Automatic mm_analyze.py call from mm_run.py should use the same platform as was used for the run

[ftclark3]

Looping over frames and updating the Context for analysis can be slow on the Reference and CPU Platforms, so if a simulation was run on a GPU, the automatic call to the analysis script should pass the same Platform argument.

Specifically, we should change
os.system(f"{sys.executable} mm_analyze.py {args.protein} -t {os.path.join(toPath, 'movie.dcd')} --subMode {args.subMode} -f {args.forces} {analysis_fasta} {additional_cmd} -c {chain} --output {info_dat_path}")
to
os.system(f"{sys.executable} mm_analyze.py {args.protein} -p simulation_platform -t {os.path.join(toPath, 'movie.dcd')} --subMode {args.subMode} -f {args.forces} {analysis_fasta} {additional_cmd} -c {chain} --output {info_dat_path}")