Tutorial set-up needs corrections about dataframes HOT 7 CLOSED

Thank you! As commented I have updated the notebook explaining the issue (I just referred the user to the documentation of Matplotlib in case of errors as there are too many possible problematic scenarios to cover).

from dataframes.

I'm just running through this tutorial again. The updated set-up notes are good for me. Using PyCall delegates the configuration responsibility to the Julia package developers, where it is safer. the MIT Computational Thinking course seems to take this strategy too.

from dataframes.

Indeed - I think that in the future we should switch from PyPlot.jl to some other plotting backend. For the time being do you think it will be enough to add a link to https://github.com/JuliaPy/PyPlot.jl#installation where the instructions how to set up PyPlot properly are provided?

from dataframes.

I have added some more instructions to the notebooks. Can you please have a look if they look clear for you.

from dataframes.

@bkamins
Thanks very much for your attention to this! I've been slow getting back to you, because I've been learning to configure too many things...

isfile.(["Project.toml", "Manifest.toml"]) behaved correctly.

the ] status and ] instantiate cells do not print any output, but if I activate the project from a repl I get:

 pkg> activate .
  Activating environment at `~/Tutorials/Julia-Tutorials/JuliaAcademy/DataFrames/Project.toml`

(DataFrames) pkg> st
      Status `~/Tutorials/Julia-Tutorials/JuliaAcademy/DataFrames/Project.toml`
  [69666777] Arrow v1.6.0
  [336ed68f] CSV v0.8.5
  [a93c6f00] DataFrames v1.2.0
  [da1fdf0e] FreqTables v0.4.4
  [38e38edf] GLM v1.5.1
  [b98c9c47] Pipe v1.3.0
  [d330b81b] PyPlot v2.9.0
  [1986cc42] Unitful v1.9.0

Which looks like the correct result, and was quite clear to me.

I tried the notebook again with the python matplotlib python package uninstalled from my enclosing conda environment, and it all instantiated correctly . But the first exercise could not find it.

from dataframes.

I think I have a simple solution which gets julia to install everything including Jupyter notebook and MatPlotLib automatically using its internal conda environment.

The user should just have to start the julia repl, activate the tutorial environment, and run notebook().

You may know how to set the julia project .toml to do most of this.

in the repl:

>
ENV["PYTHON"]=""
using Pkg
Pkg.activate()
Pkg.add("PyCall")
Pkg.build("PyCall")
Pkg.add("IJulia")
using IJulia
notebook(;dir=pwd(), detached=true)

References

1. The PyPlot module for Julia
2. Public API - IJulia

from dataframes.

I'm using this now - works for me without any environment other than julia - no need to install Jupyter notebook, MatPlotLib etc...

Project.toml looks like this

[deps]
Arrow = "69666777-d1a9-59fb-9406-91d4454c9d45"
CSV = "336ed68f-0bac-5ca0-87d4-7b16caf5d00b"
DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
FreqTables = "da1fdf0e-e0ff-5433-a45f-9bb5ff651cb1"
GLM = "38e38edf-8417-5370-95a0-9cbb8c7f171a"
IJulia = "7073ff75-c697-5162-941a-fcdaad2a7d2a"
Pipe = "b98c9c47-44ae-5843-9183-064241ee97a0"
PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"
PyPlot = "d330b81b-6aea-500a-939a-2ce795aea3ee"
Unitful = "1986cc42-f94f-5a68-af5c-568840ba703d"

Use this command line in the Tutorial directory:

PYTHON="" julia -e 'using Pkg; Pkg.activate("."); Pkg.instantiate(); using IJulia; notebook(;dir=pwd(), detached=true) ' 

PYTHON="" julia... sets the python environment to empty before running julia.

I hope that is of some use...

from dataframes.