dtomlinson/geography-anki
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

chore: fix anthem file

Browse Source
This commit is contained in:
Daniel Tomlinson
2022-06-24 00:51:15 +01:00
parent 2c3f6cb3d0
commit f0b675d4ef
1 changed files with 8 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
01_scrapy/wikipedia_country_scraper/wikipedia_country_scraper/spiders/countrydownloader.py
View File

@@ -100,7 +100,14 @@ class CountrydownloaderSpider(scrapy.Spider):
def extract_anthem_file(self, response: TextResponse, country_item: dict, urls: dict): def extract_anthem_file(self, response: TextResponse, country_item: dict, urls: dict):
anthem_text = response.xpath("//div[@id='mw-content-text']/div/p").get() anthem_text = response.xpath("//div[@id='mw-content-text']/div/p").get()
anthem_file_url = response.xpath("//tr[contains(@class, 'haudio')]//a/@href").get() _anthem_file_url = response.xpath("//tr[contains(@class, 'haudio')]//a/@href").getall()
for file in _anthem_file_url:
if file.endswith(".ogg") or file.endswith(".oga"):
anthem_file_url = file
break
else:
anthem_file_url = None
country_scrapy_item = WikipediaCountryScraperItem() country_scrapy_item = WikipediaCountryScraperItem()
country_scrapy_item["country_url"] = country_item["country_url"] country_scrapy_item["country_url"] = country_item["country_url"]