This is the Streaming Distributed Bootstrap, a statistical method for estimating uncertainty from finite sampling on bounded or unbounded data.

• The data may be bounded or unbounded.
• The data may be distributed.
• Bootstrap instances are distributed with (ie, local to) the data.
• Supports importance weighting.
• Supports ordinary bootstrap.

cd sdbootstrap
pip install [-e] .

Test it.

python
import sdbootstrap

Be sure to include the scripts in

PATH=$PATH:<path-to>/sdbootstrap/bin

See Wiki.