return from WebPage to indicate whether a link was actually crawlable and only actually crawl it if it was

README.md

@@ -1 +1,21 @@
-# Concurrent web scraper
+# Concurrent web scraper
+
+## Requirements
+
+This crawler was written in 3.7.0 to take advantage of the latest `asyncio` features.
+
+Install required modules:
+
+```bash
+pip install -r requirements.txt
+```
+
+Run:
+
+```bash
+python crawler.py -u https://urltocrawl.com
+```
+
+## Results
+
+The resulting sitemap will be output in the root of this directory as `sitemap.html`

crawler.py

@@ -5,13 +5,15 @@ Need a docstring.
 
 import argparse
 import jinja2
+import os
+import asyncio
 from datetime import datetime
 from utils.helpers import (UrlPool, WebPage, RobotsTxt, sanitise_url)
 
 
 def init_crawler(base_url=None, robots=None):
     '''
-    needs a docstring
+    Initialises the crawler by running the initial URL.
     '''
     uncrawled_urls, crawled_urls = UrlPool(), UrlPool()
     initial_page = WebPage(url=base_url, base_url=base_url, robots=robots)
@@ -48,7 +50,7 @@ def init_crawler(base_url=None, robots=None):
 
 def process_pool(base_url=None, uncrawled_urls=None, crawled_urls=None, robots=None):
     '''
-    Needs a docstring
+    Iterates over the pool of URLs and adds any discovered URLs.
     '''
     while uncrawled_urls.pool:
         # pop url from pool
@@ -56,16 +58,21 @@ def process_pool(base_url=None, uncrawled_urls=None, crawled_urls=None, robots=N
         # create a WebPage object for the URL
         current_page = WebPage(url=new_url, base_url=base_url, robots=robots)
         try:
-            current_page.run()
-            _urls = current_page.list_urls()
-            crawled_urls.add_to_pool(new_url)
+            succeeded = current_page.run()
         except Exception as e:
             print(e)
 
-        for url in _urls:
-            sanitised_url = sanitise_url(url=url)
-            if sanitised_url not in crawled_urls.pool:
-                uncrawled_urls.add_to_pool(url)
+        if succeeded:
+            _urls = current_page.list_urls()
+            crawled_urls.add_to_pool(new_url)
+
+            for url in _urls:
+                sanitised_url = sanitise_url(url=url)
+                if sanitised_url not in crawled_urls.pool:
+                    uncrawled_urls.add_to_pool(url)
+
+        print('{0} URLs crawled, {1} remaining'.format(len(crawled_urls.pool),
+                len(uncrawled_urls.pool)))
 
 
 def render_sitemap(base_url=None, crawled_urls=None, runtime=None):
@@ -84,7 +91,7 @@ def render_sitemap(base_url=None, crawled_urls=None, runtime=None):
     with open('sitemap.html', 'w') as outfile:
         outfile.write(rendered_html)
 
-    print('Sitemap available at sitemap.html')
+    print('Sitemap available at {0}/sitemap.html'.format(os.getcwd()))
 
 
 def run(args=None):
@@ -103,9 +110,6 @@ def run(args=None):
 
     render_sitemap(base_url=base_url, crawled_urls=crawled_urls.pool, runtime=runtime)
 
-    # pprint(crawled_urls.pool)
-    # print('{0} URLs crawled'.format(len(crawled_urls.pool)))
-
 
 if __name__ == '__main__':
 

notes.md

@@ -1,12 +1,56 @@
 ## Thoughts
 
   * ~~strip hashes and everything following (as they're in-page anchors)~~
-  * strip args
+  * ~~strip args~~
   * ~~use `pop()` on the set instead of `.remove()`~~
     * ~~return false once the set is empty~~
   * ~~`WebPage.parse_urls()` needs to compare startswith to base url~~
   * ~~ignore any links which aren't to pages~~
-  * better url checking to get bare domain
+  * ~~better url checking to get bare domain~~ #wontfix
+  * ~~remove trailing slash from any discovered url~~
+  * ~~investigate lxml parser~~
   * ~~remove base url from initial urls with and without trailing slash~~
-  * investigate using [tldextract](https://github.com/john-kurkowski/tldextract) to match urls
+  * ~~investigate using [tldextract](https://github.com/john-kurkowski/tldextract) to match urls~~ #wontfix
   * ~~implement parsing of [robots.txt](http://docs.w3cub.com/python~3.6/library/urllib.robotparser/)~~
+  * ~~investigate [gzip encoding](https://stackoverflow.com/questions/36383227/avoid-downloading-images-using-beautifulsoup-and-urllib-request)~~
+  * ~~implement some kind of progress display~~
+  * async
+  * better exception handling
+  * randomise output filename
+
+### Async bits
+
+in `__main__`:
+
+```python
+loop = asyncio.get_event_loop()
+try:
+    loop.run_until_complete(main())
+finally:
+    loop.close()
+```
+
+  * initialises loop and runs it to completion
+  * needs to handle errors (try/except/finally)
+
+```python
+async def run(args=None):
+    tasks = []
+
+    for url in pool:
+        tasks.append(url)
+    # for i in range(10):
+        # tasks.append(asyncio.ensure_future(myCoroutine(i)))
+
+    # gather completed tasks
+    await asyncio.gather(*tasks)
+```
+
+Getting the contents of the page needs to be async too
+
+```python
+async def get_source():
+    blah
+    blah
+    await urlopen(url)
+```

requirements.txt

@@ -4,4 +4,5 @@ certifi==2018.8.13
 chardet==3.0.4
 idna==2.7
 Jinja2==2.10
+lxml==4.2.4
 MarkupSafe==1.0

test_helpers.py

@@ -1,37 +1,35 @@
 #!/usr/bin/env python
 
 import unittest
-from utils.helpers import (clean_base_url)
+from utils.helpers import (sanitise_url)
 
 class TestUrls(unittest.TestCase):
 
-    base_url = "github.com"
-
     base_url_list = (('eu.httpbin.org', 'http://eu.httpbin.org'),
                      ('www.simonweald.com', 'http://www.simonweald.com'),
                      ('http://www.github.com/', 'http://www.github.com'),
                      ('https://www.github.com', 'https://www.github.com'))
 
-    valid_urls = ["https://www.github.com", "http://www.github.com",
-                  "github.com", "/some/url/", "index.html"]
+    urls_to_clean = (('https://www.github.com/', 'https://www.github.com/'),
+                     ('https://github.com/?foo=bar', 'https://github.com/'),
+                     ('https://github.com/#anchor', 'https://github.com/'))
 
 
-    def test_clean_base_url(self):
+    def test_sanitise_base_url(self):
         '''
         Tests whether a URL's protocol can be discovered if not provided.
         '''
         for url, target in self.base_url_list:
-            result = clean_base_url(url)
+            result = sanitise_url(url, base_url=True)
             self.assertEqual(result, target)
 
-    # def test_url_validation(self):
-    #     '''
-    #     Passes when given a valid URL. A valid URL is qualified
-    #     by being local to the domain to be crawled.
-    #     '''
-    #     for url in self.valid_urls:
-    #         result = url_validation(self.base_url, url)
-    #         self.assertTrue(result)
+    def test_sanitise_url(self):
+        '''
+        Tests whether a URL's protocol can be discovered if not provided.
+        '''
+        for url, target in self.urls_to_clean:
+            result = sanitise_url(url)
+            self.assertEqual(result, target)
 
 
 if __name__ == '__main__':

utils/helpers.py

@@ -4,8 +4,11 @@ Utilities to provide various misc functions.
 '''
 
 from bs4 import BeautifulSoup
+import aiohttp
 import urllib.request
 import urllib.robotparser
+import urllib.error
+import gzip
 from urllib.parse import (urljoin, urlsplit)
 
 
@@ -46,12 +49,16 @@ class WebPage(object):
     the data from each individual page.
     '''
 
-    headers = {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:61.0) Gecko/20100101 Firefox/61.0'}
+    # set a sane user-agent and request compression if available.
+    headers = {'Accept-Encoding': 'gzip, deflate',
+               'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:61.0) Gecko/20100101 Firefox/61.0'}
 
     def __init__(self, url=None, base_url=None, robots=None):
         self.url = url
         self.base_url = base_url
         self.robots = robots
+        self.source = None
+        self.urls_to_crawl = set()
 
 
     def get_source(self):
@@ -61,9 +68,16 @@ class WebPage(object):
 
         request = urllib.request.Request(self.url, headers=self.headers)
         page = urllib.request.urlopen(request, timeout=5)
-        headers = page.info()
-        if "text/html" in headers['content-type']:
-            self.source = page.read()
+
+        # handle the content encoding in case it needs decompressing.
+        if 'text/html' in page.info().get('Content-Type'):
+            if page.info().get('Content-Encoding'):
+                if page.info().get('Content-Encoding') == 'gzip':
+                    self.source = gzip.decompress(page.read())
+                elif page.info().get('Content-Encoding') == 'deflate':
+                    self.source = page.read()
+            else:
+                self.source = page.read()
 
 
     def find_links(self):
@@ -73,7 +87,7 @@ class WebPage(object):
         '''
         hrefs = set()
 
-        soup = BeautifulSoup(self.source, 'html.parser')
+        soup = BeautifulSoup(self.source, 'lxml')
         links = soup.find_all('a', href=True)
 
         for link in links:
@@ -90,38 +104,37 @@ class WebPage(object):
         Iterate through the list of discovered URLs and add them to the
         pool if they start with the base URL.
         '''
-        self.urls_to_crawl = set()
-
         for url in self.discovered_hrefs:
-            if url.startswith(self.url):
-                if self.robots.check(url):
-                    sanitised_url = sanitise_url(url=url)
-                    self.urls_to_crawl.add(sanitised_url)
+            if url.startswith(self.base_url) and self.robots.check(url):
+                sanitised_url = sanitise_url(url=url)
+                self.urls_to_crawl.add(sanitised_url)
 
 
     def list_urls(self):
         '''
-        Returns the contents of the
+        Returns all valid discovered URLs.
         '''
 
         return self.urls_to_crawl
 
 
     def run(self):
+        '''
+        Attempt to get the page's source and if successful, iterate through it
+        to find any links we can crawl.
+        '''
         try:
             self.get_source()
-        except Exception as e:
-            print(e)
+        except Exception:
+            # skip if we didn't retrieve the source.
+            pass
 
-        try:
+        if self.source:
             self.find_links()
-        except Exception as e:
-            print(e)
-
-        try:
             self.parse_urls()
-        except Exception as e:
-            print(e)
+            return True
+        else:
+            return False
 
 
 class RobotsTxt(object):
@@ -130,21 +143,33 @@ class RobotsTxt(object):
     '''
 
     def __init__(self, base_url=None):
+        '''
+        Manually retrieve robots.txt to allow us to set the user-agent.
+        '''
         self.base_url = base_url
+        self.headers = {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:61.0) Gecko/20100101 Firefox/61.0'}
+
+        robots_url = urljoin(self.base_url, 'robots.txt')
+        request = urllib.request.Request(robots_url, headers=self.headers)
 
         robots = urllib.robotparser.RobotFileParser()
-        robots.set_url(urljoin(self.base_url, 'robots.txt'))
+        robots.set_url(robots_url)
+
         try:
-            robots.read()
-        except Exception as e:
-            print(e)
+            response = urllib.request.urlopen(request, timeout=5)
+        except urllib.error.HTTPError:
+            robots.allow_all = True
+        else:
+            data = response.read()
+            decoded_data = data.decode("utf-8").splitlines()
+            robots.parse(decoded_data)
 
         self.robots = robots
 
 
     def check(self, url):
         '''
-        needs a docstring
+        Test if robots allows us to crawl that URL.
         '''
         return self.robots.can_fetch("*", url)