Fixing audio file version handling

There was an issue with the way load_audio_chunks_as_json_objects and sample2data identified the version of the audio file, mistaking it as V1.0. This is now fixed, but might create issues with some older code.

Added code to support new proximity files format

In the new version of the system, the server sets the badge_id on the badges, so it is no longer calculated based on the MAC address. Therefore, we can directly translate the observed id to a unique member_key using a translation table or using the member_id field now saved together with the data.

Added functions for smoothing the RSSI values, and filling gaps

v0.4.0

Bump version: 0.3.6 → 0.4.0

Minor changes to function names

v0.3.6

Bump version: 0.3.5 → 0.3.6

Adding more functions for parsing hub logs and reading sample counts for proximity and audio

v0.3.5

Bump version: 0.3.4 → 0.3.5

Adding a module for parsing hub logs

v0.3.4

Bump version: 0.3.3 → 0.3.4

Added proximity pipeline. Code cleanup

Add a pipeline for handling proximity data

Add a pipeline for handling proximity data

Updated examples and additions to proximity analysis

Merge pull request #2 from HumanDynamics/oren

Updates to notebooks used for battery testing

0.1: Merge pull request #2 from HumanDynamics/oren

Updates to notebooks used for battery testing