Quasar985 commented 4 months ago

Prerequisites

[x] Reviewed the checklist
[ ] Reviewed feedback from the "Sonar Cloud" bot. Note that you have to wait for the "CI / Unit Tests") to complete first. Failed Unit tests can be debugged by adding the label "verbose logging" to the GitHub PR.

Description of the Change

Changes include:

Minor code edits to python script, such as a fix missing import and a fix for deprecated code that was missed in previous PR. Rebuilt python package.
Functionality to copy the constellation_client.py script into the jupyter notebook's directory, should the installation of the package fail.
The REST directory path in the Constellation options menu is now functional. Now, if the REST directory option is blank (default), the rest.json secret is copied to the .iPython directory. Otherwise, rest.json is copied to the path specified in REST directory.
A small warning is shown if the REST directory option is changed, alerting the user that constellation_client will likely not be able to find rest.json.
Added functionality to constellation_client to update it on the path of rest.json. Path can either be to the directory of rest.json or the exact path to rest.json. Documentation is updated to inform users of this. Documentation in training modules will also be updated.

Alternate Designs

N/A

Why Should This Be In Core?

N/A

Benefits

Fix bug stopping package from working correctly and added functionality to reduce potential frustration.

Possible Drawbacks

None

Verification Process

FOR DEPRECATED CODE

Open Consty
Open a graph, or create a new graph
Start Jupyter Notebook

Run the following code:


import io
import os
import pandas as pd
import PIL.Image, PIL.ImageDraw, PIL.ImageFilter, PIL.ImageFont

Also import some of the notebook display methods so we can display nice things.

# from IPython.display import display, HTML, Image

This is a convenient Python interface to the REST API.

# import constellation_client

cc = constellation_client.Constellation()

df = cc.get_dataframe() df.head()

5. Verify no errors are thrown

FOR CHANGING REST PATH
1. Open Consty
2. Open a graph, or create a new graph
3. Click Setup->Options->CONSTELLATION
4. In "REST" directory (under Internal Webserver) type in a valid file path (e.g C:\Projects)
5. Observe warning message appear below REST directory input box
6. Click Apply, then OK
7. Click Tools->Start Jupyter Notebook
8. Run the following code and MAKE SURE TO CHANGE C:\Projects TO THE PATH SPECIFIED IN STEP 4:

import io import os import pandas as pd import PIL.Image, PIL.ImageDraw, PIL.ImageFilter, PIL.ImageFont

Also import some of the notebook display methods so we can display nice things.

# from IPython.display import display, HTML, Image

This is a convenient Python interface to the REST API.

# import constellation_client

cc = constellation_client.Constellation()

Tell constellation_client where rest.json is.

CHANGE C:\Projects TO THE PATH SPECIFIED IN STEP 4 ON GITHUB

cc.update_rest(r"C:\Projects")

df = cc.get_dataframe() df.head()


9. Verify no errors are thrown, and data from graph is shown (assuming current graph has data)

### Applicable Issues

<!-- Link any applicable issues here -->
https://github.com/constellation-app/constellation/issues/2073

OrionsGuardian commented 4 months ago

In the Constellation Options screen, there's a setting where you can specify the REST directory:

This setting seems to be ignored. Starting the REST server when this setting is populated didn't put the rest.json file in the specified directory. Instead, the rest.json file was put into two folders: .CONSTELLATION and .ipython

Why are we writing two copies of the file? One copy should be sufficient.

Also, we should try to find a way to use the user specified directory to hold the rest.json file. There is a possibility that some users are on workstations that have limited permissions for writing files, so they will need the ability to specify which folder to use.