- 52comments
- 198comments
- 9comments
- 45comments
- 47comments
- 23comments
- 55comments
- 44comments
- 150comments
- 26comments
- 154comments
- 58comments
- 8comments
- 34comments
- 1comments
- 346comments
- 43comments
- 128comments
- 18comments
- 97comments
- 26comments
- 8comments
- 20comments
- 38comments
- 18comments
- 253comments
- 65comments
- 1022comments
- 41comments
- —discuss
It runs in your browser like a browser extension. It scrapes the HTML and navigates from page to page.
Setup and usage instructions are in the file.
Check out https://gabrielsroka.github.io/getHNFavorites.js or to view the source code, see getHNFavorites.js on https://github.com/gabrielsroka/gabrielsroka.github.io
It's a shame favorites aren't exposed in the official HN API: https://github.com/HackerNews/API - this is a smart workaround.
Our plan is for the next version of HN's API to simply serve a JSON version of every page. I'm hoping to get to that this year.
Hi Dan,
That's great news! Is there a way to be notified (eg, via email) when this comes out?
Thanks.
If you (or anyone) want to be on an alpha-tester list, email hn@ycombinator.com and we can add you. Send a username and make sure it has the email address you want to be notified at.
That would be amazing!
I've been having some fun with the API recently building this tool: https://github.com/dogsheep/hacker-news-to-sqlite
Are there plans for an export tool, e.g. a user downloading all their comments and upvoted submissions? I tend to use the submission upvote button more than the favorite one, and an export tool wouldn't require a user API key for non-private info.
That should be an easy extension of the JSON API once we have it.
My tool here can export all of a user's comments: https://github.com/dogsheep/hacker-news-to-sqlite
Not upvoted yet as they aren't in the API I am using.
Wow, that would make it really easy to implement an alternative HN client.
Related Question: Is this the source code for HN? https://github.com/wting/hackernews
That page says the code is copied from http://arclanguage.org/install, which is indeed an old version of the HN source code. It has changed a lot over the years though.
Any chance of some newer code being released, or the addition of a place to submit patches?
I'd love to do that someday. But it would be a lot of work.
That's great! Is there any plan of exposing authenticated content through the API too? Mainly talking about upvoted stories.
Thanks Simon. I'd originally written this script to export my DVD ratings from Netflix, since there's no API for that either. It was easy to adapt it to HN.
I wanted to show people that it's possible (and easy) to get to your own data!
This is smart. I'm adding this to my HN favorites.
Thanks dvfjsdhgfv. If there's sufficient interest, I can easily turn it into a Chrome extension.
(Edit: haha, I see what you just did there. A little recursive humor.)
I made an API for favorites a while back.
https://github.com/reactual/hacker-news-favorites-api
Here's an example query:
https://reactual.api.stdlib.com/hnfavs/?id=sbr464&limit=1
Ok, now that's really smart!
Would a client using your API paginate to fetch, say, 50 pages? When I tried it using ?limit=50, I got a 504 error.
Thanks!
(Edit, never mind, I see you explain it in the readme.)
I had made it pretty quickly, the limit would be too large, thats 50*30, you may need to provide an offset and make a few requests.
Interesting that it is using x-ray, seems like x-ray is still using PhantomJS as the plugin, is PhantomJS deprecated? Would it be using Puppeteer instead?
What is HN's GDPR compliant way of requesting a copy of all stored data? Email dang?
Considering that there is no mention of GDPR in the HN FAQ or on the "legal" page, my guess is that their position is that GDPR does not apply.
According to Article 3 of the GDPR, it applies to:
1. Processing that takes place in the context of processors and controllers that are in the Union, regardless of whether or not the processing itself takes place in the Union.
2. Processing the data of subjects who are in the Union by controllers or processors who are not in the Union if the processing is related to offering goods or services to such subjects in the Union or the processing is related to monitoring the behavior of such subjects that takes place in the Union.
I don't know how HN is structured, but I've not seen any indication that they are in the Union, so #1 probably does not apply.
#2 applies if they are doing processing related to "offering goods or services to such subjects in the Union" or "monitoring the behavior of such subjects that takes place in the Union".
One of the recitals elaborates on the first branch of that:
Does HN "envisage" offering services to people in the Union? Or are they a site that is merely accessible from the Union without envisaging offering services there?
There's a recital that elaborates on the second branch, too:
Does the data HN stores about its users satisfy this? And if it does, is the behavior being monitored taking place in the Union?
Could someone convert it to Python-script?
Part of the advantage of running JavaScript in your browser is that you might already be authenticated and it can use your session. But, fetching your HN favorites doesn't require authentication.
Can someone convert it to Bel?
Thanks!
One more question: what is the best way stop it when it will reaches last page?
Actually p=17[0] is empty (as p=16 is maximum as for now).
Maybe, script should scrap pages from `1` to `infinity` UNTIL it detect next message on page[0]:
[0] https://news.ycombinator.com/favorites?id=app4soft&p=17
In Python, `range(1, 17)` produces the numbers 1-16. I hard-coded it just for your favorites.
A better way to solve it is to look at the `len()` of the results, and stop when it gets to 0:
Great!
I think this one is a little cleaner. I used some of the ideas in sbr464's code.
Go on! ;)
I had written a Python script to get saved (upvoted) links as JSON / CSV a few years ago: https://github.com/amjd/HN-Saved-Links-Export
I'm not sure if it still works as it too relied on HTML scraping. Perhaps I should update it to support favorites too.
Edit: Whoa, it's been 4 years already. I believe HN didn't have favorite feature at the time. That's why I used upvoting as my bookmarks system and created a script to export that data.
@amjd, thanks for sharing. I upgraded it from Python 2 to Python 3, but when I ran it, I got a 404 error on the `saved` endpoint. Does it work for you?
Edit: I see from the other PR it's called `upvoted`.
Edit 2: I changed it to `upvoted` and now I get a 200 OK, but the code crashed right afterwards on `tree.cssselect()`.
It's four years old so the page HTML structure must have likely changed. I'll take a look at it when I'm free and update the script. Thanks for your PR, I'll review and merge the two open PRs as well.
Btw, good work on your JS solution. It's great that it just works without requiring any download or installation. :)
This is cool, I love HN metadata, too :)
Plug for a related golang tool I wrote and use which exports favorites upvotes as structured JSON:
https://github.com/jaytaylor/hn-utils
Just
This is great! My biggest problem was I couldn’t search through my upvoted items to find the article I liked again. I used google custom search and cleaned the data as flat urls.
https://www.heyraviteja.com/post/projects/deep-search-hn/
oops - I didn’t realize that favorites != upvoted.
Can it will be done with "Scrape Similar" Chrome plugin?
Thanks for the tip. I gave the "Scraper" extension a try, and 1) I got an error, 2) it only seems to scrape 1 page -- it doesn't paginate (or, did I miss something?).
I used the jQuery selector `a.storylink`.
Is there any way to find most Favorited items on HN ?