Into the wild we go

• Support skesa, spades, velvet and megahit as assembler engines
• Use mash instead of kmc to detect genome size
• Refine the pilon correction parameters
• Attempt to control memory usage
• Less perl dependencies
• One log file