class MySpider(object):
def __init__(self, root_url):
self.parser = HtmlParser()
self.storage = DataStore()
self._get_root_urls(root_url)
def _get_root_urls(self, root_url):
if os.path.exists('job_class.json'):
pass
else:
new_urls = self.parser.get_url(root_url)
self.storage.local_store(new_urls,
'job_class.json') #存储要爬取的行业类别url
def joburl_init(self, pagenum, path='job_class.json'):
root_urls = self.storage.load_data(path)
jobs_dict = {}
for i in pagenum:
for list in root_urls:
jobs_dict[list +
str(i)] = root_urls[list] + str(i) #构造要爬取的网址链接
self.storage.local_store(jobs_dict, 'job_page_url.json') #存储构造好的网址链接
def company_url(self, path='job_page_url.json'):
company_urls = self.storage.load_data(path)
company_dicts = {}
url_get = 0 #已获取的网址总数
for company_info_url in company_urls:
print("待爬取的行业网址总数:", len(company_urls) - url_get)
url_get += 1
url = company_urls[company_info_url]
company_dicts.update(self.parser.getcompany_url(url))
self.storage.local_store(url, 'job_page_url_old.json') #存储已爬取的网址
self.storage.local_store(company_dicts,
'company_info_url_new.json') #存储公司信息的URL
def company_info(self, path='company_info_url_new.json'):
company_info_urls = self.storage.load_data(path)
url_get = 0 #以获取的公司信息网址总数
for company_name in company_info_urls:
print("待爬取的公司信息网址总数:", len(company_info_urls) - url_get)
url_get += 1
url = company_info_urls[company_name]
self.parser.getcompany_info(company_name, url)
self.storage.local_store(
url, 'compang_info_url_old.json') #存储以爬取的存储公司信息URL
#从上次断点出重新开始获取公司信息
def grab_increment(self):
new_urls = self.storage.load_data('company_info_url_new.json')
old_urls = self.storage.load_data('compang_info_url_old.json')
for company_name in new_urls:
new_url = new_urls[company_name]
if new_url not in old_urls:
self.parser.getcompany_info(company_name, url)
self.storage.local_store(
url, 'compang_info_url_old.json') # 存储以爬取的存储公司信息URL
import pickle
import tensorflow as tf
import numpy as np
from datastore import DataStore
from model_def import CombinedModel
from tensorflow.keras.optimizers import Adam
# pass the mode OLD to read the picle datastore object
MODE = "OLD"
if MODE == "NEW":
datastore = DataStore()
# interactions file
datastore.load_data("agent_interactions.jsl")
datastore.build()
with open('dataset_pick.pkl', 'wb') as o:
pickle.dump(datastore, o, pickle.HIGHEST_PROTOCOL)
else:
with open("dataset_pick.pkl", "rb") as p:
datastore = pickle.load(p)
print(datastore.actions.shape)
combined_model = CombinedModel(16, 8, 4)
optimizer = Adam(learning_rate=0.00001)
@tf.function
def train_step(train_states, prev_states, current_states, actions):
with tf.GradientTape() as tape:
