super/crawler-plugin
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

同步 Gitee 应用代码更新

Browse Source
This commit is contained in:
super
2026-05-29 19:47:23 +08:00
parent 266c0f17c1
commit 080625567b
7 changed files with 103 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
app/amazon/similar_asin.py
View File

@@ -257,22 +257,35 @@ class ChromeAmzone(ChromeAmzoneBase):
imge_ele = self.tab.ele('xpath://div[@id="imgTagWrapperId"]//img',timeout=20)
image_url = ""
min_image_url = ""
data_a_dynamic_image = imge_ele.attr("data-a-dynamic-image")
if data_a_dynamic_image:
dynamic_image_json = json.loads(data_a_dynamic_image)
self.log(f"图片信息:{dynamic_image_json}")
max_area = 0
min_area = 0
for url, (width, height) in dynamic_image_json.items():
area = width * height
if area > max_area:
max_area = area
image_url = url
if min_area == 0:
min_area = area
min_image_url = url
if area < min_area:
min_area = area
min_image_url = url
if not image_url:
image_url = imge_ele.attr("src")
if not min_image_url:
min_image_url = imge_ele.attr("src")
data["image_url"] = image_url
data["min_image_url"] = min_image_url
# sku
sku_ele_ls = self.tab.eles('xpath://ul[@class="a-unordered-list a-vertical a-spacing-mini"]',timeout=20)
@@ -340,9 +353,10 @@ class ChromeAmzone(ChromeAmzoneBase):
# 判断是否采集标题出错
new_size = "220,220"
pattern = r"\._(?:[A-Z]+)?(\d+)_\."
replacement = f"._{new_size}_.jpg"
# replacement = f"._{new_size}_.jpg"
# 使用正则替换
image_new_url = re.sub(pattern, lambda m: replacement, data["image_url"])
# image_new_url = re.sub(pattern, lambda m: replacement, data["image_url"])
image_new_url = data["min_image_url"]
print(image_new_url)