pyparsing：定制自己的解析器
在工作中，经常需要解析不同类型的文件，常用的可能就是正则表达式了，简单点的，可能会使用awk。这里要推荐一种比较小众的方式，使用pyparsing来解析文件。
pyparsing可以做些什么呢？主要可以相当方便地定制自己的tokenizer，因此可以很容易拓展，实现自己的parser
下面看一个traceview的解析例子
16803 AsyncTask #3 16804 pool-2-thread-5 16806 pool-3-thread-1 16807 uil-pool-2-thread-1 16808 uil-pool-2-thread-2 16809 uil-pool-2-thread-3 16810 uil-pool-2-thread-4 Trace (threadID action usecs class.method signature): 16736 xit 0 ..dalvik.system.VMDebug.startMethodTracingFilename (Ljava/lang/String;IIZI)V VMDebug.java 16804 xit 0 ..com.android.org.conscrypt.NativeCrypto.EVP_DigestUpdate (Lcom/android/org/conscrypt/OpenSSLDigestContext;[BII)V NativeCrypto.java 16736 xit 218 .dalvik.system.VMDebug.startMethodTracing (Ljava/lang/String;IIZI)V VMDebug.java 16736 xit 225 android.os.Debug.startMethodTracing (Ljava/lang/String;II)V Debug.java 16736 xit 230-android.os.Debug.startMethodTracing (Ljava/lang/String;I)V Debug.java 16736 xit 266-java.lang.reflect.Method.invoke (Ljava/lang/Object;[Ljava/lang/Object;Z)Ljava/lang/Object; Method.java 16804 ent 528 ..java.lang.ClassLoader.loadClass (Ljava/lang/String;)Ljava/lang/Class; ClassLoader.java 16804 ent 543 ...java.lang.ClassLoader.loadClass (Ljava/lang/String;Z)Ljava/lang/Class; ClassLoader.java 16804 ent 548 ....java.lang.ClassLoader.findLoadedClass (Ljava/lang/String;)Ljava/lang/Class; ClassLoader.java 16804 ent 567 .....java.lang.BootClassLoader.getInstance ()Ljava/lang/BootClassLoader; ClassLoader.java 16804 xit 576 .....java.lang.BootClassLoader.getInstance ()Ljava/lang/BootClassLoader; ClassLoader.java 16804 xit 681 ....java.lang.ClassLoader.findLoadedClass (Ljava/lang/String;)Ljava/lang/Class; ClassLoader.java 16804 ent 689 ....com.uc.base.aerie.hack.ClassLoaderSupport$a.loadClass (Ljava/lang/String;Z)Ljava/lang/Class; ProGuard 16804 ent 704 .....java.lang.ClassLoader.getParent ()Ljava/lang/ClassLoader; ClassLoader.java 8 16804 ent 726 ......java.lang.BootClassLoader.loadClass (Ljava/lang/String;Z)Ljava/lang/Class; ClassLoader.java 16804 ent 730 .......java.lang.ClassLoader.findLoadedClass (Ljava/lang/String;)Ljava/lang/Class; ClassLoader.java 16804 ent 734 ........java.lang.BootClassLoader.getInstance ()Ljava/lang/BootClassLoader; ClassLoader.java 16804 xit 740 ........java.lang.BootClassLoader.getInstance ()Ljava/lang/BootClassLoader; ClassLoader.java 16804 xit 754 .......java.lang.ClassLoader.findLoadedClass (Ljava/lang/String;)Ljava/lang/Class; ClassLoader.java 16804 xit 759 ......java.lang.BootClassLoader.loadClass (Ljava/lang/String;Z)Ljava/lang/Class; ClassLoader.java 16804 xit 763 .....java.lang.ClassLoader.loadClass (Ljava/lang/String;)Ljava/lang/Class; ClassLoader.java
这是一部分转换后的原始日志，格式比较标准，因此可以这么定制
import os from pyparsing import Word, nums, Combine, alphas, Literal, ZeroOrMore, Group, \ Suppress semiFlag = Literal(";") dotFlag = Suppress(Literal(".")) multiDot = ZeroOrMore(dotFlag) threadID =Word(nums, max=5) actionField = Word(alphas) usecsField = Word(nums, max=8) clsField = Word(alphas+".") methodField = Combine("(" + ZeroOrMore(Word(alphas + ";/")) + ")" + Word(alphas + "/") + semiFlag) regex = threadID + actionField + usecsField + multiDot + Group(clsField + methodField) + clsField with open(os.path.join(os.getcwd(), "StepBeforeFirstDraw_o.txt"), "rb") as f: lineno = 0 flag = 0 while 1: line = f.readline() lineno += 1 if "threadID action usecs" in line: flag = lineno continue if flag > 0: try: regex.parseString(line).toXML("") except Exception as e: pass
解析结果为：
/usr/bin/python2.7 /home/alex/workspace/virtual_space/project/calclex.py ['16804', 'ent', '528', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '543', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;Z)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '548', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '567', ['java.lang.BootClassLoader.getInstance', '()Ljava/lang/BootClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '576', ['java.lang.BootClassLoader.getInstance', '()Ljava/lang/BootClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '681', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '704', ['java.lang.ClassLoader.getParent', '()Ljava/lang/ClassLoader;'], 'ClassLoader.java'] ['16804', 'ent', '726', ['java.lang.BootClassLoader.loadClass', '(Ljava/lang/String;Z)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '730', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '734', ['java.lang.BootClassLoader.getInstance', '()Ljava/lang/BootClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '740', ['java.lang.BootClassLoader.getInstance', '()Ljava/lang/BootClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '754', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'xit', '759', ['java.lang.BootClassLoader.loadClass', '(Ljava/lang/String;Z)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'xit', '763', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'xit', '771', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;Z)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'xit', '774', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '809', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '814', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;Z)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '818', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '822', ['java.lang.BootClassLoader.getInstance', '()Ljava/lang/BootClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '827', ['java.lang.BootClassLoader.getInstance', '()Ljava/lang/BootClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '842', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '853', ['java.lang.ClassLoader.getParent', '()Ljava/lang/ClassLoader;'], 'ClassLoader.java'] ['16804', 'xit', '857', ['java.lang.ClassLoader.getParent', '()Ljava/lang/ClassLoader;'], 'ClassLoader.java'] ['16804', 'ent', '861', ['java.lang.ClassLoader.loadClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '865', ['java.lang.BootClassLoader.loadClass', '(Ljava/lang/String;Z)Ljava/lang/Class;'], 'ClassLoader.java'] ['16804', 'ent', '869', ['java.lang.ClassLoader.findLoadedClass', '(Ljava/lang/String;)Ljava/lang/Class;'], 'ClassLoader.java']
这样已经很方便去做二次处理了，而且解析规则的可读性也会比正则的强。
