Include PeerTube (Sepia Search) to about ytfzf HOT 10 CLOSED

Does anybody know on how to get peertube api? Or we can scrape the site just like youtube.

I think we can just scrape it.

Edit: Sepia search is a search engine for peertube, so I assume we can definitely scrape it like YouTube.

That being said I'd also like support for:

• LBRY
• Odysee

Maybe there can be a flag that brings up a menu that lists different video sources like YT, LBRY, etc. and when the user selects one of them, it presents a prompt where user enters search string and presses Enter and then ytfzf will behave just as it is now, but for the chosen site.

from ytfzf.

Yea flag will keep the logic modular for diff. websites. and also break less as new sites are added.

from ytfzf.

Actually, Sepia Search has a JSON API that can be used like this:

https://sepiasearch.org/api/v1/search/videos?search=stallman

I don't know where this is documented, but I found it in NewPipe’s code.

Edit: I was told that Sepia Search uses an empty PeerTube instance as a search backend. So the API is documented here.

from ytfzf.

Also invidious would be nice to be supported as an alternative source for youtube videos. It should work without javascript as well, so getting the results should be fairly simple.

from ytfzf.

Noted. We will add this to the TODO's and add sites one by one

from ytfzf.

wow, nice, I will be trying to implement this.

from ytfzf.

Sounds interesting, just found out that you can play peertube videos with mpv and ytdl. I will need to go throught the peertube api.
Does anybody know on how to get peertube api? Or we can scrape the site just like youtube.

from ytfzf.

The current script relies on the json file from youtube's search result page.
A browser would run the javascript and render the search results.
ytfzf scrapes the contents of the java script's json, and instead of running any java script, it uses jq to parse the json object and render the video results.

from ytfzf.

I just had a few minutes to think about this an wrote a few lines to produce videos_json data like in Ytfzf's code but for Sepia Search instead of YouTube.

# query string for testing
search_query="Hello World"

scrape_pt() {
     #gets a list of videos
     videos_json=$(
     curl \
         -s "https://sepiasearch.org/api/v1/search/videos" \
         -G --data-urlencode "search=$*" |
     jq '[ .data | .[] |
         {
              title: .name,
              channel: .channel.displayName,
              duration: .duration,
              views: .views,
              date: .publishedAt,
              videoID: .uuid,
              thumbs: .thumbnailUrl
          }
     ]')
}

scrape_pt "$search_query"

# print first three entries
printf "%s" "$videos_json" | jq '[.[]][:3]'

The output of the above looks like this:

[
  {
    "title": "Hello World",
    "channel": "Main requiem channel",
    "duration": 58,
    "views": 7,
    "date": "2019-12-30T18:12:19.662Z",
    "videoID": "e3fbacb5-2add-4f45-8bec-7ee565dddc8f",
    "thumbs": "https://video.hackers.town/static/thumbnails/e3fbacb5-2add-4f45-8bec-7ee565dddc8f.jpg"
  },
  {
    "title": "Hello World",
    "channel": "Main programmingfish channel",
    "duration": 6,
    "views": 16,
    "date": "2019-07-07T13:26:01.663Z",
    "videoID": "7cf70066-7775-4d2f-a9c2-44c83fd73955",
    "thumbs": "https://tube.piweb.be/static/thumbnails/7cf70066-7775-4d2f-a9c2-44c83fd73955.jpg"
  },
  {
    "title": "4- Hello World",
    "channel": "LearnAsYouLike",
    "duration": 325,
    "views": 0,
    "date": "2020-11-15T04:00:21.357Z",
    "videoID": "a2576ad6-1efb-4f06-917b-b32bc3f44fdf",
    "thumbs": "https://diode.zone/static/thumbnails/a2576ad6-1efb-4f06-917b-b32bc3f44fdf.jpg"
  }
]

from ytfzf.

Thanks to @Euro20179 this has been implemented in #281 (not yet merged to master,
to use development branch ytfzf --update-unstable)
Search peertube with -P flag

from ytfzf.