Ouroboros Chrysopoeia

I believe `pandas` natively uses the `boto3`, `s3fs`, and `fsspec` packages to interact with S3. I don't know more. Note that this issue doesn't exist with reading `csv.bz2` files, for...

> You mean when reading them from S3? Yes. > So the issue is specific to the Parquet format on S3? The issue is **specific to using `smart_open`** with pandas+boto3...

As we know, parquet is a columnar data format. If the file has _n_ columns of data, pandas could in theory try to read it in up to _n_ streams.

Enabling DEBUG level logs may or may not help, but I'll leave this to the developers.

@mrgloom `requirements.txt` is generally meant for standalone applications only, not for imported packages e.g. `hyperas`. Note that `setup.py` already documents the version-agnostic requirements as is normal. Even so, if `requirements.txt`...

OK, but can you publish it on PyPI? It's not hard.

Any update for an estimated release date for Python 3? Python 2 support and compatibility soon won't be necessary. This migration is holding up adoption of nupic. It is just...

@kyle-sorensen It is not about creating a virtualenv; that's easy. It's about integrating with other code that's already in Python 3. Also, one useful thing that I cannot do in...

It would seem that the current status is to provision NuPIC using PyTorch, as in the project https://github.com/numenta/nupic.torch

OK. Thanks. I don't know if using Numpy and Scipy will make it any shorter.