def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for link in response.xpath(
"//div[@class='view-content']/div[contains(@class,'views-row')]/div/h3/a"
):
text = link.xpath("text()").get()
url = link.xpath("@href").get()
self.log("find technology {}/{}".format(text, url),
level=logging.INFO)
patent_links.append(url)
# for next page
next_page = response.xpath("//li[@class='pager-next']/a/@href").get()
if next_page is not None:
self.log('process page {}'.format(next_page), level=logging.INFO)
yield response.follow(
url=next_page,
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
for p in patent_links:
name = self.parse_name_from_url(p)
if os.path.exists(os.path.join(self.work_directory,
name + '.json')):
self.log('{} already parsed and will skip'.format(p),
level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for link in response.xpath(
'//*[@id="formTechPub1"]/div/table[2]/tr/td/a'):
text = link.xpath("text()").get()
url = link.xpath("@href").get()
self.log("find technology {}/{}".format(text, url),
level=logging.INFO)
patent_links.append(url)
# for next page
total_result = self.statictics(response)
self.page += 1
if self.page * self.item_per_page < total_result:
self.log('process page {}'.format(self.page), level=logging.INFO)
yield response.follow(
url=self.next_page_template.format(self.page),
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
for p in patent_links:
name = self.parse_name_from_url(p)
if os.path.exists(os.path.join(self.work_directory,
name + '.json')):
self.log('{} already parsed and will skip'.format(p),
level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for link in response.xpath("//a[@class='lead-in']/@href").getall():
patent_links.append(link)
# for next page
next_page = response.xpath(
"//div[@class='nav-previous']/a/@href").get()
if next_page is not None:
self.log('process page {}'.format(next_page), level=logging.INFO)
yield response.follow(
url=next_page,
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
for p in patent_links:
name = p.split('/')[-2]
if os.path.exists(os.path.join(self.work_directory,
name + '.json')):
self.log('{} already parsed and will skip'.format(p),
level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for link in response.xpath("//h4[@class='result-title']/a"):
text = link.xpath("text()").get()
url = link.xpath("@href").get()
self.log("find technology {}/{}".format(text, url), level=logging.INFO)
patent_links.append(url)
# for next page
current_page, total_page = self.statictics(response)
if current_page < total_page:
self.log('process page {}'.format(self.page), level=logging.INFO)
yield response.follow(
url='https://otd.harvard.edu/explore-innovation/technologies/results/P{}/'.format(current_page * 10),
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
for p in patent_links:
name = self.parse_name_from_url(p)
if os.path.exists(os.path.join(self.work_directory, name + '.json')):
self.log('{} already parsed and will skip'.format(p), level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def run(self):
"""
The main function
"""
while self.running.is_set():
proxy = requests.get(self.api_url).content.decode().split(
self.token)
logging.debug('add new proxy {} to pool'.format(proxy))
for p in proxy:
POOL.add('http://' + p[:-1], 300)
time.sleep(self.delay_second)
def extract_page(self, response):
arguments = self.parse_url(response.request.url)
self.log('Process page {}'.format(response.request.url),
level=logging.INFO)
with open(
os.path.join(
self.work_directory,
'{}_{}_{}_{}.html'.format(arguments['province'][0],
arguments['round'][0],
arguments['industry'][0],
arguments.get('p', ['1'])[0])),
'wb') as fo:
fo.write(response.body)
row = response.xpath("//table[@class='ntable']/tr")
if len(row) < 1:
self.log('page {} has no data'.format(response.request.url),
level=logging.WARNING)
POOL.remove(response.request.meta['proxy'])
return
first_row = True
result = []
for r in row:
if first_row:
first_row = False
continue
image = r.xpath('td[1]/img/@src').extract_first()
project_name = r.xpath('td[2]/a/text()').extract_first()
project_url = r.xpath('td[2]/a/@src').extract_first()
investor = r.xpath('td[3]/text()').extract_first()
stage = r.xpath('td[4]/text()').extract_first()
invest_time = r.xpath('td[5]/text()').extract_first()
result.append({
'image': image,
'project': {
'url': project_url,
'name': project_name
},
'investor': investor,
'stage': stage,
'time': invest_time
})
self.log('find investment from {} to {}'.format(
investor, project_name),
level=logging.INFO)
# save the file
with open(
os.path.join(
self.work_directory,
'{}_{}_{}_{}.json'.format(arguments['province'][0],
arguments['round'][0],
arguments['industry'][0],
arguments.get('p', ['1'])[0])),
'w') as fo:
json.dump(result, fo, ensure_ascii=False)
def start_requests(self):
for url in self.start_urls:
for industry in self.industry:
for stage in self.stage:
for year in self.year:
yield scrapy.Request(
url=url,
dont_filter=True,
callback=self.apply_filter,
meta={
'proxy': POOL.get() if not self.exclusive else POOL.pop(),
'extra': {'industry': industry, 'stage': stage, 'year': year}},
errback=self.handle_failure)
def parse_list(self, response: Response):
driver = self.get_driver(response)
# wait for page to load
# wait for the redirect to finish.
patent_links = []
if os.path.exists(os.path.join(self.work_directory, 'links.json')):
patent_links = json.load(open(os.path.join(self.work_directory, 'links.json'), 'r'))
else:
while True:
time.sleep(1)
self.wait_for_element(driver, "//div[@id='ctl00_ContentPlaceHolder1_UpdateProgress1' and @style='display: block;']")
table = driver.find_element_by_xpath("//div[@class='table-body']")
for r in table.find_elements_by_xpath("div"):
cols = r.find_elements_by_xpath("div")
patent = cols[2].find_element_by_xpath('p/a')
abstract = cols[2].find_element_by_xpath('div/p')
patent_links.append({'name': patent.text, 'link': patent.get_attribute('href'), 'abstract': abstract.text})
self.log('Found technology {}'.format(patent.text), level=logging.INFO)
if not self.next_page(driver):
break
time.sleep(3)
with open(os.path.join(self.work_directory, 'links.json'), 'w') as fo:
json.dump(patent_links, fo)
for p in patent_links:
name = p['link'].split('/')[-1]
if os.path.exists(os.path.join(self.work_directory, name[:-4] + 'json')):
self.log('{} already parsed and will skip'.format(p['link']), level=logging.INFO)
continue
yield Request(
url=p['link'],
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for link in response.xpath("//ul[@id='tech-licensing']/li/a"):
text = link.xpath("text()").get()
url = link.xpath("@href").get()
if url is None:
continue
self.log("find technology {}/{}".format(text, url),
level=logging.INFO)
patent_links.append(url)
for p in patent_links:
name = self.parse_name_from_url(p)
if os.path.exists(os.path.join(self.work_directory,
name + '.json')):
self.log('{} already parsed and will skip'.format(p),
level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def start_requests(self):
for url in self.start_urls:
yield SeleniumRequest(url=url,
dont_filter=True,
callback=self.apply_filter,
meta={'proxy': POOL.get()},
errback=self.handle_failure_selenium)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
if os.path.exists(os.path.join(self.work_directory, 'links.json')):
patent_links = json.load(open(os.path.join(self.work_directory, 'links.json'), 'r'))
else:
# the id of product is provded in the <script></script>
for code in response.xpath("//script").getall():
if 'id_list' in code:
ids = re.findall(r'[0-9]+', re.findall(r'\[[0-9,]+\]', code)[0])
patent_links = [response.url + '/public/project/{}'.format(patentId) for patentId in ids]
with open(os.path.join(self.work_directory, 'links.json'), 'w') as fo:
json.dump(patent_links, fo)
for p in patent_links:
name = p.split('/')[-1]
if os.path.exists(os.path.join(self.work_directory, name + '.json')):
self.log('{} already parsed and will skip'.format(p), level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def start_requests(self):
for url in self.start_urls:
yield Request(
url=url,
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def parse(self, response):
# find the document url
link = response.xpath(
"//div[@class='ProjetInfo_title']/a/@href").extract_first()
if link is None:
self.log('{} fail to download'.format(response.url, link),
level=logging.WARNING)
# remove the invalid proxy
POOL.remove(response.request.meta['proxy'])
return
page = uuid.uuid5(uuid.NAMESPACE_URL, response.url.encode('utf-8')).hex
filename = '%s.html' % page
with open(os.path.join(self.work_directory, filename), 'wb') as f:
f.write(response.body)
self.log('{} => {}'.format(response.url, link), level=logging.INFO)
yield {
'link': response.url,
'xml': link,
}
def apply_filter(self, response):
for year in range(2018, 2003, -1):
url = response.request.url + 'y{}/'.format(year)
self.log('Process page {}'.format(url), level=logging.INFO)
yield scrapy.Request(
url=url,
dont_filter=True,
callback=self.parse,
meta={'proxy': POOL.get()},
errback=self.handle_failure)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for row in response.xpath(
"//div[@id='nouvant-portfolio-content']/div[@class='technology']"
):
title = row.xpath("h2/a/text()").get()
link = row.xpath("h2/a/@href").get()
abstract = row.xpath("p/span/text()").get()
self.log('found patent {}'.format(title), level=logging.INFO)
patent_links.append({
'title': title,
'link': link,
'abstract': abstract
})
statistics = self.statistics(response)
self.log('found {}/{} patents'.format(statistics['end'],
statistics['total']),
level=logging.INFO)
if statistics['end'] < statistics['total']:
yield response.follow(
url='/technologies?limit=50&offset={}&query='.format(
statistics['end']),
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
for p in patent_links:
name = p['link'].split('/')[-1]
if os.path.exists(os.path.join(self.work_directory,
name + '.json')):
self.log('{} already parsed and will skip'.format(p['link']),
level=logging.INFO)
continue
yield response.follow(
url=p['link'],
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def start_requests(self):
for url in self.data.keys():
if url is None or not url.startswith('http'):
continue
name = url.split('/')[-1]
if os.path.exists(os.path.join(self.work_directory,
name + '.json')):
continue
yield Request(url=url,
meta={'proxy': POOL.get()},
errback=self.handle_failure,
callback=self.parse)
def parse_category(self, response: Response):
# with javascript it would be //div[@class='split-taxonomy-4']/ul/li/a/@href
for row in response.xpath(
"//section[@id='block-taxonomy-menu-block-1']/ul/li/a/@href"
).getall():
self.log('find category {}'.format(row), level=logging.INFO)
yield response.follow(
url=row,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
callback=self.parse_list,
errback=self.handle_failure)
def parse_list(self, response: Response):
# wait for page to load
# wait for the redirect to finish.
patent_links = []
for row in response.xpath(
"//section[@id='block-system-main']/div[@class='node node-technology node-teaser clearfix']/h2/a"
):
name = row.xpath("text()").get()
link = row.xpath("@href").get()
patent_links.append({'name': name, 'link': link})
self.log('found patents {}'.format(name), level=logging.INFO)
if response.xpath("//li[@class='pager-last']/a/@href").get() is not None and\
response.url != response.xpath("//li[@class='pager-last']/a/@href").get():
# have next page
if '?page=' in response.url:
elements = response.url.split("=")
page = (int(elements[-1]) + 1)
self.log('go to page {}'.format(page), level=logging.INFO)
yield response.follow(
url='='.join(elements[:-1]) + '={}'.format(page),
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
callback=self.parse_list,
errback=self.handle_failure)
else:
self.log('go to page 2', level=logging.INFO)
yield response.follow(
url=response.url + '?page=1',
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
callback=self.parse_list,
errback=self.handle_failure)
for p in patent_links:
yield response.follow(
url=p['link'],
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
callback=self.parse,
errback=self.handle_failure)
def start_requests(self):
urls = json.load(
open(os.path.join(self.work_directory, 'links.json'), 'rb'))
for url in urls:
page = uuid.uuid5(uuid.NAMESPACE_URL, url.encode('utf-8')).hex
filename = '%s.html' % page
if os.path.exists(os.path.join(self.work_directory, filename)):
self.log('{} already exists'.format(url))
continue
yield scrapy.Request(url=url,
callback=self.parse,
meta={'proxy': POOL.get()},
errback=self.handle_failure)
time.sleep(random.random())
def parse_page(log, work_directory: str, with_proxy: bool, slug: str):
print('process {}'.format(slug))
while True:
try:
if with_proxy:
proxies = POOL.get()
response = get(slug, proxies={'http': proxies, 'https': proxies})
else:
response = get(slug)
if 200 <= response.status_code < 300:
data = response.json()
with open(os.path.join(work_directory, '{}.json'.format(slug.split('/')[-1])), 'w') as fo:
json.dump(data, fo)
break
except Exception as e:
print(e)
def parse_school_list(self, response: Response):
if os.path.exists(os.path.join(self.work_directory, 'links.json')):
school_links = json.load(open(os.path.join(self.work_directory, 'links.json'), 'r'))
else:
driver = self.get_driver(response)
school_links = []
page = 1
while True:
for row in driver.find_elements_by_xpath("//table[@id='search-results']/tbody/tr"):
title = row.find_element_by_xpath("td[@class='search-results-major']/a").text
link = row.find_element_by_xpath("td[@class='search-results-major']/a").get_attribute('href')
school_links.append({'name': title, 'link': link})
self.log('find school {}'.format(title), level=logging.INFO)
total_page = self.statistics(response)
self.log('Finish page {}/{}'.format(page, total_page), level=logging.INFO)
if page < total_page:
try:
next_page = driver.find_element_by_xpath("//img[@class='paginator-next-page paginator-button']")
if next_page is not None:
next_page.click()
page += 1
except Exception as e:
self.log('Fail to go to page {}'.format(page + 1), level=logging.ERROR)
break
else:
break
time.sleep(3)
wait = WebDriverWait(driver, 30)
try:
wait.until(lambda x: len(x.find_elements_by_xpath("//table[@id='search-results' and @style='opacity: 1;']")) > 0)
except Exception as e:
self.log('Unable to retrieve school information: {}'.format(e), level=logging.ERROR)
break
with open(os.path.join(self.work_directory, 'links.json'), 'w') as fo:
json.dump(school_links, fo)
for s in school_links:
yield Request(
url=s['link'],
callback=self.parse_list,
dont_filter=True,
meta={'proxy': POOL.get()} if self.with_proxy else {},
errback=self.handle_failure)
def parse_list(self, response):
self.log('Parse list {}'.format(response.url), level=logging.INFO)
name = response.url.split('/')[-1]
with open(os.path.join(self.work_directory, name + '.html'), 'wb') as fo:
fo.write(response.body)
# for the information of school
school = create_company()
meta = self.get_school_information(response)
if 'Name' in meta:
school['name'] = meta['Name']
if 'URL' in meta:
school['ref'] = meta['URL']
school['contact']['website'] = meta['URL']
if 'Group Type' in meta:
school['abs'] = meta['Group Type']
school['addr'] = deepcopy(self.address)
school['addr']['line1'] = school['name']
if school['name'] in self.blacklist:
return
patent_links = []
if os.path.exists(os.path.join(self.work_directory, school['name'] + '.json')):
patent_links = json.load(open(os.path.join(self.work_directory, school['name'] + '.json'), 'r'))
else:
# the id of product is provded in the <script></script>
for code in response.xpath("//script").getall():
if 'id_list' in code:
ids = re.findall(r'[0-9]+', re.findall(r'\[[0-9,]+\]', code)[0])
patent_links = ['https://www.flintbox.com/public/project/{}'.format(patentId) for patentId in ids]
with open(os.path.join(self.work_directory, school['name'] + '.json'), 'w') as fo:
json.dump(patent_links, fo)
for p in patent_links:
name = p.split('/')[-1]
if os.path.exists(os.path.join(self.work_directory, name + '.json')):
self.log('{} already parsed and will skip'.format(p), level=logging.INFO)
continue
yield response.follow(
url=p,
callback=self.parse,
dont_filter=True,
meta={'proxy': POOL.get(), 'school': school} if self.with_proxy else {'school': school},
errback=self.handle_failure)
def start_requests(self):
for url in self.start_urls:
for p in self.province:
for i in self.industry:
for r in self.round:
self.log(
'start page province={} industry={} and round={}'.
format(p, i, r),
level=logging.INFO)
yield scrapy.Request(url=url,
dont_filter=True,
callback=self.apply_filter,
meta={
'proxy': POOL.get(),
'extra': {
'province': p,
'industry': i,
'round': r
}
},
errback=self.handle_failure)
def apply_filter(self, response):
if os.path.exists(os.path.join(self.work_directory, 'category.json')):
with open(os.path.join(self.work_directory, 'category.json'), 'r') as fi:
result = json.load(fi)
links = []
for r1 in result:
for r2 in r1['children']:
if len(r2['children']) < 1:
links.append(r2['url'])
else:
links.extend([r3['url'] for r3 in r2['children']])
else:
links = self.find_filter(response)
for l in links:
url = l.split('/')
url[-2] = 'ma/' + url[-2]
l = '/'.join(url)
self.log('Process page {}'.format(l), level=logging.INFO)
yield scrapy.Request(
url=l,
dont_filter=True,
callback=self.parse,
meta={'proxy': POOL.get()},
errback=self.handle_failure)
def start_requests(self):
for url in self.start_urls:
yield SeleniumRequest(url,
callback=self.parse,
meta={'proxy': POOL.get()},
errback=self.handle_failure_selenium)
