Reactionmechanismgenerator

- IPython notebooks now found in the `ipython` folder and contains useful illustrative examples. Requires the installation of jupyter via the command `conda install jupyter`
- Cantera support in RMG:
- New module found in `rmgpy.tools.canteraModel`, which provides functions to help simulate RMG models using Cantera instead of native RMG solvers.
- Has capability to generate cantera conditions and convert CHEMKIN files to cantera models, or use RMG to directly convert species and reactions objects to Cantera objects.
- Demonstrative example found in `ipython/canteraSimulation.ipynb`
- Module for regression testing of models generated by RMG
- This new module is found at `rmgpy.tools.observableRegression` and helps identify differences between different versions of models generated by RMG, using the "observables" that the user cares about (i.e. main species mole fractions). Runs simulations via cantera.
- Automatic plotting of simulations and sensitivities when generating models
- Automatic plotting of simulations in the job's `solver` folder when `saveSimulationProfiles` is set to `True` in the input file.
- Sensitivities for top 10 most sensitivie reactions and thermo now plotted automatically and stored in the `solver` folder.
- New `rmgpy.tools.plot` module contains plotting classes useful for plotting simulations, sensitivities, and other data
- CHEMKIN files are now read literally: unlisted bath gases are not added when loading a chemkin file
- Functionality to assist in saving training reactions and thermo groups to RMG (see ipython notebook scripts in `RMG-database/scripts/` for working examples)
- Improved thermochemistry estimation (mostly for cyclics and polycyclics)
- Add rank as an additional attribute in thermo database entries to determine trustworthiness
- Fixes ring group retrieval when multiple rings or polycyclic rings are found in a single molecule
- Changes to resonance isomer generation to fixes aromatic crashes and print more debugging information
- Bug fixes:
- Crashes with a useful error message when a Reaction Library and Seed Mechanism contains more than 3 reactants or products
- Initial mole fractions get handled as floats to prevent integer division errors, and a useful message about renormalization is printed
- Training reactions now load successfully regardless of `generateSpeciesConstraints` parameters (602)
- Transport data is now saved correctly to CHEMKIN `tran.dat` file and also imports successfully
- Fixes appending of reactions to CHEMKIN file when reaction libraries are desired to be appended to output
- Fixes writing of csv files for simulation and sensitivity results in Windows
- Reverts the vertex connectivity value algorithm to prevent overflow errors
- Fixes `Reaction.draw()` function to draw the entire reaction rather than a single species
- Other fixes
- Enforces compatible versions of matplotlib and cantera in anaconda build
- Github pull requests from forked repo's will no longer fail Travis CI builds due to encrypted environment variable issues

- Pdep convergence issues in RMG-Py v1.0.2 are now fixed.
- RMG-database version information and anaconda binary version information is now recorded in the RMG log file.

- Windows users can rejoice: RMG is now available in binary format on the Anaconda platform. Building by source is also much easier now through the Anaconda managed python environment for dependencies. See the updated [Installation Page](http://reactionmechanismgenerator.github.io/RMG-Py/users/rmg/installation/index.html#installation) for more details
- Reaction filtering for speeding up model generation has now been added. It has been shown to speed up model convergence by 7-10x. See more details about how to use it in your RMG job [here](http://reactionmechanismgenerator.github.io/RMG-Py/users/rmg/input.html#filterreactions). Learn more about the theory and algorithm on the [Rate-based Model Enlarging Algorithm](http://reactionmechanismgenerator.github.io/RMG-Py/theory/rmg/enlarger.html) page.
- The RMG [native scripts](http://reactionmechanismgenerator.github.io/RMG-Py/users/rmg/modules/index.html) are now organized under the `rmgpy.tools` submodule for developer ease and better extensibility in external scripts.
- InChI conversion is now more robust for singlets and triplets, and augmented InChIs and InChI keys are now possible with new radical electron, lone pair, and multiplicity flags.
- Output HTML for visualizing models are now cleaned up and also more functional, including features to display thermo comments, display enthalpy, entropy, and free energy of reaction, as well as filter reactions by species. You can use this new visualization format either by running a job in RMG v1.0.2 or revisualizing your CHEMKIN file and species dictionary using the [visualization web tool](http://rmg.mit.edu/simulate/chemkin)