copy_posters.py is a quickly hacked together script which reads the Overview.txt file (a simple tab-delim table of Easychair abstracts mapped to talk order or poster assignment), reads the appropriate original PDF from the Easychair PDF dump, then copies to a new folder modifying the filename to the poster ID and adding a simple text box to the PDF in the upper left corner. Note there are two specific hacks to deal with the odd poster with non-standard margins or page size.
The resulting individual PDFs were combined into one using Adobe Acrobat, though one could easily use PyPDF2 to do the same.
Talk abstracts #'s were based on both day and order. Abstracts were sorted and file names modified to talk order using a simple script copy_talks.pl (yes, Perl), and then combined as above using Adobe Acrobat.
Acrobat was initially used for compression as well, but this method was not used in favor of Nomi's method as it retained the embedded images at higher resolution with better compression.
The final file is BOSC2017-complete-program-compressed.pdf.