José Morales

Hey. https://github.com/Nixtla/neuralforecast/pull/833 fixed a problem where if you had integer times you could provide anything as `freq`, e.g. `freq='hello'` and it would be internally replaced by `freq=1`. That PR made...

I think this is due to pandas 2.2 deprecating some aliases ([ref](https://pandas.pydata.org/docs/whatsnew/v2.2.0.html#deprecate-aliases-m-q-y-etc-in-favour-of-me-qe-ye-etc-for-offsets)). So `YE-DEC` works in pandas>=2.2 but doesn't on older versions.

Hey @sankarsiva123, thanks for using statsforecast. plotly was moved to be an optional dependency in #600, so I'll rename this to reference fugue instead. I believe this issue is similar...

Yes, we're going to make the package as lightweight as possible. We've moved the plotting to [utilsforecast](https://github.com/Nixtla/utilsforecast), which will be a dependency for this library but that only depends on...

We're working to keep only the minimum required dependencies in the next release. @kvnkho do you have any suggestions for @vspinu's problem?

Since we're already using ruff it may be easier to just adopt its [formatter](https://docs.astral.sh/ruff/formatter/), it's supposed to be the same as black but faster, so we can avoid an extra...

I support the pre-commit as well, it's very easy to get linter errors (especially from isort) and that would help a lot.

Just adding what you said last time about replacing isort with ruff ([full comment](https://github.com/microsoft/LightGBM/pull/5871#issuecomment-1543301789)): > I chose NOT to have it run isort, even though it can, because it currently...

Hey @tg2k, thanks for using neuralforecast. The data is converted to long format, there are 100 images of 64 pixels, which become 100 unique_ids with 64 timestamps each. I believe...

Hey @cargecla1, I just pushed some changes. Please let me know if they look good to you, if they do we can merge this.