search

tcbyrd / papers-ui

Papers for CAA 2017 at Georgia State University. Pulls data from Airtable and generates a printable PDF.
2 stars 1 forks source link

PDF should be generated automatically #1

Closed tcbyrd closed 7 years ago

tcbyrd commented 7 years ago

Right now each page needs to be printed as a PDF and combined in Acrobat before being manually uploaded to the /public/ link. This should be more automated in one of 2 ways:

To make this work, there are a couple of anomalies that had to be manually edited in the PDF:

It would also be useful to generate just the "Program" that only has the days + posters.

tcbyrd commented 7 years ago
  • When Organizers are blank, it's printing it as Organizers: , , , ,

Fixed in https://github.com/tcbyrd/papers-ui/commit/d9898d11511294bf8842c188c25227185169fb81

  • Other polish around line wrapping for the Organizers

Fixed when changing font to Arial in https://github.com/tcbyrd/papers-ui/commit/c36c2d4abe9253c6237448b5829ccd9c7ec5011b

  • GSU Logo needs to be a local resource instead of a link

Fixed in https://github.com/tcbyrd/papers-ui/commit/9e340084fc2883aed1e225050214460ee1b7e000

tcbyrd commented 7 years ago

Generated on the fly by combining all days, posters, and abstracts into one view that is "printed" together

Cloudconvert seems to be able to do this, but it's not respecting the 2 column view. When I tested on their demo site, it also kept putting Thursday before Wednesday, even after manually setting the order. Positioning should be able to be controlled through the API, but it still needs to respect the 2 column view for it to render properly.

I submitted a request on Cloudconvert's contact page to see if they have any way around this, but haven't heard back yet.

Alternatively, I looked at printing using PhantomJS, but that rasterizes the page instead of prints it, so the resulting text in the PDF is not editable. 🚫

tcbyrd commented 7 years ago

I still haven't heard back from Cloudconvert on why it's not rendering a 2 column view, but I went ahead with their implementation. I created this makepdf script that gets invoked as soon as the server starts. This means every time Heroku restarts the dyno it will build a fresh PDF with the most updated data from Airtable and store it in the /public directory so that it's accessible at http://caa-2017-program.herokuapp.com/caa-2017.pdf

I experimented with a real-time conversion, and while this is still possible with Cloudconvert, due to the asynchronous nature of the conversion process, clicking a link to download the most up to date version is not immediate and took anywhere from 2-5 seconds on a good connection. Since the data in Airtable is really only being updated a couple of times a day, I feel like building it when the dyno starts strikes a good balance between staleness and UX.

I haven't created any links in the UI to the generated PDF yet. I'd optimally like to get the 2 column view fixed before doing that since it looks much nicer and matches the printed program.