问题
I have done fair bit of searching around how to capture python ASTs with comments preserved. The suggested way includes using ast and tokenize libraries to get the job done.
I have had fair bit of success in utilizing these libraries as per my requirement but I feel there has to be a better way.
This thought stems from the fact that lib2to3 converts python2 code to python3 code with comments preserved. Also the process is stated to be Source-Code-in-Python2 -> AST -> Source-Code-in-Python3 (to be put forth in a simplified manner).
My question is how do I capture the in-between AST? I have looked at python-docs but there is no command line flag to get hold of the AST.
Just to provide you the context: I am trying to convert python source code to an XML file (with comments preserved) for some further processing
回答1:
Just to provide you the context: I am trying to convert python source code to an XML file (with comments preserved) for some further processing
An "easy" way is to use a tool that already does this, rather than reinventing it, especially if you are short on time.
Our DMS Software Reengineering Toolkit can parse Python (and many other languages), build ASTs, and capture comments, and spit out the resulting tree as XML. See example below.
A remark: XML initially seems nice, but is a clumsy way to represent/analyze/transform code. The reason that tools like DMS exist, is to provide all the machinery necessary to manipulate the parsed ASTs in ways that are more effective than XML transformation, and scale much better (e.g., to millions of lines of code): ultimately, to save engineering time and runtime.
Even if you decide on XML, where are you going to get good tools to process it? (XSLT isn't the right answer). Finally, if you intend to modify the program, and you change the XML, how do you intend to get the source code back? DMS can modify ASTs and regenerate valid source program text (including the comments).
So while DMS will export ASTs in XML (because people like you seem to insist on it), this feature is rarely used in practice. The typical use case is parse, analyze, modify the AST, then prettyprint the modified AST, all using DMS in an integrated way.
For this python program:
# A comment in the header
import sys
TOKENBLANKS=1
class MyClassNameTranslator:
# get_name looks up name
def get_name(self, name):
"""Get a translation for a real name"""
return self.realnames[name]
DMS generates the follow XML version of its AST, complete with captured comments:
C:\[snip]Python\Tools\Parser>run ..\domainparser ++XML C:\[snip]tiny.py
Python~v3_0 Domain Parser Version 2.5.15
Copyright (C) 1996-2013 Semantic Designs, Inc; All Rights Reserved; SD Confidential
Powered by DMS (R) Software Reengineering Toolkit
165 tree nodes in tree.
<?xml version="1.1" encoding="ISO-8859-1"?>
<!-- Using DMS PrintASTasXML (v.1.03) -->
<!-- XML generated on 2014/03/01 12:14:49 -->
<DMSForest>
<tree node="Python" type="1" domain="1" id="yk0x" parents="0" line="2" column="1" file="1">
<tree node="file_input" type="2" domain="1" id="yk0w" line="2" column="1" file="1">
<tree node="file_input_element_list" type="4" domain="1" id="yk0v" line="2" column="1" file="1">
<tree node="file_input_element_list" type="4" domain="1" id="yjww" line="2" column="1" file="1">
<tree node="file_input_element_list" type="4" domain="1" id="yjvc" line="2" column="1" file="1">
<tree node="file_input_element_list" type="4" domain="1" id="yjus" line="2" column="1" file="1">
<tree node="file_input_element_list" type="3" domain="1" id="ydby" line="2" column="1" file="1"/>
<tree node="file_input_element" type="5" domain="1" id="yjuq" line="2" column="1" file="1">
<tree node="NEWLINE" type="282" domain="1" id="ydbn" literal="0" line="2" column="1" file="1">
<precomment child="0" index="1"># A comment in the header</precomment>
</tree>
</tree>
</tree>
<tree node="file_input_element" type="6" domain="1" id="yjvb" line="3" column="1" file="1">
<tree node="stmt" type="7" domain="1" id="yjva" line="3" column="1" file="1">
<tree node="simple_stmt" type="9" domain="1" id="yjv9" line="3" column="1" file="1">
<tree node="small_stmt_list" type="11" domain="1" id="yjv3" line="3" column="1" file="1">
<tree node="small_stmt" type="45" domain="1" id="yjv1" line="3" column="1" file="1">
<tree node="'import'" type="305" domain="1" id="yjup" literal="0" line="3" column="1" file="1"/>
<tree node="dotted_as_name_list" type="53" domain="1" id="yjuz" line="3" column="8" file="1">
<tree node="dotted_as_name" type="60" domain="1" id="yjuy" line="3" column="8" file="1">
<tree node="dotted_name" type="61" domain="1" id="yjux" line="3" column="8" file="1">
<tree node="NAME" type="310" domain="1" id="yjuv" line="3" column="8" file="1">
<literal>sys</literal>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
<tree node="NEWLINE" type="282" domain="1" id="yjuw" literal="0" line="3" column="11" file="1"/>
</tree>
</tree>
</tree>
</tree>
<tree node="file_input_element" type="6" domain="1" id="yjwv" line="5" column="1" file="1">
<tree node="stmt" type="7" domain="1" id="yjwu" line="5" column="1" file="1">
<tree node="simple_stmt" type="9" domain="1" id="yjwt" line="5" column="1" file="1">
<tree node="small_stmt_list" type="11" domain="1" id="yjwo" line="5" column="1" file="1">
<tree node="small_stmt" type="14" domain="1" id="yjwl" line="5" column="1" file="1">
<tree node="assign_stmt" type="15" domain="1" id="yjwj" line="5" column="1" file="1">
<tree node="target_list" type="215" domain="1" id="yjvg" line="5" column="1" file="1">
<tree node="targets" type="217" domain="1" id="yjvf" line="5" column="1" file="1">
<tree node="target" type="221" domain="1" id="yjve" line="5" column="1" file="1">
<tree node="NAME" type="310" domain="1" id="yjv8" line="5" column="1" file="1">
<literal>TOKENBLANKS</literal>
</tree>
</tree>
</tree>
</tree>
<tree node="'='" type="284" domain="1" id="yjvd" literal="0" line="5" column="12" file="1"/>
<tree node="assign_rhs" type="29" domain="1" id="yjwh" line="5" column="13" file="1">
<tree node="test_list" type="30" domain="1" id="yjwf" line="5" column="13" file="1">
<tree node="tests" type="32" domain="1" id="yjwc" line="5" column="13" file="1">
<tree node="test" type="151" domain="1" id="yjwa" line="5" column="13" file="1">
<tree node="or_test" type="152" domain="1" id="yjw8" line="5" column="13" file="1">
<tree node="and_test" type="154" domain="1" id="yjw4" line="5" column="13" file="1">
<tree node="not_test" type="157" domain="1" id="yjw1" line="5" column="13" file="1">
<tree node="comparison" type="158" domain="1" id="yjvz" line="5" column="13" file="1">
<tree node="expr" type="170" domain="1" id="yjvx" line="5" column="13" file="1">
<tree node="xor_expr" type="172" domain="1" id="yjvv" line="5" column="13" file="1">
<tree node="and_expr" type="174" domain="1" id="yjvs" line="5" column="13" file="1">
<tree node="shift_expr" type="176" domain="1" id="yjvq" line="5" column="13" file="1">
<tree node="arith_expr" type="179" domain="1" id="yjvo" line="5" column="13" file="1">
<tree node="term" type="182" domain="1" id="yjvn" line="5" column="13" file="1">
<tree node="factor" type="187" domain="1" id="yjvm" line="5" column="13" file="1">
<tree node="power" type="191" domain="1" id="yjvl" line="5" column="13" file="1">
<tree node="value" type="194" domain="1" id="yjvk" line="5" column="13" file="1">
<tree node="constant" type="197" domain="1" id="yjvj" line="5" column="13" file="1">
<tree node="INTEGER" type="355" domain="1" id="yjvh" literal="1" line="5" column="13" file="1"/>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
<tree node="NEWLINE" type="282" domain="1" id="yjvi" literal="0" line="5" column="14" file="1"/>
</tree>
</tree>
</tree>
</tree>
<tree node="file_input_element" type="6" domain="1" id="yk0u" line="7" column="1" file="1">
<tree node="stmt" type="8" domain="1" id="yk0p" line="7" column="1" file="1">
<tree node="compound_stmt" type="143" domain="1" id="yk0s" line="7" column="1" file="1">
<tree node="decorators" type="144" domain="1" id="yjwx" line="7" column="1" file="1"/>
<tree node="'class'" type="330" domain="1" id="yjws" literal="0" line="7" column="1" file="1"/>
<tree node="NAME" type="310" domain="1" id="yjwy" line="7" column="7" file="1">
<literal>MyClassNameTranslator</literal>
</tree>
<tree node="':'" type="314" domain="1" id="yjwz" literal="0" line="7" column="28" file="1"/>
<tree node="block" type="115" domain="1" id="yk0q" line="7" column="29" file="1">
<tree node="NEWLINE" type="282" domain="1" id="yjx0" literal="0" line="7" column="29" file="1"/>
<tree node="INDENT" type="324" domain="1" id="yjx1" literal="0" line="10" column="1" file="1">
<precomment child="0" index="1"># get_name looks up name</precomment>
</tree>
<tree node="stmt_list" type="116" domain="1" id="yk0o" line="10" column="5" file="1">
<tree node="stmt" type="8" domain="1" id="yk0j" line="10" column="5" file="1">
<tree node="compound_stmt" type="119" domain="1" id="yk0m" line="10" column="5" file="1">
<tree node="decorators" type="144" domain="1" id="yjx5" line="10" column="5" file="1"/>
<tree node="'def'" type="326" domain="1" id="yjx4" literal="0" line="10" column="5" file="1"/>
<tree node="NAME" type="310" domain="1" id="yjx6" line="10" column="9" file="1">
<literal>get_name</literal>
</tree>
<tree node="parameters" type="121" domain="1" id="yjxe" line="10" column="17" file="1">
<tree node="'('" type="327" domain="1" id="yjx7" literal="0" line="10" column="17" file="1"/>
<tree node="optional_varargslist" type="123" domain="1" id="yjxd" line="10" column="18" file="1">
<tree node="varargslist" type="126" domain="1" id="yjw6" line="10" column="18" file="1">
<tree node="fpdef_test_list_prefix" type="131" domain="1" id="yjxk" line="10" column="18" file="1">
<tree node="fpdef_test_list_prefix" type="130" domain="1" id="yjx9" line="10" column="18" file="1"/>
<tree node="fpdef_test_comma" type="132" domain="1" id="yjxh" line="10" column="18" file="1">
<tree node="fpdef_test" type="133" domain="1" id="yjxc" line="10" column="18" file="1">
<tree node="fpdef" type="135" domain="1" id="yjxb" line="10" column="18" file="1">
<tree node="NAME" type="310" domain="1" id="yjx8" line="10" column="18" file="1">
<literal>self</literal>
</tree>
</tree>
</tree>
<tree node="','" type="297" domain="1" id="yjxa" literal="0" line="10" column="22" file="1"/>
</tree>
</tree>
<tree node="fpdef_test" type="133" domain="1" id="yjw3" line="10" column="24" file="1">
<tree node="fpdef" type="135" domain="1" id="yjvw" line="10" column="24" file="1">
<tree node="NAME" type="310" domain="1" id="yjxg" line="10" column="24" file="1">
<literal>name</literal>
</tree>
</tree>
</tree>
</tree>
</tree>
<tree node="')'" type="328" domain="1" id="yjvt" literal="0" line="10" column="28" file="1"/>
</tree>
<tree node="':'" type="314" domain="1" id="yjxl" literal="0" line="10" column="29" file="1"/>
<tree node="block" type="115" domain="1" id="yk0k" line="10" column="30" file="1">
<tree node="NEWLINE" type="282" domain="1" id="yjxf" literal="0" line="10" column="30" file="1"/>
<tree node="INDENT" type="324" domain="1" id="yjxm" literal="0" line="11" column="1" file="1"/>
<tree node="stmt_list" type="117" domain="1" id="yk0h" line="11" column="9" file="1">
<tree node="stmt_list" type="116" domain="1" id="yjyq" line="11" column="9" file="1">
<tree node="stmt" type="7" domain="1" id="yjyp" line="11" column="9" file="1">
<tree node="simple_stmt" type="9" domain="1" id="yjyo" line="11" column="9" file="1">
<tree node="small_stmt_list" type="11" domain="1" id="yjyk" line="11" column="9" file="1">
<tree node="small_stmt" type="13" domain="1" id="yjyh" line="11" column="9" file="1">
<tree node="testlist" type="255" domain="1" id="yjyf" line="11" column="9" file="1">
<tree node="test_plus" type="256" domain="1" id="yjyd" line="11" column="9" file="1">
<tree node="test" type="151" domain="1" id="yjya" line="11" column="9" file="1">
<tree node="or_test" type="152" domain="1" id="yjy7" line="11" column="9" file="1">
<tree node="and_test" type="154" domain="1" id="yjy5" line="11" column="9" file="1">
<tree node="not_test" type="157" domain="1" id="yjy3" line="11" column="9" file="1">
<tree node="comparison" type="158" domain="1" id="yjy1" line="11" column="9" file="1">
<tree node="expr" type="170" domain="1" id="yjxy" line="11" column="9" file="1">
<tree node="xor_expr" type="172" domain="1" id="yjxw" line="11" column="9" file="1">
<tree node="and_expr" type="174" domain="1" id="yjxv" line="11" column="9" file="1">
<tree node="shift_expr" type="176" domain="1" id="yjxu" line="11" column="9" file="1">
<tree node="arith_expr" type="179" domain="1" id="yjxt" line="11" column="9" file="1">
<tree node="term" type="182" domain="1" id="yjxs" line="11" column="9" file="1">
<tree node="factor" type="187" domain="1" id="yjxr" line="11" column="9" file="1">
<tree node="power" type="191" domain="1" id="yjxq" line="11" column="9" file="1">
<tree node="value" type="194" domain="1" id="yjxp" line="11" column="9" file="1">
<tree node="constant" type="200" domain="1" id="yjxo" line="11" column="9" file="1">
<tree node="string_sequence" type="208" domain="1" id="yjxj" line="11" column="9" file="1">
<tree node="STRING" type="362" domain="1" id="yjxn" line="11" column="9" file="1">
<literal>Get a translation for a real name</literal>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
<tree node="NEWLINE" type="282" domain="1" id="yjxi" literal="0" line="11" column="48" file="1"/>
</tree>
</tree>
</tree>
<tree node="stmt" type="7" domain="1" id="yk0g" line="12" column="9" file="1">
<tree node="simple_stmt" type="9" domain="1" id="yk0f" line="12" column="9" file="1">
<tree node="small_stmt_list" type="11" domain="1" id="yk09" line="12" column="9" file="1">
<tree node="small_stmt" type="39" domain="1" id="yk06" line="12" column="9" file="1">
<tree node="'return'" type="302" domain="1" id="yjyn" literal="0" line="12" column="9" file="1"/>
<tree node="testlist" type="255" domain="1" id="yk03" line="12" column="16" file="1">
<tree node="test_plus" type="256" domain="1" id="yk02" line="12" column="16" file="1">
<tree node="test" type="151" domain="1" id="yk01" line="12" column="16" file="1">
<tree node="or_test" type="152" domain="1" id="yk00" line="12" column="16" file="1">
<tree node="and_test" type="154" domain="1" id="yjzz" line="12" column="16" file="1">
<tree node="not_test" type="157" domain="1" id="yjzy" line="12" column="16" file="1">
<tree node="comparison" type="158" domain="1" id="yjzx" line="12" column="16" file="1">
<tree node="expr" type="170" domain="1" id="yjzw" line="12" column="16" file="1">
<tree node="xor_expr" type="172" domain="1" id="yjzv" line="12" column="16" file="1">
<tree node="and_expr" type="174" domain="1" id="yjzu" line="12" column="16" file="1">
<tree node="shift_expr" type="176" domain="1" id="yjzt" line="12" column="16" file="1">
<tree node="arith_expr" type="179" domain="1" id="yjzs" line="12" column="16" file="1">
<tree node="term" type="182" domain="1" id="yjzr" line="12" column="16" file="1">
<tree node="factor" type="187" domain="1" id="yjzq" line="12" column="16" file="1">
<tree node="power" type="191" domain="1" id="yjzp" line="12" column="16" file="1">
<tree node="value" type="195" domain="1" id="yjzo" line="12" column="16" file="1">
<tree node="value" type="195" domain="1" id="yjz0" line="12" column="16" file="1">
<tree node="value" type="193" domain="1" id="yjyu" line="12" column="16" file="1">
<tree node="atom" type="207" domain="1" id="yjyt" line="12" column="16" file="1">
<tree node="NAME" type="310" domain="1" id="yjyr" line="12" column="16" file="1">
<literal>self</literal>
</tree>
</tree>
</tree>
<tree node="trailer" type="228" domain="1" id="yjyz" line="12" column="20" file="1">
<tree node="'.'" type="312" domain="1" id="yjys" literal="0" line="12" column="20" file="1"/>
<tree node="NAME" type="310" domain="1" id="yjyv" line="12" column="21" file="1">
<literal>realnames</literal>
</tree>
</tree>
</tree>
<tree node="trailer" type="227" domain="1" id="yjzn" line="12" column="30" file="1">
<tree node="index" type="230" domain="1" id="yjzm" line="12" column="30" file="1">
<tree node="'['" type="358" domain="1" id="yjyy" literal="0" line="12" column="30" file="1"/>
<tree node="subscript_list" type="234" domain="1" id="yjzj" line="12" column="31" file="1">
<tree node="subscript" type="249" domain="1" id="yjzi" line="12" column="31" file="1">
<tree node="test" type="151" domain="1" id="yjzh" line="12" column="31" file="1">
<tree node="or_test" type="152" domain="1" id="yjzg" line="12" column="31" file="1">
<tree node="and_test" type="154" domain="1" id="yjzf" line="12" column="31" file="1">
<tree node="not_test" type="157" domain="1" id="yjze" line="12" column="31" file="1">
<tree node="comparison" type="158" domain="1" id="yjzd" line="12" column="31" file="1">
<tree node="expr" type="170" domain="1" id="yjzc" line="12" column="31" file="1">
<tree node="xor_expr" type="172" domain="1" id="yjzb" line="12" column="31" file="1">
<tree node="and_expr" type="174" domain="1" id="yjza" line="12" column="31" file="1">
<tree node="shift_expr" type="176" domain="1" id="yjz9" line="12" column="31" file="1">
<tree node="arith_expr" type="179" domain="1" id="yjz8" line="12" column="31" file="1">
<tree node="term" type="182" domain="1" id="yjz7" line="12" column="31" file="1">
<tree node="factor" type="187" domain="1" id="yjz6" line="12" column="31" file="1">
<tree node="power" type="191" domain="1" id="yjz5" line="12" column="31" file="1">
<tree node="value" type="193" domain="1" id="yjz4" line="12" column="31" file="1">
<tree node="atom" type="207" domain="1" id="yjz3" line="12" column="31" file="1">
<tree node="NAME" type="310" domain="1" id="yjz1" line="12" column="31" file="1">
<literal>name</literal>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
<tree node="']'" type="359" domain="1" id="yjz2" literal="0" line="12" column="35" file="1"/>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
<tree node="NEWLINE" type="282" domain="1" id="yjzl" literal="0" line="12" column="36" file="1"/>
</tree>
</tree>
</tree>
<tree node="DEDENT" type="325" domain="1" id="yk0e" literal="0" line="14" column="1" file="1"/>
</tree>
</tree>
</tree>
</tree>
<tree node="DEDENT" type="325" domain="1" id="yk0i" literal="0" line="14" column="1" file="1"/>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
</tree>
<FileIndex>
<File index="1">C:/DMS/Domains/Python/v2_6/Examples/tiny.py</File>
</FileIndex>
<DomainIndex>
<Domain index="1">Python~v3_0</Domain>
</DomainIndex>
</DMSForest>
来源：https://stackoverflow.com/questions/22002968/easy-way-to-convert-python-source-code-to-an-ast-with-comments-intact