Added option for gradient coloring and alpha in edges

[fransua]

Also a notebook to summarize options

The idea:

    'Ben': 'tab:red',
    'Betty': 'tab:blue',
    'Mike': 'tab:orange',
    'John': 'tab:green',
    'drinks': 'tab:purple',
    'fruit': 'olive',
    'meat': 'tab:brown',
    'bread': 'gold',
    ('John', 'fruit'): 'red'
}

alpha_dict = {
    'Ben': 0.65,
    'Betty': 0.65,
    'Mike': 0.65,
    'John': 0.2,
    'drinks': 0.65,
    'fruit': 0.65,
    'meat': 0.65,
    'bread': 0.65,
    ('John', 'fruit'): 1
}

ax, left_coords, right_coords = sankey(
    left=df['customer'].values[1:], right=df['good'].values[1:],
    rightWeight=weight, leftWeight=weight, aspect=20, fontsize=20, color_gradient=True,
    rightLabels=['bread', 'meat', 'fruit', 'drinks'], colorDict=color_dict, alphaDict=alpha_dict
)

[coveralls]

Pull Request Test Coverage Report for Build 75

• 33 of 56 (58.93%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-11.6%) to 87.692%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pysankey/sankey.py	33	56	58.93%

Totals
Change from base Build 74:	-11.6%
Covered Lines:	171
Relevant Lines:	195

---

💛 - Coveralls

[Pierre-Sassoulas]

Hello, thank you very much for the pull request. There is a lot to check, I'll find the time at some point. Can you rebase on the current master and add tests first so the coverage stays roughly the same?

[Pierre-Sassoulas]

I made the necessary change so the continuous integration work again but I'm going to close as it's impossible to rebase this feature on the main branch at this point. Please feel free to reopen a new merge request if you like, the alpha was interesting.