remove dependency on danfojs

[mellster2012]

This gets rid of the danfojs dependency and reduces production dependencies to 12MB (unpacked). There is more cleanup to do, I'd probably eventually get rid of the utils/ folder (and fold the utils methods into ndframe/dataframe/series) and there are a lot of eslint suggestions. But this enables the library for use in lambdas, which is the primary objective for us now. We can always continue to make gradual improvements. Note that I only verified that all the tests pass, and I have removed some non-essential functionality which danfojs provided.

[codecov-commenter]

Welcome to Codecov 🎉

Once you merge this PR into your default branch, you're all set! Codecov will compare coverage reports and display results in all future pull requests.

Thanks for integrating Codecov - We've got you covered ☂️

[chunkai1312]

@mellster2012 Thank you for this significant contribution! Given the scope of the changes, I’ve updated the target branch for this pull request to next.

Please allow me some time to review it thoroughly. I really appreciate your efforts!

[mellster2012]

There is more to do such as allowing to pass options into Stats (currently you cannot pass down the option to download the results to a file instead of viewing them in the browser), and narrowing down the array value types to numbers only where necessary, but it would be great to get this into next soon as a starting point, so we can use dev version releases from next and use them in our projects. It gets messy keeping track of all the local modifications, hopefully we can wrap up this review soon, it should be ok for next to be in-flux for a while. Thank you!

[chunkai1312]

@mellster2012 I've reviewed the changes, but I'm not entirely sure if directly importing the source code of danfojs is the best approach, even though it hasn't been updated in a while. From an initial look, the tests seem to pass without issues. I'll proceed to release this version as 1.0.0-beta.0 to make it available.

If you encounter any additional issues or have feedback, please feel free to share them promptly.

[mellster2012]

Thanks! I agree it's not optimal to heavily lean on their data types as the danfojs repo is quite dated (I did modify them quite a bit though), but we can progressively move away from danfojs and improve the data types from here. Also by making it usable in lambdas we will expose the beta versions right away to the real world of backtesting and uncover potential issues early on.