Pysindy

Major changes
This release introduces the `ConstrainedSR3` optimizer. This is an enhanced version of the `SR3` optimizer which allows one to specify *linear equality constraints* on the learned coefficients.

For example, if you know that the equations you are trying to learn look like
text
x' = a + by
y' = c - bx

(i.e. that the coefficient on y should be the negative of the coefficient on x), you could enforce that constraint using `ConstrainedSR3`.

Additionally, `ConstrainedSR3` allows you to specify a different threshold parameter for each library function coefficient. This is an improvement on our other optimizers which all take a fixed regularization constant that is applied uniformly to all coefficients. You can use this feature to choose a higher threshold for, say, quadratic interaction terms like `x*x` or `x*y` than you do for linear terms, for example.

We also have added a [notebook](https://github.com/dynamicslab/pysindy/blob/master/examples/7_plasma_example.ipynb) demonstrating these new features on a real world plasma dataset.

Other Updates
* Fix bug with extrapolating control inputs outside of original time domain with `SINDy.simulate`
* It is now possible to specify an initial guess for coefficients with the `STLSQ` and `SR3` optimizers
* Added some fancy plots to the differentiation notebook
* Minor bug fixes

This release fixes a minor issue causing `SINDy.simulate` to fail when vectors of control inputs are passed in (see 94).

The main update in this release is the addition of two objects meant to conform to the API for [Sckit-time](https://scikit-time.github.io/): `SINDyEstimator` and `SINDyModel`. We also added a notebook showing how these two objects may be used.

We also made some other minor changes:

* Add support for vector arguments for control inputs to `SINDy.simulate` (previously control inputs had to be callable)
* Removed extraneous `n_jobs` parameter from the `SINDy` class
* Added tests for SR3 trimming options
* Updated some docstrings
* Restructure examples/README content
* Minor formatting updates for the new version of black

This release implements the `SINDyDerivative` class, which allows one to use differentiation methods from the [derivative](https://derivative.readthedocs.io/en/latest/) package. This will enable the application of SINDy to much noisier datasets.

Note that `PySINDy` now _requires_ the `derivative` package.

`derivative` includes the following numerical differentiation techniques:
* Spectral derivatives (via the FFT)
* Spline-based derivatives
* Finite differences of arbitrary order
* Polynomial least-squares (Savitzky-Golay)
* Total [variation regularized derivative](https://www.hindawi.com/journals/isrn/2011/164564/) (the method recommended in the original SINDy paper)

This release also contains some improvements to the documentation:
* Example using `SINDyDerivative` in the Feature Overview notebook
* A [new notebook](https://pysindy.readthedocs.io/en/latest/examples/5_differentation.html) comparing all the differentiation options available in PySINDy
* An example showing how to optimize parameters of `derivative` objects with cross-validation

This release enables much easier cross-validation of `SINDy` models using Scikit-learn cross-validation tools. Major changes include:
- `t_default` parameter added to the `SINDy` class. This parameter specifies the default time step that should be used whenever the `t` argument of a `SINDy` method is not used.
- New notebook demonstrating cross-validation and other examples combining Scikit-learn and PySINDy objects.

More details can be found in 84.

PySINDy now requires Scikit-learn version 0.23 instead of 0.21. There are also some minor bug fixes included in this release related to checking when optimizers have been fit.