Move to using TinyDB to store results

[hannahbrucemacdonald]

Capture:

• package
• git-hash
• openmm version
• number of lambda windows
• lambda protocol (i.e. LambdaProtocol in perses)
• transformation (SMIRKS)
• simulation length
• statistical error (per leg)
• correlation times (per leg)

[hannahbrucemacdonald]

Additional default information about the simulation.

Entire conda environment for a particular environment

Clarify from the beginning on the definition of what the statistical error and the correlation times (consistently use pymbar ?)

[hannahbrucemacdonald]

Looking at the difference between variances for a single run, and between runs - look at the comparison of pymbar variance to replicate variance for a subset of simulations.

Also looking at the overlap matrix - a different structure that might be useful to do learning on. Storring all of these things will give different representations.

[hannahbrucemacdonald]

Adding a config file as a data entry

[jmichel80]

Hi @hannahbrucemacdonald did you have a chance to progress with a first version of a tinyDB spec/script to deposit sim data?

[hannahbrucemacdonald]

Sorry for being so slow - working on this today