Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: RSS Search Engine (feedi.me)
192 points by daviducolo on May 8, 2019 | hide | past | favorite | 48 comments



Honestly, this might be the best way out of the mess the web is becoming. Use RSS for "normal" content delivery. No scripts. No styling. Nothing crazy. Just content. Build tools for searching and some other typical tasks and stop when basic functionality is covered.


This would be great for consumers but won't content creators or publishers be unhappy about it? For any content you don't pay, advts is a motivation for people to create content.


They'll just demand the ads be put in the headlines.


I just submitted my own RSS feed using curl and jq, by doing this:

  $ RSS_URL="https://www.grepular.com/rss"
  $ FEEDI_TOKEN=$(curl -sX POST https://feedi.me/tokens \
    | jq -r '.data.attributes.key')
  $ curl -X POST https://feedi.me/feeds \
    -H "Authorization: Token $FEEDI_TOKEN" \
    -H "Content-Type: application/json" \
    -H "Accept: application/json" \
    -d "{\"url\":\"$RSS_URL\"}"
As per the API docs at https://github.com/davidesantangelo/feedi


This is cool a cool project. RSS needs all the support it can get. I understand this is just a demo but some things confused me.

What are the positive and negative tags for, how do they get applied?

In what order do the search results get returned? It didn't seem obvious when I added my own feed and then searched.

Is there no way to search for a particular feed or author?


> What are the positive and negative tags for, how do they get applied?

If you hover over the tag, it shows you a score that represents a "sentiment". The README says "entity extraction and sentiment analysis are provided by Dandelion API".


I like it.

It would be great to be able to use UI instead API for usage. User login, add feed, end.

FYI, without RegEx any search is insufficent.

I currently use desktop RSS QuiteRSS that has this option.

https://quiterss.org/sites/default/files/screenshots03_0.png

Also, search query should be linkable (give me URL that I can send others for them to get the same results)


anyway at the moment you can share with https://search.feedi.me/?q=ruby


"Also, search query should be linkable (give me URL that I can send others for them to get the same results)" Yes thanks for feedback.


yes but main project is Feedi API. https://github.com/davidesantangelo/feedi.

This is a small demonstration of its potential. However in the future the idea is to make a real UI


What does this do, exactly? This isn't clear to me from the title, and the page isn't doing anything for me besides showing a loading gif and linking to a Github repo that still has me wondering what this does.


you can add your RSS feed and feedi takes care of periodically importing the data, enriching it with ML and providing you with the API.


You should add some instruction copy then.

Also: all I get is a "network error" for the ajax response in the browser console.


that's cool, but what about the comments that make ecosystems like HackerNews + Reddit so great?


>Reddit

Which ones? (Subreddit)?


Thank you for sharing! Loving it already :)

One question: Are you feeding this application manually with RSS feeds from different websites or do you have a crawler in place?


I want to chime in - thanks for sharing & good work!


thanks! There is an endpoint to add manually at moment.


Great projects. Going to keep an eye on it/use it.

I've been wanting to do a web RSS thing for a while, but my idea has bloated into a full featured RSS reader and blogging Platform. It would also need a search engine, but because of the the size of the project I never got very far with actually implementing it. I need to get better at finding a minimal functioning part and start implementing it, like this search engine does.


this is a simple reactjs app around my personal project feedi take a look at https://github.com/davidesantangelo/feedi


I don't quite understand how the searching works; if I search for "Steve Kemp" I expect to find my own feed, or nothing. Yet instead I get results just including "Steve".

(Wrapping in quotes didn't make a difference, nor did adding AND in the middle.)


You should configure your Rails application correctly for production. Right now it seems like it's considering all requests as local and returning the Rails exception page to me.


if you try this https://feedi.me/search/entries?q=SEO&page=2 you receive an exception because Total are 10. I must improve exception management


can you send me an example of your request?


Absolutely. I'm talking about things that should 404 are actually raising exceptions. E.g. https://feedi.me/popular. I hit this link by mistake but that should return a 404 and not the exception page from Rails.


This generally happens because config.consider_all_requests_local is set to true in production env. By default it should only be true in development and test.


yes because the correct request is https://feedi.me/feeds/popular not "https://feedi.me/popular"


Yes I understand that. But visiting your Rails app shouldn't bubble up the debugging info to your visitor in production.


Anyone else getting this error?

    Error fetching and parsing data Error: Network Error
        at M.exports (createError.js:16)
        at XMLHttpRequest.W.onerror (xhr.js:87)
    Failed to load resource: net::ERR_SSL_PROTOCOL_ERROR


I take a look at this soon


Wish I could search for an exact phrase. I'm getting non-relevant results.


I would make the search input more obvious. My focus is dragged first to the "query, entries, per page" buttons because of the gray background.


The post is about the API, not the rough demo implementation of it...


done!


Unfortunately the page only shows the loading gif (in the left top corner) and nothing else happens. Tested in Firefox 66.0.5 (64-bit) browser.


me too. tried it on firefox and chrome


For a more general-interest version, check out https://yetigogo.com


That's very nice work ! Do you crawl sources as well or do you rely on people submitting feeds to you ?


people submit feed and the app cares to keep them updated


How can a feed be submitted to the index?


take a look at API documentation at https://github.com/davidesantangelo/feedi



just deployed a new version with various improvements! Go to https://search.feedi.me

Thanks for feedbacks!


Doesn't work in FF


Scrap that, doesn't work for me in any browser now I look again. Maybe my network is messed up.


Works perfectly well with FF + uBlock + uMatrix (without modifications!!)


FF user here, it does work for me. (v66.0.4 with ublock)


ah ok i take a look soon! Thanks




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: