k4cg / nichtparasoup Goto Github PK

nichtparasoup is a web-based hackspaces entertainment system. It continuously displays random images from Reddit & Pr0gramm in your web-browser. Image sources are highly customizable.

Home Page: https://pypi.org/project/nichtparasoup/

JavaScript 6.90% Python 75.89% Shell 0.19% Mako 2.87% CSS 1.85% HTML 12.30%

image-crawler python reddit instagram pr0gramm entertainment

nichtparasoup's Introduction

Project: nichtparasoup

nichtparasoup is a web-based hackspaces entertainment system.
It continuously displays random images from Instagram, Reddit & Pr0gramm in your web-browser.
Image sources are highly customizable.

This project includes:

The nichtparasoup python-package.
An example for a plugin of the python-package.
The source code of the Web-UI for the python-package.

nichtparasoup's People

Contributors

Stargazers

Watchers

Forkers

senden9 jkowalleck fablabnbg makefu lu-chi poschi3 lassulus arboss weisk

nichtparasoup's Issues

imgur crawler broken?

the imgur crawler seams to be broken.
does not find images, finds whole pages instead.

reset reload timer on manual reload

when pressing hotkey "j" for the next image, it may happen, that two images are loaded.
possible issue: the internal automated reload process is not canceled and the timer is not reset.

fix reddit load loop

the reddit loader currently tries to load the same images over and over again.

fix: add a pager.
the next page can be found at: ['data']['after']
example-json: (http://www.reddit.com/r/images.json)

{
  kind: "Listing",
  data: {
    modhash: "",
    children: [], // something
    after: "t3_26nrhx", // <<<< this is what we need 
    before: null
  }
}

so the next is: http://www.reddit.com/r/Images.json?after=t3_26nrhx

pythonscript exits with (1) if you comment out crawlers (in my case, soup.io)

as title says, does weird things.

fourchan plugin do not stop when reaching the last page

instead they toggle between 9 and 10, they should be raising an exception or something once they reached the end of the last page to show that there is nothing to gather anymore?

UI: Change Bar Navigation from Text to Buttons

to have a better look&feel <3

Scale big images down to only use 80% of actual screen size

especially on imgur there are many big images, not fitting into the screen

If you reload the Page, all settings are set to default, which is kinda ugly if you click on the image source the lazy way (without middle click or command click) and get back to nichtparasoup. I than must click on "play in background" again because of a crazy Tab configuration.

Would be nice if this happens automagically.

Use image url as identifier

its somehow overhead to parse alt-text of images as the identifier. Image url is also unique.

pr0gram crawler broken?

tested the pr0gram crawler today, seams to be broken.
can someone confirm?

before rewriting it, wait for my go. i am currently writing some convenience methods for the crawlers ;-)

bundler not working

when running the bundler the resulting template is not working/no images are floating in ...

UI: add image source somewhere in the image

Add imgur as image source

faster image-grep with propper filtering

according to http://faq.soup.io/post/4328678/Filtering-your-Soup
you can use http://www.soup.io/everyone?type=image to fetch only images
instead of http://soup.io/everyone

UI last seen @end-of-the-world (a.k.a. UI nonfunctional)

80 per cent max-height (which should be 90 per cent) broken
max-width (see 1)
speed (itself)
speed (initial)
halt/continue (nonfunctional)
and everything else

Investigate use of InfiniGag instead of 9gag

http://k3min.github.io/infinigag/

Remove imgur.com image Provider

Reasons because imgur provider is crap

we use /random, so image quality is bad
we get more proper results (e.g. better images) from reddit located in imgur
the scraper is pretty much a mess, we only get 1 image by request
they would have a api but need some developer access with callback urls and all this weird oauth stuff

at the end the provider is useless. we get better, more efficient and even parts of the same urls out of reddit.

SoupIO Crawler may be broken.

Instance with enabled soupio does not start, does not log.
When SoupIO was disabled, it works again.

Will verify soon.

feature: UI add boss key

add a boss key [b] to the UI.
when it is pressed, every image is hidden and the loader stopps

when pressed again: images are shown, loader starts again ...

UI: Implement flush and reset buttons

for

/flush

and

/reset

Make crawlers forget thier last crawled urls

in order to be aware of any new content appearing in the crawled sites the crawlers should forget about their lasturls after ~3 hours so we get that content as well

UI continues loading images in backgroupd and dropps them all at once

Plugins which are not ...

... Plugins 😉

The basic idea of a plugin is to 'plug it in' and have the framework figuring out the rest. When creating a new plugin it is not enough to simply create a new crawler file in crawler/ but you actually need to touch nichtparasoup.py to import the new crawler AND define the url in this file plus the logic how to handle tags.

The plugins themselve should provide have logic to return a crawler with the correct urls and optional parameters like search tags.

Besides that 👍

character "☰" in "root.html" causes build to crash

when adding "☰" to "root.html" (currently replaced with "x") build.sh crashes:
Traceback (most recent call last):
File "/home/gpunktschmitz/Apps/nichtparasoup/templates_raw/_lib/buildHTMLfromParts.py", line 84, in
print("\n".join(str(part) for part in doc_string_parts_clean))
File "/home/gpunktschmitz/Apps/nichtparasoup/templates_raw/_lib/buildHTMLfromParts.py", line 84, in
print("\n".join(str(part) for part in doc_string_parts_clean))
UnicodeEncodeError: 'ascii' codec can't encode character u'\u2630' in position 94: ordinal not in range(128)

feature: gallery

every image should be part of a gallery. when one image is clicked the state should be paused (no new images) and a modal gallery (maybe css3-only) should be started.

icon for giphy

add an icon for giphy

add instagram crawler

enhance .gitignore file

use this template as the .gitignore file:
https://github.com/github/gitignore/blob/master/Python.gitignore

Change top-line slider font and position to right-aligned

maybe use IMPACT font or something similar

hotkey space does not toggle status/hotkey bar

the hotkey behavoir of the "space" key changed. i want the old one back.

desired status:

as a user i want to see the hotkey bar (bottom), when i press the "space" key.
as a user i want to see the status bar (top), when i press the "space" key.

current status:

on key down "space" nothing happens

boss mode image

y not http://img.webme.com/pic/d/daniela2312/grinzendekatz.jpg ?

Implemet syslog possibility

to be able to get rid of strange supervisord wrappers and stuff.

Backend: Implement cooldown timer for Reset and Flush

to deny bruteforce flushing

UI: annoyence fix > image source link should be a block

currently it is hard to click on an image source, since it needs moving the cursor to the source text.

to have this fixed, the css property display of the <a> should be block.
this will put the a as wide as possible.

i hope to get this fixed today.

review the JS and make maintenance easier.
image laoder fix: just one image at a time (currently multiple are loaded sometimes.)

reddit parser: add paging

the reddit loader currently tries to load the same images over and over again.

fix: add a pager.
the next page can be found at: ['data']['after']
example-json: (http://www.reddit.com/r/images.json)

{
  kind: "Listing",
  data: {
    modhash: "",
    children: [], // something
    after: "t3_26nrhx", // <<<< this is what we need 
    before: null
  }
}

so the next is: http://www.reddit.com/r/Images.json?after=t3_26nrhx

UI bundler/builder without Java

have the UI bundler/builder work without java.

Add 9gag image in UI for 9gag content

its missing at the moment