Valeriy Mukhtarulin

Alternatively, does `yaml` solve this problem?

pandas-profiling does something similar https://github.com/pandas-profiling/pandas-profiling/blob/master/pandas_profiling/model/base.py#L122

It can be: 1. A config 2. Parameters from top-level - `arche.report_all(threshold=0.2)` 3. Environment variables - in a notebook ``` %env THRESHOLD=0.2 ... arche.report_all() ```

@ejulio Any comments or ideas?

@ejulio Do you think these threshold should be treated equally or be separated? 1. Coverage diff between jobs (currently 0.1 for errors and 0.05 for warnings) 2. Category diff between...

I tried it a kaggle kernel, and it just didn't work.

Workaround: classic notebook UI.

@Alexandr1988 What do you think?

The solution is to use `nodata` https://github.com/scrapinghub/python-scrapinghub/pull/129