[Altair] Sync exported histogram code with new bin determination

[micahtyong]

In this PR

Closes #302 by syncing up the new histogram bin determination rules from #285 into the Altair exported code.

Changes

• Modifications in altair/Histogram.py and test_vis.py

Example Output

Employee Dataset: Distribution

Following the example in #302, upon running,

df = pd.read_csv("https://raw.githubusercontent.com/lux-org/lux-datasets/master/data/employee.csv")
series = df["YearsAtCompany"]/df["TotalWorkingYears"]
series

then exporting our graph,

print(series.exported[0].to_code("altair"))

we now get the following output:

Olympic Dataset: Occurrence

The output is also as expected for displaying categorical distributions using the y-axis. For example, running

df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=True")
df

then selecting the second graph from the Occurrence tab and exporting with

print(df.exported[0].to_code("altair"))

we now get the following output:

[codecov]

Codecov Report

Merging #304 (c6a00e0) into master (14c141b) will increase coverage by 3.50%.
The diff coverage is 90.04%.

@@            Coverage Diff             @@
##           master     #304      +/-   ##
==========================================
+ Coverage   77.58%   81.08%   +3.50%     
==========================================
  Files          40       50      +10     
  Lines        2846     3569     +723     
==========================================
+ Hits         2208     2894     +686     
- Misses        638      675      +37     

Impacted Files	Coverage Δ
lux/action/column_group.py	92.59% <ø> (-3.71%)	⬇️
lux/action/enhance.py	100.00% <ø> (ø)
lux/action/filter.py	91.66% <ø> (ø)
lux/action/generalize.py	80.95% <ø> (ø)
lux/action/row_group.py	95.45% <ø> (ø)
lux/vislib/matplotlib/MatplotlibChart.py	79.31% <79.31%> (ø)
lux/core/frame.py	70.16% <80.00%> (+0.14%)	⬆️
lux/interestingness/interestingness.py	86.48% <80.00%> (-0.22%)	⬇️
lux/vislib/matplotlib/ScatterChart.py	80.30% <80.30%> (ø)
lux/vis/Vis.py	74.85% <81.25%> (+1.34%)	⬆️
... and 38 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 14c141b...86eaac2. Read the comment docs.

[dorisjlee]

Thanks @micahtyong !