
    for i in findAllFile("src/main/scala"):
        diffInfo = diffDict[i] if i in diffDict else []
        print("Add annotation for:", i, diffInfo)
        fileData, diffLen = [], len(diffInfo)
        with open(i, "r", encoding="utf-8") as f:
            lineNum, diffIndex, diffOffset = 0, 0, 0
            for line in f:
                if "auto add for Incremental Coverage" in line:
                    continue
                lineNum += 1
                if lineNum == 2:
                    fileData.append("// $COVERAGE-OFF$ auto add for Incremental Coverage\n")
                if diffIndex < diffLen and diffInfo[diffIndex][diffOffset] == lineNum:
                    if diffOffset == 0:
                        fileData.append("// $COVERAGE-ON$ auto add for Incremental Coverage\n")
                        diffOffset = 1
                    else:
                        fileData.append("// $COVERAGE-OFF$ auto add for Incremental Coverage\n")
                        diffOffset = 0
                        diffIndex += 1
                fileData.append(line)
        with open(i, "w", encoding="utf-8") as f:
            f.write("".join(fileData))

def cleanIncrementalCoverage():
    """clean Incremental Coverage"""
    for i in findAllFile("src/main/scala"):
        fileData = []
        with open(i, "r", encoding="utf-8") as f:
            for line in f:
                if "auto add for Incremental Coverage" in line:
                    continue
                fileData.append(line)
        with open(i, "w", encoding="utf-8") as f:
            f.write("".join(fileData))


if __name__ == '__main__':
    # os.getcwd() is the root directory of the submodule
    if len(sys.argv) == 2 and sys.argv[1] == "clean":
        cleanIncrementalCoverage()
        exit(0)
    if len(sys.argv) != 3:
        raise Exception("Argv not enough. Usage: python3 HandleIncrementalCoverage.py baseBranch newBranch")
    baseBranch = sys.argv[1]
    newBranch = sys.argv[2]
    diff = getDiffLines(baseBranch, newBranch, "./src/main/scala")
    preHandleIncrementalCoverage(diff)

对应的再改一下pom.xml

注意:这里的依赖只是ut用到的一些依赖和插件,不包括项目用的,请根据实际情况进行改造

    
        3.2.3
        3.1.2
        2.12
        2.12.10
        3.2.0
        2.5.0
        3.6.6
        4.5.6
        false
        true
        HEAD~1
        HEAD
        true
        true
    
    
        
            org.scalatest
            scalatest_${scala.binary.version}
            ${scalatest.version}
            test
        
        
            org.scalactic
            scalactic_${scala.binary.version}
            3.2.12
            test
        
        
            org.pegdown
            pegdown
            1.4.2
            test
        
        
            com.vladsch.flexmark
            flexmark-all
            0.35.10
            test
        
        
            org.mockito
            mockito-scala_${scala.binary.version}
            1.16.37
            test
        
        
            org.apache.hadoop
            hadoop-azure
            ${hadoop.version}
            test
        
        
            com.github.tomakehurst
            wiremock-jre8-standalone
            2.33.2
            test
        
    
    
        ${basedir}/src/main/scala
        
            
                org.apache.maven.plugins
                maven-compiler-plugin
            
            
                net.alchim31.maven
                scala-maven-plugin
            
            
                org.apache.maven.plugins
                maven-jar-plugin
                
                    ${project.parent.basedir}/target/
                
            

            
                org.scalatest
                scalatest-maven-plugin
                1.0
                
                    ${project.build.directory}/surefire-reports
                    .
                    WDF TestResult.txt
                    ${project.build.directory}/site/scalatest
                    false
                    ${skipUT}
                
                
                    
                        test
                        test
                        
                            test
                        
                    
                
            

            
                org.codehaus.mojo
                exec-maven-plugin
                
                    
                        add annotation for incremental coverage
                        compile
                        
                            exec
                        
                        
                            ${notIncrementalCoverage}
                            python
                            HandleIncrementalCoverage.py ${baseBranch} ${newBranch}
                        
                    
                    
                        remove annotation for incremental coverage
                        clean
                        
                            exec
                        
                        
                            python
                            HandleIncrementalCoverage.py clean
                        
                    
                
            
            
            
                org.scoverage
                scoverage-maven-plugin
                ${scoverage.plugin.version}
                
                    
                        test
                        test
                        
                            report
                        
                    
                
                
                    ${skipUT}
                    ${scala.version}
                    true
                    true
                    ${project.build.sourceEncoding}
                
            
        
    

计算增量覆盖率:

echo "calculate incremental coverage between master and HEAD"
mvn -P spark-3.1 test -DargLine="-DnotIncrementalCoverage=false" "-DbaseBranch=origin/master" "-DnewBranch=HEAD" "-DnotIncrementalCoverage=false"

计算全量覆盖率:

mvn -P spark-3.1 test

你可能感兴趣的:(如何在Scala上做增量覆盖率(incremental coverage)计算?)