I've been working with Prodigy on my Mac without any problems. I'm now working on a PC and am having problems getting it working. I have Prodigy successfully installed in its own environment.
In the documentation, I found that I need to run Python -m to get Prodigy to work. Here is the problem. When I run the command, nothing happens. If I stop the process and then run it again, I get the following error message:
I went into the json config file and changed the port from 8080 to 8000 (as shown here), and the problem persists. Prodigy does open on the local server but is certainly problematic as it doesn't load in the details of the dataset:
Typically the main reason you're getting this error is because you still have running another instance of Prodigy running but don't realize it (this is why you're getting a port collision).
The easiest way to resolve is to look at what's running on that port and try to shut it down.
Can you try this? I'd start first with the 8080 as my assumption is you may have a Prodigy process still running on that port and may need to kill it. Alternatively, you could restart your computer and try again. I'm a bit more hesitant on 8000 as I'm a bit worried this may be non-Prodigy/Jupyter related.
I don't understand what you mean that "it doesn't load in the details of the dataset:". Can you explain more?
Was this from opening localhost:8080 or localhost:8000?
Are you getting these same issues if you run Prodigy from a terminal, not a Jupyter notebook?
In the past, I know some users have experienced issues that are actually Jupyter related, not Prodigy. While Jupyter may make be easier for some things, we've noticed a lot of users realizing Jupyter introduces a lot of new problems so Prodigy is typically better to use on the CLI, especially if you're having issues like this.
Did you try any other port? Not sure if you have some other non-Prodigy/Jupyter related process running on 8000 too.
Fyi, you can normally specify config overrides like PRODIGY_PORT=1234 python -m prodigy ....
Can you try a few other ports to see if it works?
This can be a bit more complex when using Jupyter too:
Thanks for the quick response. I'm using Visual Studio. I changed the port back to 8080 and implemented your solution. On a Mac, I can run it through VS but not on a PC -- but, it works just fine in the terminal following the procedures you recommended.