Ji Ma

I `set_html(html)` because the HTML file is local. I run the exact code you gave but the system still hangs and RAM keeps increasing. How long does it take you...

[test.txt](https://github.com/codelucas/newspaper/files/7207502/test.txt) Thanks for the example. I tried, but still not working. Here are the code and file: ```Python from newspaper import Article article = Article('', language='en') article.download(input_html=open("test.txt", 'r').read()) article.parse() ```

Okay, I tried your code on a Windows machine, it works! I'm using a Linux server, still not working ...

I'm using `OS Ubuntu 18.04.5 LTS`. I've also updated the first post.

Ref: #86

Refer to #86

You are very welcome!

> If you use the bulkprocess feature in the dev branch, it will be several times faster for tokenizing small text. Here is an example. Calling it with pre-created documents...

@liutianling on GPU.

Just wondering: do we have a best practice for using Zotero's cite-as-write? has this function been integrated into LaTexTools?