Netpyne

- Added option to shape conn weights dynamically to create temporal patterns (issue 33)

- Store all params of synMechs exhaustively instead of by reference (issue 139)

- Added netParams.importCellParamsFromNet() to import parameters of multiple cells from existing network (issue 154)

- Added modifySynMechs function

- Added option to record from all synMechs of a type (eg. 'AMPA')

- Fixed bugs and improved efficiency of modify, modifyConns and modifyStims function

- Fixed bug plotting traces

- Improved NeuroML2 import functions (issue 12)

- Added support for homeostatic scaling synapse (requires self NetCon) (issue 141)

- Added method to convert NetParams and SimConfig objects to dicts (issue 141)

- Fixed NetParams and SimConfig constructors so arguments are converted to Dict/ODict (issue 141)

- Added sim.loadBalance() func to calculate and print load balance information (issue 77)

- Fixed bug (introduced during debugging) that prevented random seeds from working properly

- Improved NeuroML import/export functions

- Added option for cache_efficient (optimized load balance when many cores) (issue 77)

- Added sim.clearAll() function to clear all objects in memory and avoid leaks

- Cleared variables after gathering from multiple nodes to avoid memory leak

- Separated common runSim() commands into preRun() (issue 83)

- Fixed bugs in net.modifyStims() method

- Only load neuroml libraries and functions if installed to avoid error when importing sim

- Disable subcellular connectivity since wasn't modifying the synMechs correctly.

- New example using Allen Institute cells showing how to save large net and reload for simulation

- Improved export/import from/to NeuroML2 (issue 12)

- All structures (including Dict and ODict) can now be saved and loaded via json

- Added option to use separate simConfig when loading data from file

- saveData() calls gatherData() if required

- saveData() returns string with full path of saved file (issue 127)

- Fixed bugs so its possible to create and save Python structure independently of NEURON objects

- Fixed bug that created 2 instead of 1 NetStims per cell

- Fixed bug regarding mutable default args in Python funcs

- Fixed bug so NetStim stim params are set to default if missing