Adding Paragram Embedding

[tejasvaidhyadev]

I am adding Paragram Embedding and Google-drive download fetch method for datadeps in Paragram.I think it will be great to have default G-drive download fetch method in Datadeps

[oxinabox]

I am adding Paragram Embedding and Google-drive download fetch method for datadeps in Paragram.

I think it will be great to have default G-drive download fetch method in Datadeps

I made a prototype of this a few years back. It would be nice to see something more polished.
https://github.com/oxinabox/PyDrive.jl
idk if in DataDeps or in a GoogleDrive.jl package that supports DataDeps (potetially similar to how AWSS3.jl supports DataDeps)
I guess its not much code so it could go in DataDeps.
But being able to add Auth (like the PyDrive.jl demo does) the code code grows so might be better to have it else-where

I agree that it doesn't belong here.
It would be kind if nice to hold on this PR til that code found a better home.
But if you need these for a project we can look at keeping it here for now.

[tejasvaidhyadev]

For now, I don't need it for any project but would love to find a home to code.
Any suggestion how to proceed ?

[oxinabox]

I suggest you create a GoogleDrive.jl package,
based on the code you have.
(use PkgTemplates if you are not already).

Register that package, then we will add it as a dependency of Embeddings.jl

[tejasvaidhyadev]

Hi @oxinabox I created GoogleDrive.jl Package , and also opened a Registration PR
After the PR is merged I will push all the above suggested changes to this PR.

[tejasvaidhyadev]

@oxinabox as suggested GoogleDrive.jl as dependency and common.jl is added

[codecov-io]

Codecov Report

Merging #30 into master will decrease coverage by 0.81%.
The diff coverage is 88.46%.

@@            Coverage Diff             @@
##           master      #30      +/-   ##
==========================================
- Coverage   96.77%   95.95%   -0.82%     
==========================================
  Files           4        6       +2     
  Lines          93       99       +6     
==========================================
+ Hits           90       95       +5     
- Misses          3        4       +1

Impacted Files	Coverage Δ
src/Embeddings.jl	100% <100%> (ø)	⬆️
src/glove.jl	100% <100%> (+8%)	⬆️
src/Paragram.jl	80% <80%> (ø)
src/common.jl	89.47% <89.47%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 65390f1...ac45848. Read the comment docs.

[oxinabox]

I think we can also use this for FastText maybe? there is at least some code overlap.
Might need to factor out the common inner parts.

We should probably not call it a CSV since it is space seperated,
Or we call it a CSV and pass in a deliminator option

[tejasvaidhyadev]

Deliminator option is added as suggested

[oxinabox]

thanks

[tejasvaidhyadev]

Hi @oxinabox ,
I will soon implement _load_embeddings_csv for FastText as well.
sorry for delay because of my exams are going on.