Some minor tweaks for the analysis codes

[wehs7661]

This PR includes the following work:

• In the original code, free energy calculations are deactivated whenever the alchemical sampling is poor (based on the criterion of whether all trajectories have at least one round trip). However, this should not be necessary as estimated free energies could still be important in such cases, so we deactivated this mechanism.
• We modified cluster_traj to allow the clustering analysis to be performed for the entire trajectory instead of only for fully coupled configurations. We also modified the function so that when the alchemical sampling is poor, clustering analysis is deactivated and the code can normally exit (instead of erroring out due to a GROMACS error, which can be less clear).
• We modified analyze_REXEE.py to additionally pickle t_idx_list and g_list after preprocess_data is successfully executed. This allows free energy calculations to be carried out given a pickle file even when subsampling_avg=True, essentially addressing issue #35 .

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 52.95%. Comparing base (bce555f) to head (02136d6).

:exclamation: Current head 02136d6 differs from pull request most recent head 2a51ffc. Consider uploading reports for the commit 2a51ffc to get more accurate results

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #36 +/- ## ======================================= Coverage 52.95% 52.95% ======================================= Files 7 7 Lines 1571 1571 ======================================= Hits 832 832 Misses 739 739 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.