def scrape_data(self): # Create boilerplate recognizer skindex = Scraper(self.crawl_config) skindex.load_local_pages() skindex.add_template_elements() # Process all by removing boilerplate and extracting information return skindex.process_all(exclude_data=['cleaned', 'author'])
def post(self): CRAWL_CONFIG = DEFAULT_CRAWL_CONFIG.copy() CRAWL_CONFIG.update({ 'collections_path': os.path.join(os.path.expanduser('~'), 'sky_view_collections/'), # 'max_workers': 10, }) args = self.request.arguments print(args) for arg in args: value = args[arg][0].decode('utf8') if value and arg != 'url' and arg != 'checkboxcache': print('pre', arg, CRAWL_CONFIG[arg]) if isinstance(CRAWL_CONFIG[arg], list): CRAWL_CONFIG[arg] = [int(value)] if is_numeric(value) else value.split(', ') else: CRAWL_CONFIG[arg] = int(value) if is_numeric(value) else value.split(', ')[0] print('post', arg, CRAWL_CONFIG[arg]) url = self.get_argument('url', '') use_cache = self.get_argument('checkboxcache', '') domain = extractDomain(url) CRAWL_CONFIG['seed_urls'] = [url] if domain.startswith("http"): CRAWL_CONFIG['collection_name'] = domain.split("/")[2] else: CRAWL_CONFIG['collection_name'] = domain.split("/")[0] if use_cache != 'on': col_path = os.path.join(CRAWL_CONFIG['collections_path'], CRAWL_CONFIG['collection_name']) print(col_path) if os.path.exists(col_path): shutil.rmtree(col_path) crawl.start(CRAWL_CONFIG) SCRAPE_CONFIG = CRAWL_CONFIG.copy() SCRAPE_CONFIG.update({ 'template_proportion': 0.4, 'max_templates': 100, }) skindex = Scraper(SCRAPE_CONFIG) skindex.load_local_pages() skindex.add_template_elements() res = skindex.process_all(remove_visuals=True, maxn=CRAWL_CONFIG['max_saved_responses']) items = [] for num, url in enumerate(res): if num == CRAWL_CONFIG['max_saved_responses']: break dc = res[url] dc['url'] = url dc['source_name'] = domain dc['images'] = [x for x in reversed(dc['images'][:5])] # dc['blobs'] = [TextBlob(x) for x in dc['body'] if dc['body']] items.append(dc) # this is quite out of place like this print('num unique images', len(get_image_set({x['url']: x for x in items}))) if items and 'money' in items[0]: items = sorted(items, key=lambda x: len(x['money']), reverse=True) self.render('page_template.html', items=items, cached=False)
# Crawling CRAWL_CONFIG = DEFAULT_CRAWL_CONFIG CRAWL_CONFIG.update({ 'seed_urls': ['http://www.techcrunch.com/'], 'collections_path': '/Users/pascal/egoroot/sky_collections', 'collection_name': 'techie', # Optional 'crawl_filter_regexps': [], 'crawl_required_regexps': ['2015', '2014'], 'index_filter_regexps': [], 'index_required_regexps': ['2015', '2014'], 'max_saved_responses': 100, 'max_workers': 10, }) crawl.start(CRAWL_CONFIG) # Indexing SCRAPE_CONFIG = CRAWL_CONFIG.copy() SCRAPE_CONFIG.update({'template_proportion': 0.09, 'max_templates': 1000}) skindex = Scraper(SCRAPE_CONFIG) skindex.load_local_pages() skindex.add_template_elements() res = skindex.process_all(remove_visuals=True)
def __init__(self, config, cache=None): super(NewsCrawler, self).__init__(config, cache) self.scraper = Scraper(config) self.template_complete = False self.data = {} self.templates_done = 0