Better Error Handling & Interactive Mode Feature #23
Conversation
moved openai key to a function call for easier calling
allows for accepting changes of individual suggestions
spenceryonce
changed the title
|
I've added better error handling and the ability to use interactive mode. when running the script, you can now use --interactive=True to turn on interactive mode and accept individual suggestions one by one. This increases the likelihood of properly fixing a script without gpt4. |
all commands are logged in a log file with nicely formatted context and datetime
Added Proper Logging
|
Fixed the conflicts |
| print("\nChanges:") | ||
| print_diff(original_file_lines, file_lines) | ||
| except Exception as e: | ||
| raise Exception(f"Failed to apply changes: {str(e)}") |
There was a problem hiding this comment.
what exceptions were you encountering that caused you to add this? Maybe the response wasn't proper JSON (now fixed)?
There was a problem hiding this comment.
Yeah, I was getting non-real JSON in the output responses, so it wasn't parsing if I recall correctly.
|
I like the idea for interactively selecting changes! Another PR has merged that lets users choose to accept entire changes or not. Let me think about how to best combine this. |
|
Circling back around to this, been awhile, let me check out the new changes and update you. |
moved openai key to a function call for easier calling