Welcome, guest | Sign In | My Account | Store | Cart

This is my simple web crawler. It takes as input a list of seed pages (web urls) and 'scrapes' each page of all its absolute path links (i.e. links in the format http://) and adds those to a dictionary. The web crawler can take all the links found in the seed pages and then scrape those as well. You can continue scraping as deep as you like. You can control how "deep you go" by specifying the depth variable passed into the WebCrawler class function start_crawling(seed_pages,depth). Think of the depth as the recursion depth (or the number of web pages deep you go before returning back up the tree).

To make this web crawler a little more interesting I added some bells and whistles. I added the ability to pass into the WebCrawler class constructor a regular expression object. The regular expression object is used to "filter" the links found during scraping. For example, in the code below you will see:

cnn_url_regex = re.compile('(?<=[.]cnn)[.]com') # cnn_url_regex is a regular expression object

w = WebCrawler(cnn_url_regex)

This particular regular expression says:

1) Find the first occurence of the string '.com'

2) Then looking backwards from where '.com' was found it attempts to find '.cnn'

Why do this?

You can control where the crawler crawls. In this case I am constraining the crawler to operate on webpages within cnn.com.

Another feature I added was the ability to parse a given page looking for specific html tags. I chose as an example the <h1> tag. Once a <h1> tag is found I store all the words I find in the tag in a dictionary that gets associated with the page url.

Why do this?

My thought was that if I scraped the page for text I could eventually use this data for a search engine request. Say I searched for 'Lebron James'. And suppose that one of the pages my crawler scraped found an article that mentions Lebron James many times. In response to a search request I could return the link with the Lebron James article in it.

The web crawler is described in the WebCrawler class. It has 2 functions the user should call:

1) start_crawling(seed_pages,depth)

2) print_all_page_text() # this is only used for debug purposes

The rest of WebCrawler's functions are internal functions that should not be called by the user (think private in C++).

Upon construction of a WebCrawler object, it creates a MyHTMLParser object. The MyHTMLParser class inherits from the built-in Python class HTMLParser. I use the MyHTMLParser object when searching for the <h1> tag. The MyHTMLParser class creates instances of a helper class named Tag. The tag class is used in creating a "linked list" of tags.

So to get started with WebCrawler make sure to use Python 2.7.2. Enter the code a piece at a time into IDLE in the order displayed below. This ensures that you import libs before you start using them.

Once you have entered all the code into IDLE, you can start crawling the 'interwebs' by entering the following:

import re

cnn_url_regex = re.compile('(?<=[.]cnn)[.]com')

w = WebCrawler(cnn_url_regex)

w.start_crawling(['http://www.cnn.com/2012/02/24/world/americas/haiti-pm-resigns/index.html?hpt=hp_t3'],1)

Of course you can enter any page you want. But the regular expression object is already setup to filter on cnn.com. Remember the second parameter passed into the start_crawling function is the recursion depth.

Happy Crawling!

Python, 377 lines
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
class Tag:
    name = '';
    text = '';
    first_child = 0;
    parent = 0;
    next_sibling = 0;
    closed = 0;
    depth = 0;
    def get_tag_info_str(self):
        c,p,s = 'none','none','none'
        if self.first_child != 0:
            c = self.first_child.name
        if self.parent != 0:
            p = self.parent.name
        if self.next_sibling != 0:
            s = self.next_sibling.name
        return "name = {}, text = {}\nParent = {}, First Child = {}, Next Sibling = {}\nClosed = {}, Depth = {}\n".format(self.name, self.text, p, c, s, self.closed, self.depth)
      
      
from HTMLParser import HTMLParser
from htmlentitydefs import name2codepoint

class MyHTMLParser(HTMLParser):
    tag_list = []
    depth = 0;
    previous_tag = 'none';
    mode = 'silent';
  
  
    def handle_starttag(self, tag, attrs):
        if self.mode != 'silent':
            print "Start tag:", tag
            for attr in attrs:
                print "     attr:", attr
        self.depth = self.depth + 1
        t = Tag()
        t.name = tag
        t.depth = self.depth
        if self.previous_tag == 'start':
            # current tag is a first child of the last tag
            t.parent = self.tag_list[len(self.tag_list)-1]
            self.tag_list[len(self.tag_list)-1].first_child = t
        elif self.previous_tag == 'end':
            # current tag is next sibling of the last tag
          
            for x in reversed(self.tag_list):
                if x.depth == self.depth:
                    x.next_sibling = t          
                    if t.parent == 0:
                        t.parent = x.parent
                    break
        elif self.previous_tag == 'startend':
            # current tag is the next sibling of the previous tag
            t.parent = self.tag_list[len(self.tag_list)-1].parent
            self.tag_list[len(self.tag_list)-1].next_sibling = t
          
      
        self.tag_list.append(t)  
        self.previous_tag = 'start'
    def handle_endtag(self, tag):
        if self.mode != 'silent':
            print "End tag  :", tag
        for x in reversed(self.tag_list):
            if x.name == tag and x.closed == 0:
                x.closed = 1
                break
        self.depth = self.depth - 1
        self.previous_tag = 'end'
      
    def handle_startendtag(self, tag, attrs):
        if self.mode != 'silent':
            print "Start/End tag  :", tag
            for attr in attrs:
                print "     attr:", attr
        t = Tag()
        self.depth = self.depth + 1
        t.name = tag
        t.depth = self.depth
        t.closed = 1
          
        if self.previous_tag == 'start':
            # current tag is first child of the last tag
            t.parent = self.tag_list[len(self.tag_list)-1]
            self.tag_list[len(self.tag_list)-1].first_child = t
        elif self.previous_tag == 'startend':          
            # current tag is next sibling of last tag
            t.parent = self.tag_list[len(self.tag_list)-1].parent
            self.tag_list[len(self.tag_list)-1].next_sibling = t
        elif self.previous_tag == 'end':          
            # current tag is next sibling of a previous tag of depth = self.depth
            for x in reversed(self.tag_list):
                if x.depth == self.depth:
                    x.next_sibling = t          
                    if t.parent == 0:
                        t.parent = x.parent
                    break
          
        self.tag_list.append(t)  
        self.depth = self.depth - 1
        self.previous_tag = 'startend'
      
      
    def handle_data(self, data):
        if self.mode != 'silent':
            print "Data     :", data
      
        self.depth = self.depth + 1
      
        # add data to last tag in list with depth = current depth - 1
        for x in reversed(self.tag_list):
            if x.depth == self.depth - 1:
                x.text = (x.text + ' ' + data.strip(' \n\t')).strip(' \n\t')
                break
              
        self.depth = self.depth - 1
      
    def handle_comment(self, data):
        if self.mode != 'silent':
            print "Comment  :", data
    def handle_entityref(self, name):
        if self.mode != 'silent':
            c = unichr(name2codepoint[name])
            print "Named ent:", c
    def handle_charref(self, name):
        if self.mode != 'silent':
            if name.startswith('x'):
                c = unichr(int(name[1:], 16))
            else:
                c = unichr(int(name))
            print "Num ent  :", c
    def handle_decl(self, data):
        if self.mode != 'silent':
            print "Decl     :", data
      
    def print_tag_list(self, u):
        for l in self.tag_list:
            print l.get_tag_info_str()
          
    def clear_tag_list(self):
        self.tag_list.__delslice__(0,len(self.tag_list))
    
    def pretty_print_tags(self):
        for t in self.tag_list:
            s = ''
            s = s + self.get_indent_str(t.depth-1)
            s = s + self.get_tag_str(t.name)
            print s

    def get_indent_str(self, n):
        s = ''
        while(n != 0):
            s = s + '    '
            n = n - 1          
        return s
          
    def get_tag_str(self, name):
        return '<{}>'.format(name)
      
    def find_first_tag(self, name):
        r = 0
        for t in self.tag_list:
            if t.name == name:
                r = t
                break
        return r
      
    def print_first_tag_info(self, name):
        t = self.find_first_tag(name)
        if t == 0:
            print "Tag: {} not found".format(name)
        else:
            print t.get_tag_info_str()











import urllib
import socket
socket.setdefaulttimeout(10)
import httplib

class WebCrawler:
    """A simple web crawler"""
      
    link_dict = {};
    initial_depth = 0;
    #filter_list = [];
    parser = 0;
    re_compiled_obj = 0;
    
    class PageInfo:
        """ i store info about a webpage here """
        has_been_scraped = 0;
        word_dict = {};
                 
          
    def __init__(self,re_compiled_obj):      
        #self.filter_list.append(self.Filter(1,'.cnn.'))
        self.parser = MyHTMLParser()
        self.re_compiled_obj = re_compiled_obj
          
    def get_page(self,url):
        """ loads a webpage into a string """
        page = ''
        try:
            f = urllib.urlopen(url=url)
            page = f.read()
            f.close()
        except IOError:
            print "Error opening {}".format(url)
        except httplib.InvalidURL, e:
            print "{} caused an Invalid URL error.".format(url)
            if hasattr(e, 'reason'):
                print 'We failed to reach a server.'
                print 'Reason: ', e.reason
            elif hasattr(e, 'code'):
                print 'The server couldn\'t fulfill the request.'
                print 'Error code: ', e.code        
                
        return page
  
    def check_filters(self,url):
        """ If the url_str matches any of the
        enabled filter strings
        then put the url in the dictionary """
        
        
        match = self.re_compiled_obj.search(url)
        #print "match = {}".format(match)
        return match
  
      
    def find_h1_tag(self,s,pos):
        """ finds the first <h1> tag """
        start = s.find('<h1>',pos)
        end = s.find('</h1>',start)
        return start, end

    def save_tag_text(self, tag, d):
        """ stores each word in the tag in a dictionary """
        if tag != 0:
            token_list = tag.text.split(' ')
            for token in token_list:
                #print 'token = {}'.format(token)
                if d.has_key(token):
                    d[token] = d[token] + 1
                else:
                    d[token] = 1
        return d
      
    def save_page_text(self,page_str):
        """ Save all important text on the page """
        offset = 0
        d = {}
      
        while offset != -1:
            start,end = self.find_h1_tag(page_str,offset)
            offset = end
      
            if start != -1 and end != -1:
                h1_tag = page_str[start:end+5]
                #print h1_tag
                self.parser.clear_tag_list()
                # turn text into linked list of tags
                # only feed part of the page into the parser
                self.parser.feed(h1_tag)                                 
                #self.parser.pretty_print_tags()
                tag = self.parser.find_first_tag('h1')
                # add words from tag into the dictionary
                d = self.save_tag_text(tag,d)
        return d
      
    def save_all_links_on_page(self,page_str,limit=60):
        """ Stores all links found on the current page in a dictionary """
        d = {}
        offset = 0
        i = 0
        num_pages_filtered = 0
        num_duplicate_pages = 0
        while offset != -1:
            if i == limit:
                break
            offset = page_str.find('<a href="http',offset)
            if offset != -1:
                start = page_str.find('"', offset)
                end = page_str.find('"',start+1)
                link = page_str[start+1:end]
                # don't just save all the links
                # filter the links that match specified criteria
                if self.check_filters(link):
                    if link not in self.link_dict:
                        # adding link to global dictionary
                        self.link_dict[link] = self.PageInfo()
                        # adding link to local dictionary
                        d[link] = self.PageInfo()
                    else:
                        num_duplicate_pages = num_duplicate_pages + 1
                else:
                    num_pages_filtered = num_pages_filtered + 1
                offset = offset + 1
            i = i + 1
        print "{} out of {} links were filtered".format(num_pages_filtered,i)
        print "{} out of {} links were duplicates".format(num_duplicate_pages,i)
        #print "{} links are being returned from save_all_links".format(len(d))
        return d
  
  
  
  
    def save_all_links_recursive(self,links,depth):
        """ Recursive function that
            1) converts each page (link) into a string
            2) stores all links found in a dictionary """
        d = {}
      
        print "We are {} levels deep".format(self.initial_depth - depth)
      
        if depth != 0:
            depth = depth - 1
            urls = links.viewkeys()
            #print "There are {} urls".format(len(urls))
            for url in urls:
                print "trying to get {} over the internet".format(url)
                page_str = self.get_page(url)
                print "done getting {} over the internet".format(url)
                self.link_dict[url].word_dict = self.save_page_text(page_str)
                d = self.save_all_links_on_page(page_str)
                self.link_dict[url].has_been_scraped = 1
                # d contains all the links found on the current page
                self.save_all_links_recursive(d,depth)

    def start_crawling(self,seed_pages,depth):
        """ User calls this function to start crawling the web """
        d = {}
        self.link_dict.clear()
       
        # initialize global dictionary variable to the seed page url's passed in
        for page in seed_pages:           
            self.link_dict[page] = self.PageInfo()
            d[page] = self.PageInfo()
        self.initial_depth = depth
        # start a recursive crawl
        # can't pass in self.link_dict because then i get a RuntimeError: dictionary changed size during iteration
        self.save_all_links_recursive(d,depth)
      
    def print_all_page_text(self):
        """ prints contents of all the word dictionaries """
        for i in range(len(self.link_dict)):
            page_info = self.link_dict.values()[i]
            url = self.link_dict.keys()[i]
            print 'url = {}, has_been_scraped = {}'.format(url,page_info.has_been_scraped)
            d = page_info.word_dict
            for j in range(len(d)):
                word = d.keys()[j]
                count = d.values()[j]
                print '{} was found {} times'.format(word,count)
      
import re
                
cnn_url_regex = re.compile('(?<=[.]cnn)[.]com')                


# (?<=[.]cnn)[.]com regular expression does the following:
# 1) match '.com' exactly
# 2) then looking backwards from where '.com' was found it attempts to find '.cnn'


                
w = WebCrawler(cnn_url_regex)
w.start_crawling(['http://www.cnn.com/2012/02/24/world/americas/haiti-pm-resigns/index.html?hpt=hp_t3'],1)

I make no guarantees that the code is 100% bug free. I have run it on real live web sites and I have tried to fix any bugs I run into. I have by no means exhaustively tested this code.

The code is just a simple framework. You might say this code doesn't do much. Well it is up to you to make it do something special.

I am not a code guru. My coding style is all about "getting the job done" and not elegant or "pythonic". If any of you more experienced coders have critiques please comment. I too hope to learn from you.

2 comments

James Mills 12 years ago  # | flag

http://code.activestate.com/recipes/576551-simple-web-crawler/?in=user-4167757

This is my implementation and works quite well. :)

cheers James

Suresh 11 years ago  # | flag

Hey Bro.... can u plzzz upload a video or anything else u like explaining how the code works.... would be a great help ..........