Create URL list for Google Cloud Storage Transfer Service, see: https://cloud.google.com/storage/transfer/create-url-list

make_TsvHttpData.pl

#!/usr/bin/perl
use strict;

my $http_base = shift;
my $path_base = shift;
my $prev_file = shift;

if ( ! $http_base || ! $path_base ) {
  print STDERR <<"HERE";
USAGE:
$0 <HTTP base URL> <Filesystem base path> [<Previous output file>]

EXAMPLE:
$0 http://my.hostname.org/ ~/public_html/sync/ ~/TsvHttpData.pre

SYNOPSIS:
This script is used to generate a 'TsvHttpData' URL list for Google Cloud
Storage Transfer Service, see:

  https://cloud.google.com/storage/transfer/create-url-list

TsvHttpData file is tab-delimited and contains 3 columns:

* URL
* Object size, in bytes
* Base64 encoded MD5 checksum of object

MD5 checksumming is I/O intense, and to improve efficience for repeated runs
of this script, we take advantage of each file's modification time to determine
if the previously calculated MD5 checksum can be reused, or needs to be
recalculated.  As such, output format contains 4 columns:

* URL
* Object size, in bytes
* Base64 encoded MD5 checksum of object
* Object modification time, in seconds since Unix epoch

TsvHttpData format can be created from this script's output like:

  cat ~/TsvHttpData.pre | cut -f 1,2,3 > ~/public_html/TsvHttpData.tsv
HERE

  exit(1);
}
if ( ! -d $path_base ) {
  die "Not a directory: $path_base";
}
if ( ! -f $prev_file ) {
  print STDERR "no previous file";
}

#recover previously cached object modification times
#and MD5 checksums
my %ent = ();
open( P, $prev_file );
<P>; #skip header
while ( my $line = <P> ) {
  chomp $line;
  my ( $url, $size, $md5, $mod ) = split /\t/, $line;
  my $path = $url;
  $path =~ s/$http_base/$path_base/;

  $ent{ $path } = [ $url, $size, $md5, $mod ];
}

print qq(TsvHttpData-1.0\n);

#iterate over files currently in path to sync.
#reuse cached data if it exists and isn't stale.
foreach my $path ( `find $path_base -type f` ) {
  chomp $path;

  #modification time
  my $m = `stat -c %Y "$path"`;
  chomp $m;

  if ( $ent{ $path } && $m == $ent{ $path }[3] ) {
    print join "\t", $ent{ $path }[0], $ent{ $path }[1], $ent{ $path }[2], $ent{ $path }[3];
    print "\n";
    next;
  }

  my $o = $path;
  $o =~ s/$path_base//;
  $o = $http_base . $o;

  #TODO make escaping more robust.
  $o =~ s/%/%25/g;

  #size in bytes
  my $s = -s $path;

  #base64 encoded md5 checksum
  my $h = `openssl md5 -binary "$path" | openssl enc -base64`;
  chomp $h;

  print "$o\t$s\t$h\t$m\n";
}

Hey! thanks for that! :)
Just a heads up. My gc file transfer failed when I tried to upload the file generated directly by the output.
I had to erase the last bit of the file, after the md5 hash to make work. I don't know exactly what it stands for though.
I'm transfer files from an ubuntu machine, 17.10

can you share an example how you used this script?
Thanks