Paddy/Betterzon
1
0
Fork 0
mirror of https://github.com/Mueller-Patrick/Betterzon.git synced 2024-12-23 04:05:12 +00:00
Code Issues Projects Releases Wiki Activity

BETTERZON-58: Fixing API endpoint of the crawler

Browse Source 
- The list of products in the API request was treated like a string and henceforth, only the first product has been crawled
This commit is contained in:
Patrick Müller 2021-05-17 17:53:20 +02:00
parent 73effffc89
commit 63cbac5490
3 changed files with 9 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
Crawler/api.py
View File

@ -10,8 +10,8 @@ api = Api(app)
# To parse request data # To parse request data
parser = reqparse.RequestParser() parser = reqparse.RequestParser()
parser.add_argument('key') parser.add_argument('key', type=str)
parser.add_argument('products') parser.add_argument('products', type=int, action='append')
class CrawlerApi(Resource): class CrawlerApi(Resource):

9
Crawler/crawler.py
View File

@ -73,9 +73,14 @@ def __crawl_amazon__(product_info: dict) -> tuple:
try: try:
price = int(soup.find(id='priceblock_ourprice').get_text().replace(".", "").replace(",", "").replace("€", "").strip()) price = int(soup.find(id='priceblock_ourprice').get_text().replace(".", "").replace(",", "").replace("€", "").strip())
except RuntimeError: except RuntimeError:
price = '' price = -1
except AttributeError:
price = -1
return (product_info['product_id'], product_info['vendor_id'], price) if price != -1:
return (product_info['product_id'], product_info['vendor_id'], price)
else:
return None
def __crawl_apple__(product_info: dict) -> tuple: def __crawl_apple__(product_info: dict) -> tuple:

1
Crawler/sql.py
View File

@ -54,7 +54,6 @@ def getProductLinksForProduct(product_id: int) -> [dict]:
cur = conn.cursor() cur = conn.cursor()
query = 'SELECT vendor_id, url FROM product_links WHERE product_id = %s' query = 'SELECT vendor_id, url FROM product_links WHERE product_id = %s'
print(query)
cur.execute(query, (product_id,)) cur.execute(query, (product_id,))
products = list(map(lambda x: {'product_id': product_id, 'vendor_id': x[0], 'url': x[1]}, cur.fetchall())) products = list(map(lambda x: {'product_id': product_id, 'vendor_id': x[0], 'url': x[1]}, cur.fetchall()))