Anypytools

**Changed:**
- Update documentation to use markdown.
- Add error message when no files are found with
`search_subdir` arguments (fixes [69](https://github.com/AnyBody-Research-Group/AnyPyTools/issues/69)).

**Fixed:**

- Fix deprecation warnings in pytest plugin.

**Added:**

- AnyPyTools can now be used on linux if the AnyBody Modeling System is installed using wine.

**Changed**:

- The `to_dataframe()` methods have been updated.
They now return a dataframe without an index by default.

They also now support interpolation of the data.

python
app = AnyPyProcess()
results = app.start_marco(macro_list)

df = results.to_dataframe(
interp_var="Main.MyStudy.Output.Abscissa.t",
interp_val=linspace(0,1,50)
)


**Added**:

- Documentation on how to use the `to_dataframe()` method has been added to the tutorials.

Add methods for exporing simuation output as a pandas dataframe.

python
app = AnyPyProcess()
results = app.start_marco(macro_list)

df = results.to_dataframe(index_var="Main.MyStudy.Output.Abscissa.t")


The data has too be consistent across all macros. That means that the
`index_var` must be present in all simulations. Also note that since
pandas dataframes are 2D any 3D data will be flattened. For example 3D vectors
which will be saved into three columns in the dataframe. E.g. `r[0]`, `r[1]`, `r[2]`.

Ensure that 'nan' values returned from AnyBody are treated as "float('nan')" when returned to Python.