🔧 improve lsf queries to config every job

When using lsf, every job queries lsf for configuration at least 3 times. This release allows to pass TOIL_CONTAINER_LSF_PER_CORE env variable with values Y/N to define if LSF total memory is define per job or per core, thus avoiding any lsf query for getting this info. And it removes the dynamic query to lsf units. It assume always that if an integer is passed to memory= then the units are bytes. If other unit is desired, it can be specified as a string. ie. memory="8Gb".

🔧 freeze toil version at 5.5.0

Freeze toil at 5.5.0 as greater versions no longer support python versions below 3.7.

🍩 Support toil >= 5.5

Support toil 5.5.0 after it brings breaking changes from toil 5.4

Toil 5.x upgrade

🔌 Fix custom_lsf to be compatible with toil >= 5.4.0
🐍. Change CustomLsf to custom_lsf to keep new toil snake_case style for batchSystem options.
🌈 Add coloredlogs to improve logging with ansi output
✅ Fix tests
🐳 upgrade to support docker>=4.3

Thanks to @multimeric for the contributions

🚀 support Singularity 2.6.1

v1.1.6

💎 bump to 1.1.6

🐛 actually fix encapsulated job use for LSF

Patch of 1.1.4 fix

🐛 fix bug that was impeding the use of encapsulated jobs on LSF

thanks @mflevine

🚀add random time delays between retries

v1.1.3

💎bump to 1.1.3

🐛 import subprocess from toil

v1.1.2

📘 minor doc update

🚀 add support for resource retry

CustomLSF now supports retry when jobs are killed by memory or runtime resource usage. Configure with the following environment variables:

option	description
TOIL_CONTAINER_RUNTIME	set a default runtime in minutes
TOIL_CONTAINER_PER_SLOT	not needed with toil > 3.18.0, else set to "Y"
TOIL_CONTAINER_RETRY_MEM	retry memory in integer GB (default "60")
TOIL_CONTAINER_RETRY_RUNTIME	retry runtime in integer minutes (default "40000")
TOIL_CONTAINER_RUNTIME_FLAG	bsub runtime flag (default "-W")