class CategoryCrawl(object):
CATEGORY_ID = 50000000 # Default value
COLLECTION = 'category'
_DELIMITER = 'category?catId='
_PATH_TOKEN = '#'
def __init__(self):
# 크롤링 설정 정보 관리 - singleton
self.crawl_config: CrawlConfiguration = ConfigManager().crawl_config
# Database manager - 데이터 조회 및 저장을 여기서 합니다. - singleton
self.database_manager = DatabaseManager()
# 중복 데이터 확인을 위해 미리 저장된 결과 list를 조회한다.
self._category_list: list = list(
self.database_manager.find_all_mongo(self.COLLECTION))
self.CATEGORY_ID = self.crawl_config.category_id
def _update(self, cid, name, paths: str):
# Database manager - 데이터 조회 및 저장을 여기서 합니다. - singleton
self.database_manager = DatabaseManager()
_query = self.database_manager.find_query('cid', cid)
_update_data = dict()
_update_data['name'] = name
_update_data['paths'] = paths
_update_data['update_time'] = datetime.now()
return self.database_manager.update(self.COLLECTION, _query,
{"&set": _update_data})
def _insert(self, cid, name, paths: str, is_root: bool = False):
""" Mongo Database Insert """
_is_exists: bool = False
for item in self._category_list:
_name = item['name']
_cid = item['cid']
_paths = item['paths']
if is_root:
if eq(_name, name):
self._category_list.remove(item)
return
else:
if eq(_cid, cid):
if eq(_name, name) and eq(_paths, paths):
self._category_list.remove(item)
return
else:
self._update()
self._category_list.remove(item)
return
_category_document = dict()
_category_document['cid'] = cid
_category_document['name'] = name
_category_document['paths'] = paths
_category_document['insert_time'] = datetime.now()
return self.database_manager.insert_one_mongo(self.COLLECTION,
_category_document)
def _is_exists(self, field, value: str):
"""MongoDB에 cid 값을 조회하여 조건에 맞는 document가 있는지 확인"""
_query = self.database_manager.find_query(field, value)
return self.database_manager.count_document('category', _query) > 0
def _parse_category(self, element: HtmlElement, root_paths: str):
ul_tag: HtmlElement = element.find('ul')
if ul_tag is not None:
li_tags = ul_tag.findall('li')
li: HtmlElement
for li in li_tags:
li_a_tag = li.find('a')
if li_a_tag is not None:
_name = li_a_tag.text
_href = li_a_tag.get('href')
_cid = Utils.separate_right(_href, self._DELIMITER)
_paths = Utils.join_path(self._PATH_TOKEN, root_paths,
_name)
self._insert(_cid, _name, _paths)
div_tag = li.find('div')
if div_tag is not None:
self._parse_category(div_tag, _paths)
if li.find('ul') is not None:
self._parse_category(li, _paths)
def parse(self):
for category_id in range(self.CATEGORY_ID, self.CATEGORY_ID + 11):
_url = 'https://search.shopping.naver.com/category/category/{0}'
logging.info("PID >> %s | CategoryID >> %d " %
(os.getpid(), category_id))
request = requests.get(_url.format(category_id))
Utils.take_a_sleep(0, 1)
# 상태 체크
if request.status_code != 200:
return
try:
_content = request.content
tree: HtmlElement = html.fromstring(_content)
header_xpath = '//*[@id="__next"]/div/div[2]/h2'
_root_name = tree.xpath(header_xpath)[0].text
self._insert(str(category_id), _root_name, None, True)
xpath = '//*[@id="__next"]/div/div[2]/div/div'
elements: [HtmlElement] = tree.xpath(xpath)
element: HtmlElement
for element in elements:
if element.find('div') is not None:
a_tag: HtmlElement = element[0].find('h3').find('a')
_name = a_tag.find('strong').text
_href = a_tag.get('href')
_cid = Utils.separate_right(_href, self._DELIMITER)
_paths = Utils.join_path(self._PATH_TOKEN, _root_name,
_name)
self._insert(_cid, _name, _paths)
self._parse_category(element[0], _paths)
else:
logging.info('Element is not Exists')
except Exception as e:
logging.error(str(e))
# 더이상 필요없는 카테고리 아이템들 제거
for item in self._category_list:
_query = self.database_manager.find_query('_id', item['_id'])
self.database_manager.delete_one(self.COLLECTION, _query)
def run(self):
pass
class ProductCrawl:
'''
상품목록에서 상품 정보를 수집하는 crawler 입니다.
Attributes:
driver
'''
PRODUCT_COLLECTION = "product"
CRAWL_CONFIG_COLLECTION = "crawl_config"
_excepted_data_count = 0
def __init__(self):
logging.info('start product crawl')
# Database manager - 데이터 조회 및 저장을 여기서 합니다. - singleton
self.database_manager = DatabaseManager()
self.crawl_config: CrawlConfiguration = ConfigManager().crawl_config
# start Page Default
self._paging_start: int = 1
self._view_size: int = 80 # self.crawl_config.crawl_count
# 먼저 확인해야함. 다시 수집시 DB->Config 정보 셋
# TODO: 나중에 처리하도록 수정
# self._check_crawl_configuration()
self._result: list = []
self._category: dict = None
self.productInfo_arr = []
self._current_page: int = 0
self.last_crawled_date_time = datetime.datetime.now()
def _upsert_crawl_configuration(self, start_page):
"""모든 분석이 끝나고 Config 정보 update"""
# 조건
_filter = {}
# 변경 데이터
_config = dict()
_config['start_page'] = start_page
self.database_manager.update(self.CRAWL_CONFIG_COLLECTION, _filter, _config)
def _check_crawl_configuration(self):
"""Config 정보 set"""
_config: dict = self.database_manager.find_one(self.CRAWL_CONFIG_COLLECTION)
if _config.get('start_page') is not None:
self._paging_start = _config['start_page']
if _config.get('crawl_category_list') is not None:
self.crawl_config.crawl_category = _config['crawl_category_list']
def _category_getter(self, crawl_category: list) -> list:
""" 카테고리 목록 조회해서 분석
:return category 목록들"""
_categories: list = []
if crawl_category is None:
crawl_category = self.crawl_config.crawl_category
for item in crawl_category:
query = self.database_manager.keyword_query('paths', item)
_categories.extend(list(self.database_manager.find('category', query=query)))
return _categories
def make_url(self, paging_index: int, frm: str = "NVSHMDL", _filter: str = "") -> str:
"""category id, 페이지 사이즈, 페이지 넘버를 조합하여 url 생성"""
_url = ("https://search.shopping.naver.com/search/category?catId={0}&frm={1}{2}&origQuery&pagingIndex={3}&pagingSize={4}&productSet=model&query&sort=rel×tamp=&viewType=list")
_cid = self._category['cid']
return _url.format(_cid, frm, _filter, paging_index, self._view_size)
async def parse(self, identifier: str, context: dict, crawl_category: list = None):
""" 외부에서 파싱을 하기 위해 호출하는 함수 """
_categories: list = self._category_getter(crawl_category)
for category in _categories:
await asyncio.sleep(1)
jobs = context['jobs']
job_info = jobs[identifier]
job_info['status'] = 'in Progress'
job_info['category'] = category.get('name')
self._category = category
"""파싱 프로세스 시작"""
self._current_page = 0
# Default = 1
_url = self.make_url(paging_index=1)
_total_count, _filter = self._get_base_data(_url)
# Page 조건 변경 필요
_is_oversize = _total_count > 8000
# Page 계산
_page_size = Utils.calc_page(_total_count, self._view_size)
if _is_oversize:
self._filter_parse(_filter)
else:
await self._execute_parse(_page_size)
logging.info('>>> end childCategory: ' + self._category.get('name') + ' Pg.' + str(self._current_page))
job_info['status'] = 'done'
def _make_list(self, _min, _max, _half):
result = []
a = [_min, _half]
b = [_half, _max]
result.append(a)
result.append(b)
return result
def _filter_parse_recursive(self, min_value, max_value):
_param = ("&maxPrice={0}&minPrice={1}".format(str(max_value), str(min_value)))
_url = self.make_url(1, "NVSHPRC", _param)
_total_count, _filter = self._get_base_data(_url)
_is_oversize = _total_count > 8000
_page_size = Utils.calc_page(_total_count, self._view_size)
if _is_oversize:
half_price = math.ceil((min_value + max_value) / 2)
_range = self._make_list(min_value, max_value, half_price)
for value in _range:
self._filter_parse_recursive(value[0], value[1])
else:
await self._execute_parse(_page_size, _param)
pass
def _filter_parse(self, filters: list):
# 한번만 호출된다.
for _filter in filters:
_filterAction = _filter.get('filterAction')
_separator = "-" # default = -
_paramName = None
if _filterAction is not None:
_separator = _filterAction.get('separator')
# price split
_value: str = _filter.get('value')
_param = ""
_min = 0
_max = 0
if _value is not None:
_min, _max = (int(_price) for _price in _value.split(_separator))
logging.info("Filter Parse >> min{0} / max{1}".format(_min, _max))
self._filter_parse_recursive(_min, _max)
async def _execute_parse(self, page_number, filter_param: str = ""):
for page_number in range(1, page_number):
try:
_url = self.make_url(page_number, _filter=filter_param)
self.parse_data(self._get_product_json(_url))
logging.info(">>> URL : " + _url)
logging.info('>>> start parsing: ' + self._category.get('name') + ' Pg.' + str(page_number))
self._current_page = page_number
except Exception as e:
logging.debug(">>> Category Collect Err " + str(self._current_page)
+ " name: " + self._category.get('name') + " Err :" + str(e))
def _get_product_json(self, url) -> dict:
"""
상품 정보 가져오기
:arg
:param url: request URL
:return: data_dict 상품 정보
"""
# header 추가 필요.
try:
_headers = {'Content-Type': 'application/json;'}
req = requests.get(url, _headers)
html = req.text
soup = BeautifulSoup(html, 'html.parser') # html.parser를 사용해서 soup에 넣겠다
json_data = soup.find('script', text=re.compile('application/json'))
data_dict = json.loads(str(json_data.contents[0]))
except Exception as e:
data_dict = None
# 슬립 시간 조정 필요 - 8초가 부족할 수 있음.
time.sleep(8)
# 비정상적인 요청이 감지됨 - 다시 URL을 요청한다.
logging.error("no find Data request Error >> {0} | URL >> {1}".format(e, url))
self._get_product_json(url)
return data_dict
def parse_data(self, data_dict):
""" 데이터 파싱 """
product_info: dict = self._get_data(data_dict, 'products')
if product_info is not None:
'''수집된 데이터가 있는 경우'''
product_list: list = product_info.get('list')
self._excepted_data_count = 0
logging.info("수집 시작 - 상품 데이터 수: " + str(len(product_list)))
if len(product_list) > 0:
for product in product_list:
product_data = dict()
product_item = product.get('item')
if product_item.get('adId') is None:
'''광고 데이터가 아닌 경우에만 수집'''
# 카테고리 정보 Setting
self._set_category_info(product_data)
# 상품 정보 Setting
self._set_product_info(product_data, product_item)
self._insert_product_info(product_data)
else:
self._excepted_data_count += 1
else:
logging.error('!!! Exception: 상품 정보가 없습니다.')
# if len(product_list) != len(products_data) + self._excepted_data_count:
# logging.error("!!! Exception: 데이터 수 확인이 필요 합니다.")
# logging.info("수집된 데이터 수: " + str(len(products_data)))
# logging.info("수집 제외된 데이터 수: " + str(self._excepted_data_count))
else:
logging.error('!!! Exception: 데이터가 수집되지 않았습니다.')
def _set_category_info(self, product_data: dict):
'''상품정보에 카테고리 정보 셋팅
arg:
products_data: 상품 정보 객체
'''
product_data['n_cid'] = self._category.get('cid')
product_data['cid'] = self._category.get('_id')
product_data['paths'] = self._category.get('paths')
product_data['cname'] = self._category.get('name')
def _set_product_info(self, product_data: dict, product_item):
product_data['n_id'] = product_item.get('id')
product_data['imageUrl'] = product_item.get('imageUrl')
product_data['title'] = product_item.get('productTitle')
product_data['price'] = product_item.get('price')
# product_data['option'] = {}
_attribute: str = product_item.get('attributeValue', "")
_attribute_value: str = product_item.get('characterValue', "")
if (_attribute != "") and (_attribute_value != ""):
# 옵션 정보가 있는 경우
product_option_key: list = product_item.get('attributeValue').split('|') # 옵션 키값
product_option_value: list = product_item.get('characterValue').split('|') # 옵션 벨류값
product_data['option'] = dict(zip(product_option_key, product_option_value))
def _insert_product_info(self, value: dict):
"""db data insert"""
try:
# TODO: 값 비교는 어디서 하지?
_selection = self.database_manager.find_query("n_id", value.get("n_id"))
self.database_manager.update(self.PRODUCT_COLLECTION, _selection, value)
pass
except Exception as e:
logging.error('!!! Fail: Insert data to DB: ', e)
def _get_base_data(self, url):
_data = self._get_product_json(url)
_total_count = 0
value_filters: Optional[dict] = None
if _data is not None:
products = self._get_data(_data, 'products')
if products is not None:
_total_count = products.get('total')
if _total_count is not None:
_total_count = int(_total_count)
else:
_total_count = 0
filters = self._get_data(_data, 'mainFilters')
if filters is not None:
value_filters = self._get_filter(filters)
return _total_count, value_filters
def _get_data(self, data: dict, _type: str):
return data.get('props', {}).get('pageProps', {}).get('initialState', {}).get(_type)
def _get_filter(self, main_filters: dict) -> dict:
value_filters = None
for _filter in main_filters:
_filterType: str = _filter.get('filterType')
if (_filterType is not None) and (eq(_filterType, 'price')):
value_filters = _filter.get('filterValues')
if value_filters is not None:
break
return value_filters
class CategoryCrawl(object):
URL = 'https://search.shopping.naver.com/category/category/{0}'
CATEGORY = 50000000
DELIMITER = 'cat_id='
COLLECTION = 'category'
def __init__(self):
# 크롬 selenium Driver - singleton
self.driver = Selenium().driver
# 크롤링 설정 정보 관리 - singleton
self.crawl_config: CrawlConfiguration = ConfigManager().crawl_config
# Database manager - 데이터 조회 및 저장을 여기서 합니다. - singleton
self.database_manager = DatabaseManager()
# 중복 데이터 확인을 위해 미리 저장된 결과 list를 조회한다.
self._category_list: list = list(
self.database_manager.find_all_mongo(self.COLLECTION))
def _update(self, cid, name, paths: str):
_query = self.database_manager.find_query('cid', cid)
_update_data = dict()
_update_data['name'] = name
_update_data['paths'] = paths
_update_data['update_time'] = datetime.now()
return self.database_manager.update(self.COLLECTION, _query,
{"&set": _update_data})
def _insert(self, cid, name, paths: str, is_root: bool = False):
""" Mongo Database Insert """
_is_exists: bool = False
for item in self._category_list:
_name = item['name']
_cid = item['cid']
_paths = item['paths']
if is_root:
if eq(_name, name):
self._category_list.remove(item)
return
else:
if eq(_cid, cid):
if eq(_name, name) and eq(_paths, paths):
self._category_list.remove(item)
return
else:
self._update()
self._category_list.remove(item)
return
_category_document = dict()
_category_document['cid'] = cid
_category_document['name'] = name
_category_document['paths'] = paths
_category_document['insert_time'] = datetime.now()
return self.database_manager.insert_one_mongo(self.COLLECTION,
_category_document)
def _is_exists(self, field, value: str):
"""MongoDB에 cid 값을 조회하여 조건에 맞는 document가 있는지 확인"""
_query = self.database_manager.find_query(field, value)
return self.database_manager.count_document('category', _query) > 0
def parse(self):
self.driver.get(self.URL)
try:
for category in self.driver.find_elements_by_xpath(
'//*[@id="home_category_area"]/div[1]/ul/li'):
time.sleep(1)
self._parse_root(category)
# 더이상 필요없는 카테고리 아이템들 제거
for item in self._category_list:
_query = self.database_manager.find_query('_id', item['_id'])
self.database_manager.delete_one(self.COLLECTION, _query)
except Exception as e:
logging.error(str(e))
def _parse_root(self, category: WebElement):
# Root 이름
root_name: str = category.text
# root_name = text.replace('/', '-')
logging.info('rootName : ' + root_name)
for exclude_category in self.crawl_config.exclude_category:
if eq(root_name, exclude_category):
return None
class_att = category.get_attribute('class')
click_xpath = '//*[@id="home_{0}"]'.format(class_att)
self.driver.implicitly_wait(5)
# 먼저 클릭해봄.
self.driver.find_element_by_xpath(click_xpath).send_keys(Keys.ENTER)
# class_att 맞춰 내부 xPath 설정
time.sleep(1)
xpath_cate = '//*[@id="home_{0}_inner"]/div[1]'.format(class_att)
# Root Category
element: WebElement = None
while 1:
if element is not None:
break
else:
# 클릭 이벤트가 정상적으로 안들어오면 계속 클릭하자..
self.driver.find_element_by_xpath(click_xpath).send_keys(
Keys.ENTER)
self.driver.implicitly_wait(4)
time.sleep(1)
element = self.driver.find_element_by_xpath(xpath_cate)
self._insert(None, root_name, None, True)
# Root -> sub
co_col_elements = element.find_elements(By.CLASS_NAME, 'co_col')
self._parse_co_col(co_col_elements, root_name)
def _parse_co_cel(self, co_cel_elements, root_name):
co_cel: WebElement
for co_cel in co_cel_elements:
# href
sub_href = co_cel.find_element_by_tag_name('a').get_attribute(
'href')
# cid
_cid = Utils.separate_right(sub_href, self.DELIMITER)
sub_element: WebElement = co_cel.find_element_by_tag_name('strong')
# name
_name = sub_element.find_element_by_tag_name('a').text
_name = re.sub("전체보기", "", _name)
# paths
_paths = Utils.join_path(token='#', source=root_name, value=_name)
# cid, name, paths
self._insert(_cid, _name, _paths)
# 하위 카테고리 리스트
child_items: [WebElement] = co_cel.find_elements(By.TAG_NAME, 'li')
self._parse_child(child_items, _paths)
pass
def _parse_co_col(self, sub_category, root_name):
co_col: WebElement
for co_col in sub_category:
time.sleep(1)
# 중간 카테고리
co_cel_elements = co_col.find_elements_by_class_name('co_cel')
self._parse_co_cel(co_cel_elements, root_name)
def _parse_child(self, child_items, sub_paths):
child_item: WebElement
for child_item in child_items:
time.sleep(1)
# href
_href = child_item.find_element_by_tag_name('a').get_attribute(
'href')
# cid
_cid = Utils.separate_right(_href, self.DELIMITER)
# name
_name = child_item.text # 이름
# paths
_paths = Utils.join_path(token='#', source=sub_paths, value=_name)
self._insert(_cid, _name, _paths)