I've made the following changes to the TDMS reader class:

• Units are now read from the TDMS channel properties (unit_string) rather than hardcoded. Current is converted from A to mA to follow ixdat convention.
• Phase_E unit is now read from the file (rad, radians) instead of None.
• ISO 8601 datetime is included in the measurement metadata.
• Macro file parsing: if a *.EC_Macro file is present in the same directory as the .tdms file, it is automatically parsed and included in the metadata as a list of experiment steps (hardware control, gas, CV, hold parameters). The user is warned if no macro file is found, or if multiple are found.
• Minor fixes: removed unused import, replaced print() with warnings.warn(), fixed a scope bug where tseries_EIS could be undefined when Phase_E was present without Z_E.

The following points raised in the review are left for a future PR as they concern the broader ixdat API:

• Renaming tstamp to unix_time in exports
• Supporting multiple units in ECMeasurement.calibrate (e.g. A_el_m2)
• Making the default potential symbol configurable
• Replacing "anodic"/"cathodic" with "positive"/"negative" in docstrings

CI tests for python 3.7 are currently failing due to unrelated issues, which are being addressed separately in PR #194

Everything else look great!
Resolving the conflict with NEXT_CHANGES.rst and then merging 🎉