#%% 從IEEE與Google學術爬取論文引用
from selenium import webdriver
from urllib import parse
from time import sleep

class GetBibs():
  def __init__(self, driver_path, option_path, ie_search_url, gg_search_url) -> None:
    self.ie_search_url = ie_search_url
    self.gg_search_url = gg_search_url
    # 啟用帶插件的瀏覽器
    option = webdriver.ChromeOptions()
    option.add_argument("--user-data-dir="+option_path)
    self.browser = webdriver.Chrome(executable_path = driver_path, options = option)   # 打開chrome瀏覽器
    self.browser.set_window_size(800,800) #不要隨意修改，太窄會導致按鈕的隱藏，模擬點擊失效

  def get_bib_from_IEEE(self, paper_title):
    strto_pn=parse.quote(paper_title)
    url = self.ie_search_url + strto_pn
    self.browser.get(url)  
    compare_title = ''.join(list(filter(str.isalnum, paper_title))).lower()    
    #等待詞條載入
    for i in range(100):
      try:
        elements=self.browser.find_elements_by_css_selector("[class='List-results-items']")
        elements[0].get_attribute('id')
        break
      except:
        sleep(0.1)
    #掃描所有詞條，是否存在所需文獻
    paper_url = r'//ieeexplore.ieee.org/document/'
    for i in elements:
      s_title = i.text.split('\n')[0]
      s_title = ''.join(list(filter(str.isalnum, s_title))).lower()
      if s_title == compare_title:
        paper_url += i.get_attribute('id')
        break
    if paper_url == r'//ieeexplore.ieee.org/document/': #沒找到
      return ''
    # 進入文獻頁面
    self.browser.get(paper_url) 
    # 等待載入bib按鈕     
    for i in range(100):
      try:
        element=self.browser.find_element_by_css_selector("[class='layout-btn-white cite-this-btn']")
        element.click()
        break
      except:
        sleep(0.1)
    # 點擊bibtex
    for i in range(100):
      try:
        element=self.browser.find_element_by_css_selector("[class='modal-dialog']")
        element=element.find_elements_by_css_selector("[class='document-tab-link']")[1]
        element.click()
        break
      except:
        sleep(0.1) 
    for i in range(100):
      try:    
        self.browser.find_element_by_css_selector("[class='text ris-text']")
        break
      except:
        sleep(0.1) 
    sleep(2)
    bib = self.browser.find_element_by_css_selector("[class='text ris-text']").text
    return bib 

  def get_bib_from_google_scholar(self, paper_title):
    strto_pn=parse.quote(paper_title)
    url = self.gg_search_url + strto_pn
    self.browser.get(url)      
    #等待詞條載入
    for i in range(100):
      try:
        element=self.browser.find_element_by_css_selector("[class='gs_r gs_or gs_scl']")
        element=element.find_element_by_css_selector("[class='gs_fl']")
        element=element.find_element_by_css_selector("[class='gs_or_cit gs_nph']")
        element.click()
        break
      except:
        sleep(0.1)
    for i in range(100):
      try:
        element=self.browser.find_element_by_id("gs_citi")
        element=element.find_element_by_css_selector("[class='gs_citi']")
        element.click()
        break
      except:
        sleep(0.1)
    for i in range(100):
      try:
        bib = self.browser.find_element_by_tag_name('pre').text
        break
      except:
        sleep(0.1)
    return bib

  def get_bib(self, paper_title):
    bib = self.get_bib_from_IEEE(paper_title)
    if bib != '':
      return "IEEE", bib
    return "Google", self.get_bib_from_google_scholar(paper_title) 

driver_path = r'C:/chromedriver/chromedriver' # 瀏覽器驅動位置
option_path = r"C:/Users/Administrator/AppData/Local/Google/Chrome/User Data/" # 使瀏覽器能用你自定義的設置，否則Selenium創建的瀏覽器對象是默認設置，一些插件就不能用了
ie_search_url = r'//ieeexplore.ieee.org/search/searchresult.jsp?newsearch=true&queryText=' # 在執行程式碼之前，先打開IEEE官網的搜索頁面，把類似的網址複製到這裡，等號=後面就是一會兒要搜索的內容
gg_search_url = r'//scholar.google.com/scholar?hl=zh-CN&as_sdt=0%2C5&inst=1597255436240989024&q=' # Google學術也是一樣
get_bibs = GetBibs(driver_path, option_path, ie_search_url, gg_search_url)
#%% **********************以上定義爬蟲對象，以下開始爬取*******************************
paper_titles = { # 要爬取的論文，key用於標記，value是論文題目。下面是一些樣例
  "ESPCN":'Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network',
  "Sparse_Coding":'Image Super-Resolution Via Sparse Representation',
  "ESRGAN":'ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks',
  "EnhanceNet":'EnhanceNet: Single Image Super-Resolution Through Automated Texture Synthesis',
  'Meta-SR': 'Meta-SR: A Magnification-Arbitrary Network for Super-Resolution',
  'SAN': 'Second-Order Attention Network for Single Image Super-Resolution',
}

for k in paper_titles.keys():
  source, bib = get_bibs.get_bib(paper_titles[k]) 
  print(source+":",k)
  print(bib)
  print()