Tom Augspurger

Python Array API Standard

28

Hi all, A group of folks from various projects / companies have put together a draft of an Array API standard, and are looking for feedback. - Blog post: https://data-apis.org/blog/array_api_standard_release...

Maintain a Grafana Chart

5

Pangeo is publishing some grafana dashboards of their cluster usage, including their Dask clusters, at http://grafana.us-central1-b.gcp.pangeo.io/grafana/d/a_khcwhWk/dask?orgId=1&refresh=5s. This is based off of (I think) @ig248's chart at https://grafana.com/grafana/dashboards/11522. ![Screen Shot 2020-08-06...

Project Permissions

7

We want to avoid cases where a project has a sole maintainer role on a service (GitHub, PyPI, conda-forge, etc.). https://docs.google.com/spreadsheets/d/1lIF6kDWWoL7CqOLbTg2hChCzvHLQW-gMZofGrAoKtMo/edit?usp=sharing has a spreadsheet with things asof today. I've bolded...

Update README.md

2

I believe these docs were out of date.

ValueError ValueError: numpy.ufunc has the wrong size, try recompiling. Expected 192, got 216

6

I think there's an issue with the Conda-forge package for NumPy 1.15.x ``` conda create -n tst -c conda-forge --force python=3.6.7 fastparquet numpy=1.15.4 ``` ``` bash-5.0$ python -c "import fastparquet"...

Should `partition_on` columns be included in the pandas_metadata?

2

I think they should be. Just checking if there was a reason they weren't @martindurant ```python In [20]: import pandas as pd In [21]: import fastparquet as fp In [22]:...

CV works

1

Closes #21 Closes #31

Check column names before passing on

5

e.g. if they're ints, xgboost will refuse them. ```python import pandas as pd import numpy as np import dask.dataframe as dd import dask_xgboost as xgb from distributed import Client df...

Given the relatively low traffic here, it'd be nice to regularly run the tests to catch changes in XGBoost / dask.

Add collections conformance

1

(hopefully this is the right place to report this) In https://github.com/stac-utils/pystac-client/issues/320, @system123 noticed that pystac-client's `.get_collections()` method was slow. It relies on STAC API advertising a `/collections` endpoint (see https://github.com/stac-utils/pystac-client/blob/5d7510c5307f04cdbd3bd694f4b919a9bbf9c275/pystac_client/client.py#L182-L1850)....