- Unified
- Many implementations, one way of access.
- Map
- Apply a function to a list of arguments and collect the results -- serial, parallel or distributed.
This package provides reasonably simple syntax for a frequent programming task which is implemented in various places (built-in, standard library, external libraries). Here are three descriptions of this task:
- Map a list of inputs to a list of results via a user-provided function.
- Apply a given function to a list of arguments to get a list of return values in the same order.
- A black-box function is evaluated with different inputs and the outputs are collected.
In general, this is a so-called "pleasingly parallel problem" (aka "embarrassingly parallel" or "pleasingly parallel") because it is straightforward to separate it into independend subtasks. For this reason, and due to its frequent occurrence, it is recognized as a programming idiom ("parallel map", "parallel for loop") in parallel computing that can equally simple be applied in distributed computing. This package allows to do so with a focus on simplicity of use.
Documentation | GitHub Page |
Source code | GitHub |
Package | PyPI |
Authors | Robert Haas |