Alt: [email protected]
Has gone suprisingly well.
Tesseract failed in some places, making some of the sub-headings come out in what looks like Klingon. HF have varied their paper stock dimensions as well, which caused a few things to be clipped.
Acceptable output for manual corrections.
Preview:
Expect a DM soon-ish.
Good news, they went through the ADF without too much trouble. Images just need a little contrast adjustment.
The copier used also OCR’d the whole lot, so they’ll be searchable.
Welcome.
Didn’t realise just how many I had - a whole box file full and overflow from that.
I’ve had a nice morning so far sorting, de-duping and remembering the time I coated half the kitchen in sesame seeds…
Been meaning to do it for ages so cheers for the incentive!
Python, Tesseract, OpenAI and my 3 remaining brain cells have now combined to form a working script that will rename the scan file names to whatever it reads in a certain section of the card.
Doing them by hand would be a nightmare 😅
Now I am wondering the same.
And looking down the rabbithole of Tesseract OCR haha.
I’ve done less useful things with a Monday morning before. PDFs, and I will try to suss a way to set the file names programatically.
A handful may come with ‘pre-printed accidents’ but all will be legible :)
As long as the scanner can handle the slightly thicker paper stock they use, we should be golden.
If you don’t mind them auf Englisch, I’ve got a hoard of them and time with a duplex scanner.
So they do.
That’s really neat!
Site either respects system light/dark theme or is dark by default. Either way, it’s a plus for me that is often forgotten.
Guides are simple enough and the illustrations are cute.
You’ve got the balance of plugging your own instance while reminding visitors that there are others just right.
All useful info and not so much of it that it is overbearing.
I’ve had some decent results with this:
I lean on Irfanview+plugins for this kind of processing - it’s the bedrock of some of my workflows.
Will definitely give this a try though, thank you for sharing :)
Server seized by the feds, but zero useful info retrieved by them.
The gold standard.
The USP here seems to be the traffic mangling used to disguise the connection. The client is open source, which is always good, but that means it’s probably easier for a bad actor to analyse and break that USP.
Very limited payment options and macOS only currently. Feels like they’ve made a pretty webpage & rushed launch, instead of polishing the product/product availability first.
Unfortunate that the term ‘Obscura’ is dominated by the band in search results.
I will stick with stuffing an envelope with cash and posting it to mullvad, but this is definitely one to watch.
I buy and sell drives occasionally. Even shitty old IDE drives with 5, 10 years of runtime are worth something to someone.
Bought one recently just to get an old lathe PLC (really just an old PC with DOS software control) running again.
It would be better if the server could detect bots and send them down a rabbit hole
Already possible: Nepenthes.
I just treat their letters as scrap paper or kindling. They are very carefully worded to give the illusion of power where there is practically none. Capita are masters in mismanagement, so I’m not surprised your declarations have been ignored in the past. Just don’t bother.
If you’re truly not doing anything required to have a licence, then they can’t prove you do. Licence dodgers are usually clever enough to not give it away too.
Don’t answer the door to them on the rare off chance they come prospectively calling. If you do, just close it on the scum without a word, and go about your day. No warrant = no entry.
Sounds like a good way for Mr Pirate Fire Stick Man to sell more of his £50/year IPTV packages.
It’s already easy enough for the average joe to do this. Having the BBC double-dip their Netflix sub may well be the push that has them cancel subs & go ‘elsewhere’.
If we want a public broadcaster, it should be recognised as a national service and funded directly through tax. Unfortunately, that would put a lot of Crapita licence fee collector dogs out of a job and we simply can’t have that(!).
Just look at the crap you get shoved through your door if you decide not to play their game: http://www.bbctvlicence.com/
Before anyone mentions that you can submit a ‘no licence needed’ declaration - they’ve got no business even knowing my name, let alone my media choices.
Monetizing is what ruins other places.
I like the way my home instance does financial backing through an open model, and that’s part of why I chose it.
An ideal is enough contributors to keep the lights on and to reimburse the admins for their time spent in keeping it afloat. Moderation should always be a volunteer position for those that want to support their individual communities.
Any excesses in finance I would hope go towards future running costs (to a point), feature development and then charitable donations in that order. Non-profit on paper and in practice.
This is viable for a small instance. Maybe even larger ones if the users are altruistic enough as a whole.