bitfox-git / hackernewspaper

The HackerNewsLetter Paper

hackernewspaper

This is the hackernewsletter but in the format of a magazine, a pdf edition. Well, at least I tried to look like a magazine 😄 The latex based pdf contains the hackernewsletter edition including the categories, screenshots of all the url's, some metadata, and the 'gather' text from the url, but don't expect too much of the text sofar.

As an example: HackerNewsPaper-667. All the hackernewspaper editions can be found here, in the releases of this repo.

Motivation

I ❤️ the hackernewsletter. But I also cannot remember the links I've already opened, viewed, readed, watched. Only the name of the link in the email is not enough (for me..) to memorize what i've seen and not. But I do remember, as soon as the page opens in the browser, at that moment, i can exactly recall most of it. I just wanted to have a quicker overview of the newsletter. And the screenshots of the url's helps me achieve this.

But ended up with much more....

Screenshots

I use playwright to generate screenshots of the links. Or pythumb in case of a YouTube link, to get the main thumb of the video.

Credits

Praise and credits for Kale Davis, as he is doing all the hard work of handpicking/organizing the articles each week. If you like this hackernewspaper, please also subscribe to the original hackernewsletter at hackernewsletter.com! Disclaimer: Hackernewsletter and Hackernewspaper are not affiliated with Y Combinator in any way.

Want to help?

I'm not a Latex expert. And there are also a lot of nice to have refinements to think of:

• Parsing the top content (some times Kale leaves a message) on the top of his newsletter, is not done.
• Parsing the sponsor content is sometimes off..
• Nicer output , more layouts in the template
• Solve the GDPR Cookie problems on some screenshot.
• Better text formatting. Maybe with the help of a LLM?
• For non-html links, devise another method to generate a screenshot and extract text and metadata. The link in the newsletter is not always pointing to html content (think of direct links to MP4 files, PDF files, etc. )
• [X/50%] In case of YouTube / Vimeo url's , get the duration as metadata + description from the platform. Which requires a Youtube Video API v3 key...

hackernewspaper's Issues

Add YouTube / Vimeo text from corresponding API's

For YouTube posts, the description/text is not available as it can only be requested by the corresponding API's.

Add PDF support

See the HackerNews #679 , article with title 'Charlie Munger has died'. The link references directly to a PDF file.
The paper doesn't contain a better screenshot (some JPG/screenshot of first page) and no content is presented (extract text from PDF).

Auto new hackernewsletter detector to trigger main build

The current hackernewsletter Github action is run manually, after updating the issue variable in the settings.
Normally, a news hackernewletter is send around friday.

Develop a second, scheduled, Github action, that tries to determine if a new newsletter is already published and if so, update the issue-variable in settings+trigger the main Github build action. Schedule it to run at friday night.

Solve warnings in latest Github actions

Solve the warnings (nodejs version) in the reports of the latest Github actions