[BUG] build() sometimes returns a bunch of URLs, sometimes not

[MrsBookik]

Describe the bug

build() seems not appear to be reliable. Sometimes it does what you would expect, sometimes not. Sometimes it returns tons of articule URLs, and some seconds later not, providing the same input parameter. When I scrape cnn.com, or edition.cnn.com, which has been used in the official examples, it constatenly returns different result und subsequent calls. On the first attempt, it returns 100s of article URLs. On the second call only 5, on the third only 2 and from then on, zero results. Waiting a day, starting from scratch, repeats it in a similar way.

To Reproduce Steps to reproduce the behavior, please post any code you used and the website you tried to parse/process:

cnn_paper = newspaper.Source('https://cnn.com')

print(cnn_paper.size())  # no articles, we have not built the source

cnn_paper.build()
print(cnn_paper.article_urls())

print(cnn_paper.size())

Expected behavior

it would always return a bunch of listed URLs on the cnn.com site.

Screenshots

System information

• OS: [Windows / Linux / Macos]
• Python version [e.g. 3.6, 3.9]
• Library version [e.g. 0.9.0]

Additional context Add any other context about the problem here.

[AndyTheFactory]

Hi, can you try out with the new release (0.9.3)?

Also, make sure you test with memorize_articles=False?

https://newspaper4k.readthedocs.io/en/latest/user_guide/api_reference.html?highlight=memorize_#newspaper.configuration.Configuration.memorize_articles

just set it in config, or as a parameter for the Source constructor