[PYTHON] Procedure to set hydrogen of atom (virtual environment)

Introduction

Do you want to execute code halfway while developing with atom? When I looked it up, there was a package called hydrogen that can use Atom like jupyter, so I will write how to set it. A beginner like me couldn't solve the error easily, and when I could solve it, I cried honestly. Lol I would appreciate it if you could point out any mistakes in this post for the first time and self-taught.

Target person

・ Those who want to set atom hydrogen in a virtual environment

-If you get the following error statement when setting hydrogen with atom (this is an error when you try to execute hydrogen with atom) Traceback (most recent call last): File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.7_3.7.2032.0_x64__qbz5n2kfra8p0\lib\runpy.py", line 193, in _run_module_as_main "main", mod_spec) File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.7_3.7.2032.0_x64__qbz5n2kfra8p0\lib\runpy.py", line 85, in _run_code exec(code, run_globals) File ・ ・ ・ ・ ・ ・ ・ Omitted ・ ・ ・ ・ ・ ・ ・ ・ "C: \ Users \ username \ AppData \ Local \ Packages \ PythonSoftwareFoundation.Python.3.7_qbz5n2kfra8p0 \ LocalCache \ local-packages \ Python37 \ site-packages \ jupyter_core \ paths.py ", line 361, in win32_restrict_file_to_user import win32api ImportError: DLL load failed: �w � 肳 ꂽ���W ��� [����������

About the error I got

When I ran hydrogen on atom after completing all the settings, I was able to get the above error code, and I had a really hard time. As a result of various trials, the solution was as follows. (For details, see step ⑧ below) -This is because the json of the base environment was pasted instead of the json of the virtual environment during the procedure of pasting the json of the jupyter Kernel to the "startup code" and "Kernel Gateways" that can be set from Atom / environment settings / hydrogen / settings. It seems that there was an error. (The base environment is python3.7.)

environment

・ Windows 10 ・ Atom 1.48.0 ・ Hydrogen 2.14.3 ・ Virtual environment conda python3.6.10

• It is assumed that anaconda, atom, and hydrogen are already installed. The reason why I built a virtual environment is that hydrogen does not work well with python3.7? Since there were several articles such as, I decided to build a virtual environment for python3.6. (The truth is unknown because I have not actually tried the setting with python3.7)

procedure

① Build a virtual environment

・ Start anaconda prompt

・ Execute the following code

Anaconda_prompt

>conda create -n py36_fxpractice python=3.6

• The part of py36_fxpractice is the part that becomes the virtual environment name, so you can use any name.
• Since I want to build a virtual environment of python3.6 this time, I specify the version of python when building the virtual environment.

・ After various sentences are displayed after execution, the following sentences are displayed.

Proceed ([y]/n)? 

Enter y and ʻEnter`

②: Confirmation of the constructed environment

Execute the following code

Anaconda_prompt

>conda info -e

Output result

# conda environments:
#
base                  *  C:\Users\UsersName\Anaconda3
py36_fxpractice          C:\Users\UsersName\Anaconda3\envs\py36_fxpractice

• Use this code to check if the virtual environment created earlier is created and where you are.
• From the output result, you can see that there is a base environment and the py36_fxpractice environment created earlier. You can also tell that you are currently in the base environment because there is a*mark in the base environment.

③: Switching to a virtual environment

Execute the following code

Anaconda_prompt

>activate py36_fxpractice

After execution

(py36_fxpractice) C:\Users\Kawahara>

• After execution, the environment will switch to the virtual environment. It will be replaced with (base) ⇒ (py36_fxpractice).

④: Installation of jupyter

Execute the following code

Anaconda_prompt

>conda install jupyter

• Install jupyter in a virtual environment. When you install jupyter, what you need at the same time is displayed like "Do you want to install this together?" After executing the code of ④.

・ After various sentences are displayed after execution, the following sentences are displayed.

Proceed ([y]/n)? 

Enter y and ʻEnter`

⑤: Create a jupyter kernel

Execute the following code

Anaconda_prompt

ipython kernel install --user --name py36_fxpractice --display-name py36_fxpractice

• Since $ ipython kernel install --user --name = environment name --display-name = environment name, the part of environment name can be arbitrary.

⑥: Check the jupyter kernel

Execute the following code

Anaconda_prompt

jupyter kernelspec list

Output result

Available kernels:
  py36_fxpractice    C:\Users\UserName\AppData\Roaming\jupyter\kernels\py36_fxpractice
  python3            C:\Users\UserName\AppData\Roaming\jupyter\kernels\python3

• If something like the path of the created virtual environment is displayed, it is successful.

⑦ Installation of necessary libraries

Anaconda_prompt <example>

conda install numpy　

You can install it with conda install [library name]. ①④ Similarly, you will be asked Proceed ([y] / n)?, So enter y and press ʻEnter. The above installs numpy` as an example.

⑧: Start jupyter notebook

Execute the following code

Anaconda_prompt

jupyter notebook　

After executing the code, jupyte will start in the browser. (The following screen will be displayed.)

-Click New in the above image and select the display-name of the kernel created in ⑤. Try running the code and see if jupyter works fine. After confirmation, close jupyter and anaconda prompt. (I'm closed with a cross in the upper right corner of the screen.)

⑨: Get json of jupyter kernel

Open the anaconda prompt again and execute the following code

Anaconda_prompt

　jupyter kernelspec list --json

• The json code will be displayed after executing the code. The json is surrounded by a lot of [], and the path etc. are displayed inside. If there is a virtual environment name created by the above procedure or a display name in it, it seems to be successful.

⑩: Atom settings

Open Atom from anaconda prompt (execute the following code)

• It seems that hydrogen will not start unless you open it from the anaconda prompt.

Anaconda_prompt

　atom .

After starting Atom, paste the jupyter Kernel json to "startup code" and "Kernel Gateways" that can be set from Atom / Preferences / hydrogen / settings.

Exit atom once and open Atom again in the same way. Open the appropriate file and press shift + Enter and the code should be executed like jupyter on Atom like the screen.

Reference URL

I have referred to the HP of the following people, so I will post it.

Jupyter related

・ [Qiita] Miniconda ・ Python environment construction using Atom Hydrogen

About virtual environment

[Qiita] [For beginners] Try creating a virtual environment with Anaconda

path setting

This is not written in the article, but if the above steps do not work, please refer to this article and try Path in a virtual environment. ・ [Qiita] Atom procedure memo for running Hydrogen