It seems, pdf.js itself requesting whole byte range requests of a PDF file. Instead, is it possible to request only 5 pages on PDF load, On scroll can able to load another set of 5 pages, like that.. Is there a way to achieve this by using pdf.js ?
Long story short - No.
PDF is not a contiguous storage format. If the PDF file is formatted for fast web view then you can get it to show page 1 whilst other pages are still streaming in, but you can't ask to start at a specific page or page range. Internally pdf uses a bunch of sections, links/pointers between them and digests. Think of them as wooden blocks with bits of string between them. You can't render anything until you have 'enough' of the file to provide the parts you need, but the organisation of the internal sections is pretty much random as far as your question is concerned.
The only way to get specific pages would be to have a server-side component split them out of the PDF file for you and make a new PDF file containing just those parts, but paging on to page 6 would mean opening a new document, etc.
Edit: There are startup params for Acrobat viewer that could allow you to set the first page to be displayed, and other viewers may offer this feature, but unless you have some very smart client-server interaction this would still require the entire PDF document to be present in the client first.
Edit 2: As per comment from #async5, PDF.js 'may' be able to do page-range loading. See this section of the PDF.js docs. But note that there are requirements on the web server that is serving the PDF file.
As described in an issue here, old versions of PDF.js did not handle linearized PDF files properly(as described by Peter in comment, when you try to load page 1000 it loads page 1-1000).
It seems the problem has been resolved at some point (I dont know specific version #) and now the behaviour when you set those params correctly (namely disableAutoFetch and disableStream both to true) and load page 1000, it would only load page 1000.
Related
I want to download hundreds of pdf documents from a site. I have tried tools such as SiteSucker and similar, but it does not work, because there appears to be some "separation" between the files and the page that links to them. I don't know how to describe this in a better way, since I don't know that much about website programming or scraping. Any advice on what this can be and how one can circumvent it?
More specifically, I am trying to download pdfs of UN resolutions, stored on pages like this one: http://www.un.org/depts/dhl/resguide/r53_en.shtml
It appears there is an in-built "search function," on the UN site, which makes dummy scraping, like SiteSucker, not work as intended.
Are there other tools that I can use?
Clicking a link on the page you mentioned redirects to a page composed by two frames (html). The first one is the "header" and the second one loads a page to generate the PDF file and embed it inside. The URL of the PDF file is hard to guess. I don't know of free tool that could scrap this type of page.
Here is an example of the url in the second frame that ends to the PDF file:
http://daccess-dds-ny.un.org/doc/UNDOC/GEN/N99/774/43/PDF/N9977443.pdf?OpenElement
I found this link that describes how to add content to the PDF. So as I understood right the pages will be generated based on input NSString. So the count of pages depends on the length of the string.
I need to implement the same but with the images. So for example I have 10 images, and all of them will be places on by one on each page. Like first image on the first page, second images on the second page and etc.
I found also this library that can generate PDF files but seems I don't have chance to set count of pages. Is there any good solution how to implement it?
I can upload large document as pdf file into web page no problem. but i want to use arrows to navigate the book pages not to upload the whole book at once as this may take long.
can any one help how to do this in mvc app with or without database? if database is necessary would Mongodb be a better choice? i do not want people to download the book; they can just read it online?
First you cannot prevent people to download your content if you visually display it BUT you can discourage them by making it difficult to do so.
That being said you wouldn't have a need a database to do what you want to do. You can but it's not necessary. You can simply find some library online that handle PDF such as iTextSharp cut the book in 1 PDF per page with it when it get uploaded so you have bunch of small files.
Then the trick is simple you query the PDF library to load the file Page1.PDF (arbitrary name) extract text format and output as text nicely has HTML. when the person click the link Page 2 then reload the page with the new PDF to use for display.
Doing so prevent the user from seeing or having access to the PDF file itself and if he want to download it all he will have to copy paste every single page manually or by code if he's a dev. Most common user wont go around copy pasting manually 300 pages because of laziness.
What i would personally do is each file uploaded i would create a folder with the name of the book and call the files 1.pdf, 2.pdf .... per page. Like that if i query the listing of directories i get the list of all books, and if i check the count of files in it i know the total page number. That would allow me to run all that without database.
Let's say I have a 400MB PDF, on a web server, and I want an ipad user to be able to open it and to start looking at it as soon as possible, without downloading the file entirely first. What are the options?
Is Safari able to stream a large PDF like this? Can it start showing the first pages while the file is still downloading?
Is there a way to build a native app to achieve this? If so, should the PDF be "spitted" first on the server? How?
Any tips on how to open a large PDF, in a friendly way and on an ipad, would be appreciated. Bonus points if the index of the PDF is accessible!
To clear things out:
Linearized PDF has its order of objects reordered, so that all required data to display page are in front of the others. You can read more details in PDF specification in Annex F (page 683)
In case You haven't found application, that supports Linearized PDF, the best solution from my perspective would be rendering pages at the server side and create custom protocol to transfer only a single page with the table of contents.
Unless You want to create custom PDF viewer with Linearized PDF, You just create small portable "page viewer", which can quickly show the table of contents and then ask server for exact page number.
Example Server Interface:
class Pages
{
string title;
int number;
}
Stream GetDocumentPage(Page n);
List<Page> GetDocumentPages();
For this, one can use open-source solutions (such as Ghostscript) to parse pdf file and render the required page. Then send the binary data over Your custom protocol and display page at Your small portable "page viewer".
This solution does not require downloading whole document as it can cause significant network traffic.
Hope this helps.
Is it possible to display a pdf from a partial download?
I need only the first page of a pdf for my app. The problem is all the PDF online are 25mb or more in size. Optimizing for the app is not an option :(
The entire PDF will need to be downloaded to display and save it, but I want to show a preview first.
A similar question, but for android:
How to Display first page of PDF before downloading is completed
I do understand downloading of data in iOS, but how can I tell where in the PDF's data the page ends, so I can just display that.
Yes you can do this but the PDF needs to be pre-constructed in a linearized format. This is something that is part of the PDF specfication and is sometimes known as fast-web-view.
Linearized PDF is the same as normal PDF but the objects in the document are ordered in a particular way and with certain extra information which makes it possible to work with partial data.
In particular the objects for the first page are included at the start of the file specifically so that the first page can be displayed quickly.
So I see no reason you shouldn't download the objects at the start of the PDF and use those to display the first page. You could use the hint tables for fast access to selected other pages but that would be quite complicated.
However the essence is that you need to pick up the group-one objects for the first page. These should run from the "%PDF" header through to the first "%%EOF". I'm not sure whether your environment will complain about the missing (but not required) objects but if it does you will need to blank them out on a binary level so that you have an internally consistent page one PDF.
For full details on PDF linearization see the Adobe PDF Specification.
My answers may feature concepts based around ABCpdf .NET. It's what I work on. It's what I know. :-)