search

JuliaText / TextAnalysis.jl

Julia package for text analysis
Other
373 stars 95 forks source link

Added ROUGE Score to TextAnalysis.jl #156

Closed djokester closed 5 years ago

djokester commented 5 years ago

Added the ROUGE Score functions originally implemented in the paper ROUGE: A Package for Automatic Evaluation of Summaries, Chin-Yew Lin

Implemented the following functions in rouge.jl ROUGE-N ROUGE_L (Summary Level) ROUGE_L (Sentence Level)

Implemented the following functions in utils.jl Weighted LCS Jackknifing F Measure Based on Weighted LCS.

Added tests for ROUGE.jl

djokester commented 5 years ago

@Ayushk4 Is there something like pytest for Julia which lets me run tests locally without creating this mess all over the commits?

Ayushk4 commented 5 years ago

Once a package is under development with dev, you can simply use pkg> test <package-name> for testing a package. Please go through the docs of Pkg.jl for details.

To clean, what you are calling as mess over commits, you may squash the commits and force push in git, please look this up as well.

Happy Learning :)

djokester commented 5 years ago

@Ayushk4

Once a package is under development with dev, you can simply use pkg> test <package-name> for testing a package. Please go through the docs of Pkg.jl for details.

This is what I needed thanks for this.

To clean, what you are calling as mess over commits, you may squash the commits and force push in git, please look this up as well.

This is unwanted advice, no one asked you for. You are simply showing off. What I do with my commits is none of your business.

Happy Learning :)

This is to suggest that you know more than the person sending the PR and that you love goading

I will give you a piece of advice. Don't do this. Just answer what is asked. Don't do this especially when the other person is a senior from your college. It is rude and puts the other person in a very awkward position as to how to respond to you.

I quietly saw all of the changes you suggested and here are my 2 cents for you. Don't showboat. Let the person take his time to commit changes and make the tests pass. Don't bombard it with changes the minute a PR drops.

Also, remember what goes around, comes around.

aviks commented 5 years ago

Hi @djokester Thanks for doing this, really appreciate the effort you've put into this. It's a large PR, so I'm taking a while to review.

Re: Ayush's comment, I think it's easy to misunderstand each other on the internet. So I always go by the maxim of "always assume good intentions". Even if sometimes it feels someone is talking down to you, they are probably only trying to help. Don't assume malice at the beginning.

As an open source maintainer, doing code reviews are hard. We are are necessarily forced to be negative to someone who has just put in a large amount of free work -- it's psychologically draining. So I do appreciate it when someone else reviews the code, and helps spread that load. And you don't have to always agree with a review comment, mine, or anyone else.

But mutual reviews help make the code better over time. I would be very concerned if I put up a large PR and no-one reviewed it -- I would not be confident of the quality without that. Even if I don't agree with things a reviewer says.

But once again, thank you for writing this, this will be really useful for the project.

djokester commented 5 years ago

@aviks made the changes Kindly have a look