Jupyter support

[sebhoerl]

It would be nice to have functions to run things with synpp from Jupyter, for instance stage by stage and to examine the results directly, e.g.

pipeline = synpp.setup("config_ile_de_france.yml")
df_census = pipeline.run("data.census.cleaned")
# Do something with df_census

df_census = pipeline.run("data.census.cleaned", stages = {
  "data.census.raw": SOME_DATA # Override stage output
})

[davibicudo]

With stage decoration maybe a workaround could be:

df_census = synpp.run([synpp.stage("data.census.cleaned", **{"data.census.raw": synpp.stage(lambda: SOME_DATA)})]) 

The dependency "data.census.raw" is replaced by another fake stage produced with the lambda.

I haven't tested it though, it would probably break the cache...