v2.2

New features:

• thermo: added TRAM estimation.
• thermo: added plotting feature for implied timescales.
• thermo: added Jupyter notebook examples: :ref:ref-notebooks.
• thermo: show convergence progress during estimation.

Fixes:

• clustering: fix parallel cluster assignment with minRMSD metric.
• base: during estimation the model was accessed in an inappropriate way,
  which led to the crash "AttributeError: object has no attribute '_model'" #764.
• coordinates.io: fixed a bug when trying to pyemma.coordinates.load certain MD formats.
  The iterator could have returned None in some cases #790.
• coordiantes.save_traj(s): use new backend introduced in 2.1, speed up for non random
  accessible trajectory formats like XTC. Avoids reading trajectory info for files not
  being indexed by the input mapping. Fixes #788.

v2.1.1

Fixes

• some issues in parallel environments (eg. shared file systems on clusters)
  in conjunction with enabled trajectory cache [default]).
• File-logging disabled by default
• No files are written to your home directory until you request it (eg. by setting the
  config directory [pyemma.config.cfg_dir = "path" or invoke pyemma.config.save(filename)]

Improvements

• Parallel assignment of cluster centers via threads.

v2.1

[setup] enforce higher joblib version. Fixes a bug with multiple jobs.

Fix type check in thermo.util.umbrella check

Merge pull request #706 from Upabjojr/long_python3

fix type check in thermo.util._ensure_umbrella_center

latest fixes

• fix database corruption on OSX, safari in notebook sessions during kernel restart
• added FAQ about bitness on Windows
• fixed callbacks in progress reporter
• fixed some minor bugs

v2.1rc1

First release candidate of 2.1 containing major new features like:

• multi-ensemble Markov models
• improved HMM code
• Re-designed coordinates package