Some functions to help with analysing cpp files.

cpp_code_helpers.py

import os

class ddict(dict):
    def __init__(self, **kwds):
        self.update(kwds)
        self.__dict__ = self
        
def to_markers(lefts,
               rights):
    return [ddict(lhs  = i,
                  lenl = len(i),
                  rhs  = j,
                  lenr = len(j) ) for i,j in zip(lefts,
                                                 rights)]
    
def single_spacing(txt, also_ln=False):
    txt_old = None
    while(txt_old != txt):
        txt_old = txt
        
        txt = txt.replace('  ',' ')
        txt = txt.replace('\t',' ')
        if also_ln:
            txt = txt.replace('\n',' ')
    return txt
    
def remove_whitespace(txt, also_ln=False):
    N = None
    while(N!=len(txt)):
        N = len(txt)
        
        txt = txt.replace(' ','')
        txt = txt.replace('\t','')
        if also_ln:
            txt = txt.replace('\n','')
    return txt

    
def scrub_nonvarchars(txt):
    return ''.join([' ' if i not in '\n_0123456789'
                                    'abcdefghijklmnopqrstuvwxyz'
                                    'ABCDEFGHIJKLMNOPQRSTUVWXYZ' else i for i in txt])
    
def scrub_all_except_newline(txt):
    return ''.join([' ' if i != '\n' else i for i in txt ])


def scrub_comments_and_strings(txt):
    """
    A pre-process to make scraping easier. This function turns all the comments and strings
    into empty text (spaces):
    FROM: /* // */ cout << "hello \"world\"!" << R"(bla)"; //chingching
    TO  :          cout << "                " << R"(   )";             
    """
    
    markers = to_markers(['/*', '"', 'R"('],
                         ['*/', '"', ')"' ])
    
    #Remove any escaped \" or \\\", but not \\" or \\\\" (uneven vs. even slashes)
    txtout = txt
    i = -1
    while i<len(txtout)-1:
        i+=1
        
        if txtout[i] == '\\':
            nrslashes = 0
            for j in range(i, len(txtout)):
                
                if txtout[j] == '\\':
                    nrslashes += 1
                    
                elif txtout[j] == '"':
                    if nrslashes%2==1:
                        txtout = txtout[:j-1]+'  '+txtout[j+1:]
                    i=j
                    break
                else:
                    i=j
                    break

    
    #Match lefts with righs and clear the text inbetween
    #be aware of // comments!!!!
    i = -1
    while(i<len(txtout)-1):
        i+=1
        
        nxtiter = False
        for m in markers:
            if m.lhs == txtout[i:i+m.lenl]:
                #was the last seen \" farther back than the last seen //? then 
                #we are in a comment, skip this event
                if txtout.rfind('//', 0, i+1) > txtout.rfind('\n', 0, i+1):
                    break #--,
        #<-------------------'       
        
                i+=m.lenl
                for j in range(i, len(txtout)):
                    if m.rhs == txtout[j:j+m.lenr]:
                        txtout = (txtout[:i] +
                                  scrub_all_except_newline(txtout[i:j]) +
                                  txtout[j:])
                        
                        i = j+m.lenr-1 #will ++ just now
                        
                        nxiter = True
                        break #-----+
                if nxiter:          #
                    break           #  
        #<--------------------------+
    
    #clear the // commented text and remove the lefover /* and */ signs
    lines = txtout.split('\n')
    for i, line in enumerate(lines):
        idx = line.find('//')
        if idx >= 0:
            lines[i] = line[:idx] + ' '*(len(line)-idx)
    txtout = '\n'.join(lines)
    
    txtout = txtout.replace('/*', '  ')
    txtout = txtout.replace('*/', '  ')
    
    return txtout


def place_back_strings(txt_scrubbed,
                       txt_original):
    """
    This function placed back the strings that was scrubbed away,
    so you end up with only the comments scrubbed.
    """
    markers = to_markers(['/*', '"', 'R"('],
                         ['*/', '"', ')"' ])
    
    
    #Match lefts with righs and clear the text inbetween
    txtout = txt_scrubbed
    i = -1
    while(i<len(txtout)-1):
        i+=1
        
        nxtiter = False
        for m in markers:
            if m.lhs == txtout[i:i+m.lenl]:   
                i = i+m.lenl
                for j in range(i+1, len(txtout)):
                    if m.rhs == txtout[j:j+m.lenr]:
                        txtout = (txtout[:i] +
                                  txt_original[i:j] +
                                  txtout[j:])
                        
                        i = j+m.lenr-1 #will ++ just now
                        
                        nxiter = True
                        break #-----+
                if nxiter:          #
                    break           #  
        #<--------------------------+
    
    return txtout

def txt_views(txt):
    v = ddict()
    v.orig = txt
    v.clean = scrub_comments_and_strings(txt)
    v.nocomments = place_back_strings(v.clean, txt)
    v.vars = scrub_nonvarchars(v.clean)
    
    return v

def view_lnsplit(v):
    return ddict(**{k:v.split('\n') for k,v in v.items()})
    

def isint(txt):
    try:
        int(txt)
        return True
    except: return False
    
def split(txt):
    splits = []
    iskeep = False
    for i, char in enumerate(txt):
        if not iskeep and not char in (' ', '\t', '\n'):
            iskeep = True
            splits.append(ddict(i=i))
        elif iskeep and char in (' ', '\t', '\n'):
            iskeep = False
            splits[-1].j = i
            
    for s in splits:
        s.str = txt[s.i:s.j]
        
    return splits

Untitled.ipynb