With the configuration files we can re-architect the module to be a little more "R"-y. I propose the following interface:
Before
py = plotly(username, key)
trace0 = list(x=...y=...)
py$plotly(trace0)
After
plotly:::set_credentials(username, key)
plotly:::plot(trace0)
Or, if you prefer to send your credentials on a per-call basis (e.g. you're using the R demo account):
plotly:::plot(trace0, username=..., key=....)
And similarly for the other functions:
plotly:::ggplotly(..., plotly_options)
plotly:::plot(..., plotly_options)
plotly:::export_image(..., plotly_options)
plotly:::get_figure(..., plotly_options)
With additional functionality ...
figure = plotly:::get_figure(file_owner, file_id, plotly_options) # returns a list of lists of lists of ... (plotly's JSON graph JSON)
data = plotly:::get_data(figure) # doesn't make a request to plotly, it just removes strips keys out of the named list
A few questions I have (because I'm not an R expert)
1 - Is it the R convention to have nested modules/functions?
e.g. in Python we do:
import plotly.plotly as py
# all the functions in py, call the plotly servers
py.plot(...)
py.iplot(...)
py.get_figure(...)
py.image.save_as(...)
# other functions, like tools, are in a separate namespace:
```python
import plotly.tools as tls
tls.embed(figure)
Do we do a similar thing in R? Or do we keep all the functions at the same "level", e.g.
This:
or this:
plotly:::tools:::embed(figure) #??
With the configuration files we can re-architect the module to be a little more "R"-y. I propose the following interface:
Before
After
Or, if you prefer to send your credentials on a per-call basis (e.g. you're using the R demo account):
And similarly for the other functions:
With additional functionality ...
A few questions I have (because I'm not an R expert)
1 - Is it the R convention to have nested modules/functions?
e.g. in Python we do:
Do we do a similar thing in R? Or do we keep all the functions at the same "level", e.g.
This:
or this: