Update Darshan integration

[mtitov]

• [x] 1. Use Darshan config file (move env variables into this file if possible)

https://www.mcs.anl.gov/research/projects/darshan/docs/darshan-runtime.html#_configuring_darshan_library_at_runtime example from Shane:

# enable DXT modules if interested in fine-grained
# trace data -- off by default
MOD_ENABLE      DXT_POSIX,DXT_MPIIO

# allocate 4096 file records for POSIX and MPI-IO modules
# (darshan only allocates 1024 per-module by default)
MAX_RECORDS     4096      POSIX,MPI-IO

# the '*' specifier can be used to apply settings for all modules
# in this case, we want all modules to ignore record names
# prefixed with "/lib" and any names with a ".pyc" or ".so" suffix
NAME_EXCLUDE    ^/lib              *
NAME_EXCLUDE    .pyc$,.so$         *

# bump up Darshan's default memory usage to 8 MiB
MODMEM  8

# avoid generating logs for git and ls binaries
APP_EXCLUDE     git,ls

• [x] 2. Change the way Darshan appears for the EnTK executable

follow Andre's comment: https://github.com/radical-cybertools/radical.entk/pull/661#discussion_r1532202265

• [x] 3. Add docstrings