Show media uploads to Discourse

#21

Yes, I’m a muppet. This works. Nicely done Tom.

1 Like
#22

Aces :slight_smile:
That SQL stuff is powerful voodoo

SELECT
CONCAT('https://discourse.southlondonmakerspace.org/p/', posts.id) AS post,
CONCAT('https://discourse.southlondonmakerspace.org', uploads.url) AS image
FROM post_uploads, uploads, posts, topics
WHERE DATE(current_date) - DATE(uploads.created_at) < 7
AND post_uploads.upload_id = uploads.id
AND topics.archetype = 'regular'
AND post_uploads.post_id = posts.id
AND posts.topic_id = topics.id
ORDER BY uploads.created_at DESC

I’ve temporarily bumped it up to 150 days so you can trawl recent history

1 Like
#23

Ah, any chance of user name in there too? Would be good to credit.

FWIW, this is the Python to get the data. Missing the accept header threw me for a loop.

#! /usr/bin/env python3

import sys
from urllib.request import Request, urlopen
from urllib.parse import urlencode
import json

if len(sys.argv) != 2:
    sys.exit('API key missing')

SLMS_DISCOURSE_QUERY_URL = 'https://discourse.southlondonmakerspace.org/admin/plugins/explorer/queries/1/run'
SLMS_DISCOURSE_VALUES = { 'api_key': sys.argv[1], 'api_username': 'makerspace'}

data = urlencode(SLMS_DISCOURSE_VALUES)
request = Request(SLMS_DISCOURSE_QUERY_URL,
                    data=data.encode('utf-8'),
                    headers={
                        'Content-Type':'application/x-www-form-urlencoded;charset=utf-8',
                        'Accept': '*/*'
                        },
                    )

with urlopen(request) as f:
    result = f.read().decode('utf-8')
    result = json.loads(result)
1 Like
#24

Ok, this also outputs usernames now

#25

This will get a bit meta:

Screen Shot 2016-03-22 at 14.09.57.png1256×366 41.9 KB

#26

This is a hackattack. Do not judge me.

https://gist.github.com/tobyspark/654528d9e3ae023a746a

#27

Pfft. It’s probably nicer code than my amateur-hour SQL :smiley:

#28

Heh. One more request. Haven’t seen any EXIF date info yet, and the creation date is my download. I’d wager these aren’t originals (resized and/or just stripped).

There’s more media types than just photos tho’, so regardless of whether they’re originals or not, could you add a column for posted date pretty please?

#29

Done. On my phone. On the bus. :sunglasses:

1 Like
#30

Close but no cigar?

Columns are –
['html$user', 'html$post', 'html$image']

#31

You’re requesting query no. 6?
Looks like you’re asking for no.1

#32

Query #6 indeed has the date (when you change query number?). And it tidies up the html cruft I had to work around, even better. But… now the post title has disappeared.

#33

You just want the title, or the link to the post as well?

#34

I just used the title for the file name. Long term, both might be handy.

(and realising what’s happened, I extracted URL and title from the HTML link)

#35

how about now then?

#36

Happily downloading via an updated https://gist.github.com/tobyspark/654528d9e3ae023a746a

1 Like
#37

Flickr now has 53 new photos, and some lovely @marketing can be done.

(It’s also just got five project posters back from E&C Mini Maker Faire, that I realised weren’t there when I went to link to Yulia’s jewellery)

#38

https://discourse.southlondonmakerspace.org/admin/plugins/explorer?id=1
(click run query)

#39

Speaking of this, I got behind on my Flickr-updating, and I’ve got a gap. I’ll need the request period increased or a special one-off, but hadn’t posted to say so as I’ve been busy with other things. Its… (goes to check)

#40

…I’m missing
2016-12-05-KylieJudd---Lasercut_lamp
to
2017-01-08-Jonathan---14_hour_flight_with_a_2_and_4_year_old