generate_docstrings.py

利用C

#!/usr/bin/env python

__author__ = "Johan Hake & Prabhu Ramachandran"
__license__ = "bsd style"

import os

from xml.dom import minidom
import re
import textwrap
import sys
import types

def my_open_read(source):
    if hasattr(source, "read"):
        return source
    else:
        return open(source)

def my_open_write(dest):
    if hasattr(dest, "write"):
        return dest
    else:
        return open(dest, 'w')


class Doxy2SWIG:    
    """Converts Doxygen generated XML files into a file containing
    docstrings that can be used by SWIG-1.3.x that have support for
    feature("docstring").  Once the data is parsed it is stored in
    self.pieces.

    """    
    __author__ = "Prabhu Ramachandran"
    __license__ = "BSD style"
    # This code is implemented using Mark Pilgrim's code as a guideline:
    #   http://www.faqs.org/docs/diveintopython/kgp_divein.html

    def __init__(self, src):
        """Initialize the instance given a source object (file or
        filename).

        """
        f = my_open_read(src)
        self.my_dir = os.path.dirname(f.name)
        self.xmldoc = minidom.parse(f).documentElement
        f.close()

        self.pieces = []
        self.pieces.append('\n// File: %s\n'%\
                           os.path.basename(f.name))

        self.space_re = re.compile(r'\s+')
        self.lead_spc = re.compile(r'^(%feature\S+\s+\S+\s*?)"\s+(\S)')
        self.multi = 0
        self.ignores = ('inheritancegraph', 'param', 'listofallmembers',
                        'innerclass', 'name', 'declname', 'incdepgraph',
                        'invincdepgraph', 'programlisting', 'type',
                        'references', 'referencedby', 'location',
                        'collaborationgraph', 'reimplements',
                        'reimplementedby', 'derivedcompoundref',
                        'basecompoundref', 'argsstring')
        #self.generics = []
        self.include_function_definition = False
        
    def generate(self):
        """Parses the file set in the initialization.  The resulting
        data is stored in `self.pieces`.

        """
        self.parse(self.xmldoc)
    
    def parse(self, node):
        """Parse a given node.  This function in turn calls the
        `parse_<nodeType>` functions which handle the respective
        nodes.

        """
        pm = getattr(self, "parse_%s"%node.__class__.__name__)
        pm(node)

    def parse_Document(self, node):
        self.parse(node.documentElement)

    def parse_Text(self, node):
        txt = node.data
        txt = txt.replace('\\', r'\\\\')
        txt = txt.replace('"', r'\"')
        # ignore pure whitespace
        m = self.space_re.match(txt)
        if m and len(m.group()) == len(txt):
            pass
        else:
            self.add_text(textwrap.fill(txt))

    def parse_Element(self, node):
        """Parse an `ELEMENT_NODE`.  This calls specific
        `do_<tagName>` handers for different elements.  If no handler
        is available the `generic_parse` method is called.  All
        tagNames specified in `self.ignores` are simply ignored.
        
        """
        name = node.tagName
        ignores = self.ignores
        if name in ignores:
            return
        attr = "do_%s" % name
        if hasattr(self, attr):
            handlerMethod = getattr(self, attr)
            handlerMethod(node)
        else:
            self.generic_parse(node)
            #if name not in self.generics: self.generics.append(name)

    def parse_Comment(self, node):
        return

    def add_text(self, value):
        """Adds text corresponding to `value` into `self.pieces`."""
        if type(value) in (types.ListType, types.TupleType):
            self.pieces.extend(value)
        else:
            self.pieces.append(value)

    def get_specific_nodes(self, node, names):
        """Given a node and a sequence of strings in `names`, return a
        dictionary containing the names as keys and child
        `ELEMENT_NODEs`, that have a `tagName` equal to the name.

        """
        nodes = [(x.tagName, x) for x in node.childNodes \
                 if x.nodeType == x.ELEMENT_NODE and \
                 x.tagName in names]
        return dict(nodes)

    def generic_parse(self, node, pad=0):
        """A Generic parser for arbitrary tags in a node.

        Parameters:

         - node:  A node in the DOM.
         - pad: `int` (default: 0)

           If 0 the node data is not padded with newlines.  If 1 it
           appends a newline after parsing the childNodes.  If 2 it
           pads before and after the nodes are processed.  Defaults to
           0.

        """
        npiece = 0
        if pad:
            npiece = len(self.pieces)
            if pad == 2:
                self.add_text('\n')                
        for n in node.childNodes:
            self.parse(n)
        if pad:
            if len(self.pieces) > npiece:
                self.add_text('\n')

    def space_parse(self, node):
        self.add_text(' ')
        self.generic_parse(node)

    do_ref = space_parse
    do_emphasis = space_parse
    do_bold = space_parse
    do_computeroutput = space_parse
    do_formula = space_parse

    def do_compoundname(self, node):
        self.add_text('\n\n')
        data = node.firstChild.data
        self.add_text('%%feature("docstring") %s "\n'%data)

    def do_compounddef(self, node):
        kind = node.attributes['kind'].value
        if kind in ('class', 'struct'):
            prot = node.attributes['prot'].value
            if prot <> 'public':
                return
            names = ('compoundname', 'briefdescription',
                     'detaileddescription', 'includes')
            first = self.get_specific_nodes(node, names)
            for n in names:
                if first.has_key(n):
                    self.parse(first[n])
            self.add_text(['";','\n'])
            for n in node.childNodes:
                if n not in first.values():
                    self.parse(n)
        elif kind in ('file', 'namespace'):
            nodes = node.getElementsByTagName('sectiondef')
            for n in nodes:
                self.parse(n)

    def do_includes(self, node):
        self.add_text('C++ includes: ')
        self.generic_parse(node, pad=1)

    def do_parameterlist(self, node):
        self.add_text(['\n', '\n', 'Parameters:', '\n'])
        self.generic_parse(node, pad=1)

    def do_para(self, node):
        self.add_text('\n')
        self.generic_parse(node, pad=1)

    def do_parametername(self, node):
        self.add_text('\n')
        self.add_text("%s: "%node.firstChild.data)

    def do_parameterdefinition(self, node):
        self.generic_parse(node, pad=1)

    def do_detaileddescription(self, node):
        self.generic_parse(node, pad=1)

    def do_briefdescription(self, node):
        self.generic_parse(node, pad=1)

    def do_memberdef(self, node):
        prot = node.attributes['prot'].value
        id = node.attributes['id'].value
        kind = node.attributes['kind'].value
        tmp = node.parentNode.parentNode.parentNode
        compdef = tmp.getElementsByTagName('compounddef')[0]
        cdef_kind = compdef.attributes['kind'].value
        
        if prot == 'public':
            first = self.get_specific_nodes(node, ('definition', 'name'))