Try ai library flake tool on RHV data #14
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
| @@ -0,0 +1,1504 @@ | |||
| { | |||
There was a problem hiding this comment.
Do you want to link to the original cockpit code here as a reference, so that it is clear where a lot of the code and comments are coming from?
Reply via ReviewNB
There was a problem hiding this comment.
Added the AI library code reference.
| @@ -0,0 +1,1504 @@ | |||
| { | |||
There was a problem hiding this comment.
note that they are also different in the fact that these log messages are short while some in the real data set can be huge.
Reply via ReviewNB
| @@ -0,0 +1,1504 @@ | |||
| { | |||
There was a problem hiding this comment.
This output might be easier to understand if you reformatted it as a pandas dataframe here.
Also, if these are the input "features" for our model, is it the case that they are all the same besides the log message and the row number? If so, you might want to just note that the current implementation only uses the log feature.
Reply via ReviewNB
There was a problem hiding this comment.
Good Idea, converting it to DF and adding that it's only using log message as a feature.
| @@ -0,0 +1,1504 @@ | |||
| { | |||
There was a problem hiding this comment.
What does this output mean? that we get a non-zero value when comparing 2 different logs based on their cached ncd values?
Reply via ReviewNB
There was a problem hiding this comment.
Oh It's just a demo showing what the calculate function returns. Added that in comments.
| @@ -0,0 +1,1504 @@ | |||
| { | |||
There was a problem hiding this comment.
There was a problem hiding this comment.
The idea was that trying with one type of log may eliminate any noise that would be there because of training with different types of logs. Adding this in markdown. Let me know if that seems illogical or if some other experiment may be required at this point and I'll update them.
|
JH stopped, will push the new changes in comments once it's back. |
And it's back :D |
In this PR, I add a notebook with AI library flake analysis model applied to RHV dataset. I also explain the working with a dummy dataset. The PR addresses issue #12.