Jit compile simple Pandas dataframe to candas

For the supported subset of Pandas supported we would like to make a jit decorator in Python that uses the type information provided in the function signature, and the pandas code included in the function body, to generate candas code as a callable extension module. We should at least suppert to function signatures:

1. Filename of csv to read
2. numpy arrays for the data to fill into the candas dataframe

Pseudo code
``` python
@candas.jit(int, float, float) # Type information needed
def my_analysis(filename: str) -> float:
df = pandas.read_csv(filename)
return df.sum(axis=1)
```