research/resources
Archived
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

modify scripts

Browse Source
This commit is contained in:
oscarz
2025-03-18 18:30:14 +08:00
parent 33e3155599
commit f4dacaa9ea
2 changed files with 4 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
aabook/src/fetch.py
View File

@ -25,7 +25,8 @@ def fetch_book_list():
# 获取书籍列表 # 获取书籍列表
list_data, next_url = scraper.parse_book_list(soup, url=url) list_data, next_url = scraper.parse_book_list(soup, url=url)
for item in list_data: for item in list_data:
row_id = db_tools.insert_or_update_common(item, db_tools.tbl_name_books) #row_id = db_tools.insert_or_update_common(item, db_tools.tbl_name_books)
row_id = db_tools.insert_books_index(item)
if row_id: if row_id:
logging.debug(f"insert one book. row_id: {row_id}, name: {item['name']}") logging.debug(f"insert one book. row_id: {row_id}, name: {item['name']}")
else: else:

3
aabook/src/sqlite_utils.py
View File

@ -130,10 +130,11 @@ def insert_or_update_common_lower(data, tbl_name, uniq_key='href'):
def insert_books_index(data): def insert_books_index(data):
try: try:
# 查询是否存在以及是否需要更新 # 查询是否存在以及是否需要更新
cursor.execute(f"SELECT id FROM books WHERE href = ? and update_time >= ?", (data['href'], data['update_time'], )) cursor.execute(f"SELECT id FROM {tbl_name_books} WHERE href = ? and update_time >= ?", (data['href'], data['update_time'], ))
existing_book = cursor.fetchone() existing_book = cursor.fetchone()
if existing_book: # **如果演员已存在** if existing_book: # **如果演员已存在**
logging.debug(f"book {data['href']} already exist. id: {existing_book[0]}")
return existing_book[0] return existing_book[0]
# 不存在，或者需要更新 # 不存在，或者需要更新