abhi3700/semantic-hashing-demo

Convert analysis to notebook

Closed this issue · 5 comments

jfrank-summit commented

Currently the analysis isn't very straight forward to conduct. I'm not sure what I need to actually run to get to the final output. It is also not easily repeatable. I think two improvements should be made:

  • Improved documentation
  • Add python notebook to allow for data exploration and saving of results
abhi3700 commented

Actually, the README currently covers all the steps to run the 4 scripts as used so far.

Still, I will convert all into a jupyter notebook.

jfrank-summit commented

The below is not sufficient documentation. I don't think they are in the correct order and they don't actually explain what is done. Also, in my environment huak run ... always tells me that ... does not exist. Is there a step missing in the documentation?

# pre-process input data to buckets
huak run preprocess

# detect similar text
huak run detect

# generate texts (source, variant)
huak run generate

# process generated texts (source, variant)
huak run post_generate
abhi3700 commented

Actually, you needed to ensure this first.
It would be helpful for me to diagnose the issue, if you could help me with the error following 🔝 .

The sequence is accordingly set in README.

jfrank-summit commented

I do not think the readme documentation is adequate. I should be able to clone the repo and know exactly what I need to do to run the analysis, including what each step does, which are optional, etc. I should be able to do this with somewhat limited context.

  • 👀1
abhi3700 commented

I do not think the readme documentation is adequate. I should be able to clone the repo and know exactly what I need to do to run the analysis, including what each step does, which are optional, etc. I should be able to do this with somewhat limited context.

Added in commit: 6d2fd28