def run(fname):
stopws = stopwords.words('english')
bf = pybloom_live.BloomFilter(capacity=100000, error_rate=0.001)
with open("data/" + fname, "r") as f:
start = time.time()
print("Starting {0}.".format(fname))
for i, l in enumerate(f):
reviews_dict = defaultdict(int)
scores_dict = defaultdict(float)
count_dict = defaultdict(int)
revs = json.loads(l)["allReviews"]
for rev in revs:
score = rev[0]
wlist = break_rev(rev[1], stopws, bf)
for w in wlist:
count = r.hget("#counts", w)
if count is None: continue
if int(count) < 1000 or int(count) > 7000: continue
reviews_dict[w] = reviews_dict[w] + 1
scores_dict[w] = scores_dict[w] + score
count_dict[w] = count_dict[w] + 1
for key in reviews_dict:
reviews_dict[key] = reviews_dict[key] / int(r.hget("#counts", key))
results = [(x[0], x[1], (scores_dict[x[0]] / count_dict[x[0]]), count_dict[x[0]]) for x in reviews_dict.items()]
results.sort(key=lambda x: x[1])
if 'title' in l:
print(json.loads(l)["title"])
print(list(map(lambda x: x[0], results[::-1][:10])))
if i % 100 == 0:
print("{2} {0} products, refreshing bf, took {1} seconds since last print".format(i, time.time() - start, fname))
bf = pybloom_live.BloomFilter(capacity=100000, error_rate=0.001)
start = time.time()
if i == 10: break
def initialize_bloom_filter(item, bloom_error_rate):
md5, tokens = item
bloom_filter = pybloom_live.BloomFilter(capacity=len(set(tokens)),
error_rate=bloom_error_rate)
for token in tokens:
bloom_filter.add(token)
return (md5, bloom_filter)
def queue_arrange():
m = 0
url_total = []
url_total = list_open.url_list_read()
url_next_num.put(url_total[len(url_total) - 1])
b = pybloom_live.BloomFilter(capacity=90000000, error_rate=0.01)
with open('urllist.txt', 'r') as file_object:
for line in file_object:
b.add(line.rstrip())
while 1:
try:
url_current = []
if not url_current_num.empty(
): #对url_total队列进行去重，并将url_current_num压入总列表和队列
while not url_current_num.empty():
url_current.append(url_current_num.get())
for num1 in url_current:
if num1 not in b:
#url_total.append(num1)
url_next_num.put(num1)
b.add(str(num1))
m = m + 1
print('url管理进程运行了%d次' % (m))
print('total列表有%d个元素' % (len(url_total)))
#b.sync()
#time.sleep(2)
except Exception:
continue
def run(fname):
out_file = "wordset/" + fname + "-wordset"
stopws = set(stopwords.words('english'))
bf = pybloom_live.BloomFilter(capacity=1000000, error_rate=0.001)
with open("data/" + fname, "r") as f:
with open(out_file, "w") as out_f:
# Timer
start = time.time()
print("Starting {0}.".format(fname))
tokenized_words = set()
for i, l in enumerate(f):
d = json.loads(l)
# Break reviews
revs = d["allReviews"]
for rev in revs:
for w in break_string(rev[1], stopws, bf):
tokenized_words.add(w)
# Break title
if "title" in d:
for w in break_string(d["title"], stopws, bf):
tokenized_words.add(w)
# Break desc
if "description" in d:
for w in break_string(d["description"], stopws, bf):
tokenized_words.add(w)
# Timer
if (i + 1) % 100 == 3:
bf = pybloom_live.BloomFilter(capacity=1000000,
error_rate=0.001)
print(
"{2} {0} products, refreshing bf, took {1} seconds since last print"
.format(i,
time.time() - start, fname))
start = time.time()
for w in tokenized_words:
out_f.write(w)
out_f.write("\n")
def load_bfobj(self):
"""load bloom file obj"""
files_list = os.walk(defaults.BLOOM_FILE_PATH)
files_list_str = str(list(files_list))
# bloom file not exist --> first run
if defaults.BLOOM_FILE_NAME not in files_list_str:
print('init')
return pybloom_live.BloomFilter(capacity=self.capacity,
error_rate=self.error_rate)
# bloom file exist
with open(defaults.BLOOM_FILE_PATH + defaults.BLOOM_FILE_NAME,
'rb') as fp:
print('load_bfobj',
defaults.BLOOM_FILE_PATH + defaults.BLOOM_FILE_NAME)
return pybloom_live.BloomFilter.fromfile(fp)
def __init__(self,
url,
deep=4,
thread_num=100,
url_beyond=100,
suspend=5,
timeout=5,
capacity=1000000,
error_rate=0.001):
# 爬虫起始url
self._url = url
if url.endswith('/'):
self._url = url[:-1]
# 顶级域名
self.top_domain = tldextract.extract(url).domain
# 所有爬取到url
self.__urls_dict = {}
# 临时存储URL数据 {'url链接':[此链接中提取到的URL]}
self._urls_dict_tmp = {}
self.urls_dict_tmp = url
# 爬虫深度
self.deep = deep
self.bloom = pybloom_live.BloomFilter(capacity=capacity,
error_rate=error_rate)
self._req_tds = []
# 超出个数时暂停
self.url_beyond = url_beyond
# 暂停时间
self.suspend = suspend
# 线程数
self.thread_num = thread_num
# 请求超时时间
self.timeout = timeout
class Vieclam24hQlSpider(scrapy.Spider):
name = 'vieclam24h_QL'
start_urls = [
'https://vieclam24h.vn/tim-kiem-viec-lam-nhanh/?hdn_nganh_nghe_cap1=&hdn_dia_diem=&hdn_tu_khoa=&hdn_hinh_thuc=&hdn_cap_bac=',
]
ur = pybloom_live.BloomFilter(capacity=2097152, error_rate=0.005)
collection_name = 'News'
client = pymongo.MongoClient(settings.MONGO_URI)
db = client[settings.MONGO_DATABASE]
collection = db[collection_name]
Y_in_db = list(
collection.find({}, {
"title": 1,
"company": 1,
"address": 1,
"_id": 0
}))
no_duplicate_items = 0
def parse(self, response):
for tn in response.xpath('//div[@class="list-items "]/div/div/span'):
src = tn.xpath('a/@href').extract_first()
src = response.urljoin(src)
add_url = self.ur.add(src)
if add_url is False:
yield scrapy.Request(src, callback=self.parse_src)
next_pages = response.xpath('//li[@class="next"]/a/@href').extract()
next_page = next_pages[len(next_pages) - 1]
if next_page is not None:
next_page = response.urljoin(next_page)
yield scrapy.Request(next_page, callback=self.parse)
def parse_src(self, response):
self.item = JobItem()
self.item["url"] = response.request.url
title = response.xpath(
'//div[@class="col-xs-12"]/h1[@class="text_blue font28 mb_10 mt_20 fws title_big"]/text()'
).extract()
x_title = title[0]
self.item["title"] = x_title
#Cong ty
company = response.xpath(
'//p[@class="font16"]//a[@class="text_grey3"]/text()').extract()
x_company = company[0]
self.item['company'] = x_company
#Noi lam viec
addresses = response.xpath(
'//span[@class="pl_28"]//a[@class="job_value text_pink"]/text()'
).extract()
address = ', '.join([
address.replace("Việc làm", "").replace("TP.HCM",
"Hồ Chí Minh").strip()
for address in addresses
])
self.item['address'] = address
#Check duplicate
data_need_check = DataReduction(
3, [[job['title'], job['company'], job['address']]
for job in self.Y_in_db])
if data_need_check.is_match([x_title, x_company, address]):
self.no_duplicate_items += 1
print(self.no_duplicate_items)
return
#Luong
salary = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Mức lương")]/span/text()'
).extract()
if len(salary) > 0:
salary_str = " ".join(salary)
salary_need_normalize = Normalize_salary()
salary_normalized = salary_need_normalize.normalize_salary(
salary_str)
self.item["salary"] = salary_normalized
else:
self.item["salary"] = 'Thỏa thuận'
pass
#Kinh nghiem
experience = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Kinh nghiệm")]/span/text()'
).extract()
if len(experience) > 0:
self.item["experience"] = experience[0]
else:
self.item["experience"] = 'Không yêu cầu'
# Bang cap
diploma = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Yêu cầu bằng cấp")]/span/text()'
).extract()
if len(diploma) > 0:
self.item["diploma"] = diploma[0]
else:
self.item['diploma'] = 'Không yêu cầu'
#So luong
amount = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Số lượng cần tuyển")]/span/text()'
).extract()
if len(amount) > 0:
self.item["amount"] = amount[0]
else:
self.item['amount'] = 'Không yêu cầu'
#Nganh nghe
career = response.xpath(
'//div[@class="line-icon mb_12"]//h2[contains(text(),"Ngành nghề")]//a/text()'
).extract()
career_need_nomarlize = Normalize_careers()
career_normalized = career_need_nomarlize.normalize_careers(career)
self.item["career"] = career_normalized
# Chuc vu
position = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Chức vụ")]/span/text()'
).extract()
if len(position) > 0:
self.item["position"] = position[0]
else:
self.item['position'] = 'Không yêu cầu'
#Hinh thuc lam viec fulltime/parttime
category = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Hình thức làm việc")]/span/text()'
).extract()
if len(category) > 0:
self.item["category"] = category[0]
else:
self.item['category'] = 'Không yêu cầu'
#Thoi gian thu viec
trial_time = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Thời gian thử việc")]/span/text()'
).extract()
if len(trial_time) > 0:
self.item["trial_time"] = trial_time[0]
else:
self.item['trial_time'] = 'Không yêu cầu'
#Yeu cau gioi tinh
sex = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Yêu cầu giới tính")]/span/text()'
).extract()
if len(sex) > 0:
self.item["sex"] = sex[0]
else:
self.item['sex'] = 'Không yêu cầu'
#Yeu cau tuoi
age = response.xpath(
'//div[@class="col-xs-6"]//p//span[contains(text(),"Yêu cầu độ tuổi")]/span/text()'
).extract()
if len(age) > 0:
self.item["age"] = age[0]
else:
self.item['age'] = 'Không yêu cầu'
#Mo ta
description = response.xpath(
'(//div[@id="ttd_detail"]//div[@class="item row"])[1]//p[@class="col-md-9 pr_0 mb_0 word_break"]/text()'
).extract()
self.item["description"] = " ".join(
[des.strip() for des in description])
#Quyen loi duoc huong
benefits = response.xpath(
'(//div[@id="ttd_detail"]//div[@class="item row"])[2]//p[@class="col-md-9 pr_0 mb_0 word_break"]/text()'
).extract()
self.item["benefits"] = " ".join(
[benefit.strip() for benefit in benefits])
#Yeu cau khac
require_skills = response.xpath(
'(//div[@id="ttd_detail"]//div[@class="item row"])[3]//p[@class="col-md-9 pr_0 mb_0 word_break"]/text()'
).extract()
self.item["require_skill"] = " ".join(
[require_skill.strip() for require_skill in require_skills])
#Thong tin lien he
per_contact = response.xpath(
'(//div[@class="job_description bg_white pl_24 pr_24 mt_16 pb_18 box_shadow"]//div[@class="item row pt_14 pb_14"])[1]//p[@class="col-md-9 pr_0 mb_0"]/text()'
).extract()
add_contact = response.xpath(
'(//div[@class="job_description bg_white pl_24 pr_24 mt_16 pb_18 box_shadow"]//div[@class="item row pt_14 pb_14"])[2]//p[@class="col-md-9 pr_0 mb_0"]/text()'
).extract()
contact = u"Người liên hệ: " + per_contact[0].strip(
) + u" Địa chỉ liên hệ: " + add_contact[0].strip()
self.item["contact"] = contact
#Han nop ho so
expired = response.xpath(
'(//span[@class="text_pink"])[1]/text()').extract() #Het han
if len(expired) > 0:
self.item["expired"] = expired[0]
pass
#Ngay tao hoso
created = response.xpath(
'(//p[@class="text_grey2 font12 mt8 mb12"]//span)[3]/text()'
).extract() #Ngay tao
if len(created) > 0:
created_at = created[0][14:]
self.item["created"] = created_at
if self.item["title"] != "":
yield self.item
class CareerbuilderSpider(scrapy.Spider):
name = 'careerbuilder'
start_urls = ['https://careerbuilder.vn/viec-lam/tat-ca-viec-lam-vi.html/']
ur = pybloom_live.BloomFilter(capacity=2097152, error_rate=0.005)
collection_name = 'News'
# Y_in_db = []
client = pymongo.MongoClient(settings.MONGO_URI)
db = client[settings.MONGO_DATABASE]
collection = db[collection_name]
Y_in_db = list(collection.find({}, {"title":1,"company":1, "address":1, "_id":0}))
no_duplicate_items = 0
def parse(self, response):
# client = pymongo.MongoClient(self.settings.get("MONGO_URI"))
# db = client[self.settings.get("MONGO_DATABASE")]
# collection = db[self.collection_name]
# self.Y_in_db = list(collection.find({}, {"title":1,"company":1, "address":1, "_id":0}))
# Time start
for tn in response.xpath('//h3[@class="job"]'):
src = tn.xpath('a/@href').extract_first()
src = response.urljoin(src)
add_url = self.ur.add(src)
if add_url is False:
yield scrapy.Request(src, callback=self.parse_src)
pass
next_pages = response.xpath('//a[@class="right"]/@href').extract()
next_page = next_pages[len(next_pages) - 1]
if next_page is not None:
next_page = response.urljoin(next_page)
yield scrapy.Request(next_page, callback=self.parse)
def parse_src(self, response):
self.item = JobItem()
self.item["url"] = response.request.url
#Nganh nghe
career= response.xpath("//div[@id='showScroll']/ul/li/p/span[contains(text(), 'Ngành nghề')]/following-sibling::b/a/text()").extract()
print(career)
career_need_nomarlize = Normalize_careers()
career_normalized = career_need_nomarlize.normalize_careers(career)
self.item["career"] = career_normalized
#Tieu de
title = response.xpath('//div[@class="top-job-info"]/h1/text()').extract()
x_title = title[0]
self.item["title"] = x_title
company = response.xpath('//div[@class="tit_company"]/text()').extract()
x_company = company[0]
self.item["company"] = x_company
#Noi lam viec
address = response.xpath("//div[@id='showScroll']/ul/li/p/span[contains(text(), 'Nơi làm việc')]/following-sibling::b/a/text()").extract()
add = ", ".join(address)
self.item["address"] = ", ".join(address)
data_need_check = DataReduction(3, [[job['title'], job['company'], job['address']] for job in self.Y_in_db])
#Check duplicate
if data_need_check.is_match([x_title, x_company, add]):
self.no_duplicate_items += 1
print(self.no_duplicate_items)
return
# Luong
salary = response.xpath("//div[@id='showScroll']/ul/li/p/span[contains(text(), 'Lương')]/following-sibling::label/text()").extract()
if len(salary) > 0:
salary_str = " ".join(salary)
salary_need_normalize = Normalize_salary()
salary_normalized = salary_need_normalize.normalize_salary(salary_str)
self.item["salary"] = salary_normalized
else:
self.item["salary"] = 'Thỏa thuận'
pass
#Kinh nghiem
experience = response.xpath("//div[@id='showScroll']/ul/li/p/span[contains(text(), 'Kinh nghiệm')]/../text()").extract()
if len(experience) > 0:
self.item["experience"] = experience[0].strip()
elif not experience:
self.item["experience"] = "Không yêu cầu"
descriptions = response.xpath("//div[@class='MarBot20']/h4[contains(text(),'Mô tả Công việc')]/following-sibling::div[@class='content_fck']//text()").extract()
self.item["description"] = ' '.join([description.replace('-', '').strip() for description in descriptions])
#Thong tin khac bao gom bang cap, do tuoi, va hinh thuc lam viec
info_others = response.xpath("//div[@class='MarBot20']/h4[contains(text(),'Thông tin khác')]/following-sibling::div[@class='content_fck']/ul/li//text()").extract()
# Bang cap
diploma = [info_other.strip() for info_other in info_others if
"bằng cấp" in info_other.lower() or "tốt nghiệp" in info_other.lower()] + [
description.strip() for description in descriptions if "bằng cấp" in description.lower() or "tốt nghiệp" in description.lower()]
if len(diploma) > 0:
self.item['diploma'] = diploma[0].split(':')[-1].strip()
else:
self.item['diploma'] = 'Không yêu cầu'
#So 0
amount = ""
self.item["amount"] = amount
# Chuc vu
position = response.xpath('(//p[@class="fl_right"])[1]//label/text()').extract()
if len(position) > 0:
self.item["position"] = position[0]
pass
#Hinh thuc lam viec fulltime/parttime
category = [info_other.strip() for info_other in info_others if "hình thức" in info_other.lower()] + [description.strip() for description in descriptions if "hình thức" in description.lower()]
if len(category) > 0:
self.item['category'] = category[0].split(':')[-1].strip()
else:
self.item['category'] = 'Không yêu cầu'
#Thoi gian thu viec
trial_time = [info_other.strip() for info_other in info_others if "thời gian thử việc" in info_other.lower()]
if len(trial_time) > 0:
self.item["trial_time"] = trial_time
else:
self.item['trial_time'] = 'Không yêu cầu'
#Yeu cau gioi tinh
sex_male = "Nam" if len([info_other for info_other in info_others if "nam" in info_other.lower()]) > 0 else ""
sex_female = "Nữ" if len([info_other for info_other in info_others if "nữ" in info_other.lower()]) > 0 else ""
sex = sex_male + ("", "/")[sex_male != "" and sex_female != ""] + sex_female
if sex == "":
self.item['sex'] = "Không yêu cầu"
else:
self.item['sex'] = sex.strip()
ages = [other.strip() for other in info_others if "tuổi" in other] + [description.strip() for description in
descriptions if "tuổi" in description]
if len(ages) > 0:
self.item["age"] = ages[0].split(":")[-1].strip()
else:
self.item["age"] = 'Không yêu cầu'
#Mo ta
#Quyen loi duoc huong
benefits = response.xpath("//div[@class='MarBot20 benefits-template']/h4[contains(text(),'Phúc lợi')]/following-sibling::ul/li/text()").extract()
self.item["benefits"] = ', '.join(benefits).strip()
#Yeu cau khac
require_skills = response.xpath("//div[@class='MarBot20']/h4[contains(text(),'Yêu Cầu Công Việc')]/following-sibling::div[@class='content_fck']//text()").extract()
self.item["require_skill"] = " ".join([skill.replace('-', '').strip() for skill in require_skills])
#Thong tin lien he
per_contact = response.xpath('(//p[@class="TitleDetailNew"]//label)[3]//strong/text()').extract()
add_contact = response.xpath('(//p[@class="TitleDetailNew"]//label)[2]/text()').extract()
if len(per_contact) > 0 :
pers_contact = re.sub(r'<.*?>', ' ', per_contact[0])
pers_contact = re.sub(r'\n', ' ', pers_contact)
pers_contact = re.sub(r'\r', ' ', pers_contact)
pass
else:
pers_contact = ""
if len(add_contact) > 0:
addr_contact = re.sub(r'<.*?>', ' ', add_contact[0])
addr_contact = re.sub(r'\n', ' ', addr_contact)
addr_contact = re.sub(r'\r', ' ', addr_contact)
pass
else:
addr_contact = ""
# contact = pers_contact + "\n" +addr_contact
contact = u"Người liên hệ: " + pers_contact.strip() + u" Địa chỉ liên hệ: " + addr_contact.strip()
self.item["contact"] = contact
#Han nop ho so
expired = response.xpath('(//p[@class="fl_right"])[3]/text()').extract() #Het han
if len(expired) > 0:
self.item["expired"] = expired[0]
pass
#Ngay tao hoso
created = response.xpath('//div[@class="datepost"]//span/text()').extract() #Ngay tao
if len(created) > 0:
self.item["created"] = created[0]
if self.item["title"] != None:
yield self.item
import pybloom_live
import requests
if __name__ == '__main__':
texts = requests.get('http://www.gutenberg.org/files/2852/2852-0.txt')
bloom = pybloom_live.BloomFilter(capacity=10000, error_rate=0.1)
for each in texts.text.split():
bloom.add(each)
print(len(bloom))
print('the' in bloom)
print('luo_wei_' not in bloom)
'connection': "keep-alive",
'cache-control': "no-cache",
'upgrade-insecure-requests': "1",
'user-agent':
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.95 Safari/537.36",
'accept':
"text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
'accept-language': "zh-CN,en-US;q=0.8,en;q=0.6"
}
city_home_pages = []
city_ids = []
dirname = 'mafengwo_notes/'
# 创建 Bloom Filter
download_bf = pybloom_live.BloomFilter(1024 * 1024 * 16, 0.01)
def download_city_notes(id):
for i in range(1, 999): #遍历每个城市的每一页游记，再下载每一页的游记
url = 'http://www.mafengwo.cn/yj/%s/1-0-%d.html' % (id, i)
if url in download_bf:
continue
print("open url %s" % url)
download_bf.add(url)
req = urllib.request.Request(url, headers=request_headers)
response = urllib.request.urlopen(req)
htmlcontent = response.read().decode("utf-8")
city_notes = re.findall(r'href="/i/\d{7}.html', htmlcontent)
# 如果导航页错误，该页的游记数为0，则意味着 1-0-xxx.html 已经遍历完，结束这个城市
