lionel wrote:
I don’t understand why you write that. To access Mainbullau airfield, you click on “AD Flugplätze”, then “M”, then “Mainbulla EDFU”. I would say it is more access by ICAO code that is missing.
I don’t care about the aerodrome part, because I do have another source for them, but not being able to search the documents? Com’on it’s 2023. I just wrote a batch job to pull all of the AIP, feed it through OCR and generate aggregated searchable documents. What a waste of time only because they did not simply put the PDF there.
I wonder whether one can use a web crawler to download all the images and then put them into a PDF…
MichaLSA wrote:
I don’t care about the aerodrome part, because I do have another source for them, but not being able to search the documents? Com’on it’s 2023. I just wrote a batch job to pull all of the AIP, feed it through OCR and generate aggregated searchable documents. What a waste of time only because they did not simply put the PDF there.
Looks like MichaLSA already did something similar…
wbardorf wrote:
Looks like MichaLSA already did something similar…
Framegrabber and via an OCR before building the PDFs. It is super easy, but not very satisfying because so needless to OCR a picture done from a PDF to generate a PDF … feels silly indeed.
MichaLSA wrote:
Framegrabber and via an OCR before building the PDFs. It is super easy, but not very satisfying because so needless to OCR a picture done from a PDF to generate a PDF … feels silly indeed.
Agreed although presumably the OCR should therefore work 100%
50% of freelancer.com does website scraping.
It is the second oldest profession.
There are even free spider tools. I used to have one and it was brilliant but didn’t do HTTPS and I have not needed it since.
Reports from German pilots claim that the DFS has sued airports which had their VFR plates on their website (only for the said airport) so maybe one needs to step carefully, or keep it as a private site which then defeats the point.
If you run a private site then you may as well host Jepp VFR Europe and sidestep all this BS 
So the charts seem to have been scanned at 150 dpi which is not great but ok.
I really wonder why they published the AIP but did so in such a shoddy way. This is basically state-financed trolling.
There are two explanations,
I am more inclined to think it’s the former
The amount of spilled taxpayers money On this is just huge, I am sure BASULM or Pooleys VFR plates cost less than 20milions…
This is basically state-financed trolling.
Yes – it is really outrageously cynical way of complying with some directive from above. It’s like your 5 year old is banned from eating his 5th ice cream that day, so he licks it instead.
150dpi is not a lot but if it is ex-PDF (a PDF has “infinite” font resolution) then OCR should work well enough for extracting text for searching. The data published would look better as 150dpi graphics though.
Technically this is straightforward to do; scrape the site and re-host the data, but I would not do a straight spider job and just replicate the site layout, because like I say above DFS could go after you on copyright.
It will just take someone many hours to do a decent job, plus 5-10 a month for hosting it. It needs to be HTTPS (to avoid increasingly dire browser warnings; some ISPs actually block HTTP websites – a wonderful demo of obscession with “privacy” coupled with human stupidity) so the certificate needs to be sorted out. And unless you front it with say Cloudflare, this certificate will give you regular hassle when it expires and the cron job you set up breaks (it always does). Also remember that the site will be scraped, and copies of it will appear elsewhere in Germany. There is a server in Germany which is diligently maintaining a copy of our airport database Why? I don’t have much of an idea (but I do have “some” idea, given previous posts here and on the German domestic site, pressing EuroGA to allow unrestricted republication rights) but they can’t openly republish it. Your site will also be attacked, roughly once a second or more, so you need up to date OS / web server / competently done coding (the standard anti-XSS stuff, as a minimum). “Everybody” knows PHP + Maria DB but few know the whole process. Our airport database was ~200 man-hours.
A smarter way is to host the 150dpi images from the open DFS site but doing OCR on the PDFs from the subscription DFS site. Nobody will discover you did that. I never wrote this 
Ibra wrote:
DFS never published VFR AIP as it was orginally in this c***p format
It is charitable that you entertain that idea, but I assure you that all these years when I phoned AIS-C and they sent me by email the pages I needed for a certain flight (so in a way, access has been free at least since I started flying, only inconvenient), I always got PDFs that contained vector graphics and no bitmap images at all. Even the text was converted to vector shapes, not in the PDF as text & fonts.
