Implementation of a computation graph
Project description
Pensieve 2.1
"One simply siphons the excess thoughts from one's mind, pours them into the basin, and examines them at one's leisure. It becomes easier to spot patterns and links, you understand, when they are in this form."
—Albus Dumbledore (Harry Potter and the Goblet of Fire by J. K. Rowling)
Pensieve for Data
In J. K. Rowling's amazing world of magic, "a witch or wizard can extract their own or another's memories, store them in the Pensieve, and review them later. It also relieves the mind when it becomes cluttered with information. Anyone can examine the memories in the Pensieve, which also allows viewers to fully immerse themselves in the memories" 1.
Dealing with data during data wrangling and model generation in data science is like dealing with memories except that there is a lot more of back and forth and iteration when dealing with data. You constantly update parameters of your models, improve your data wrangling, and make changes to the ways you visualize or store data. As with most processes in data science, each step along the way may take a long time to finish which forces you to avoid rerunning everything from scratch; this approach is very error-prone as some of the processes depend on others. To solve this problem I came up with the idea of a Computation Graph where the nodes represent data objects and the direction of edges indicate the dependency between them.
After using Pensieve for some time myself, I have found it to be beneficial in several ways:
- error reduction, especially for data wrangling and model creation
- data object organization
- easy transfer of data
- coherent data processing and data pipelines
- data and model reproducibility
- most importantly relieving the mind
Installation
pip install pensieve
Usage
Pensieve stores memories and functions that define the relationship between memories.
from pensieve import Pensieve
# initiate a pensieve
pensieve = Pensieve()
# store a "memory" (with 1 as its content)
pensieve.store(key='one', content=1)
# create a new memory made up of a precursor memory
pensieve.store(key='two', precursors=['one'], function=lambda x: x + x)
There are two types of memories:
- independent memories (without precursors)
- dependent memories (with precursors)
Independent Memories
An independent memory does not have any precursors and instead of a function, which would define the relationship with the precursors, has content.
from pensieve import Pensieve
pensieve = Pensieve()
pensieve.store(key='integers', content=list(range(10)))
Dependent Memories
A dependent memory is created from running a function on the contents of its precursors. When there is only one precursor to a memory, the function can be defined as a lambda with one input which is accessed directly within the function, e.g., lambda x: x + 1.
# the precursor, 'integer' is accessed within the lambda under the label: numbers
pensieve.store(
key='odd_integers', precursors=['integers'],
function=lambda numbers: [x for x in numbers if x%2==1]
)
Memory with Two or More Precursors
If a memory has multiple precursors, its function should still have one input but the precursors should be accessed as items in the input, as if the input is a dictionary of precursors.
For example, if a function adds two precursors x and y, it should be defined as: lambda x: x['x'] + x['y']. In the following example, the function gets a set of integers and odd integers and by filtering out the odd integers from integers, it finds all even integers in the set. This function has only one input, which is called precursors for clarity (but can be called anything) and the precursors are accessed within the function as items 'integers' and 'odd_integers' like a dictionary.
pensieve.store(
key='even_integers',
precursors=['integers', 'odd_integers'],
function=lambda precursors: [
x for x in precursors['integers']
if x not in precursors['odd_integers']
]
)
Retrieving a Memory
Retrieving the content of a memory is like getting an item from a dictionary as shown below.
pensieve['integers']
# output: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
pensieve['even_integers']
# output: [0, 2, 4, 6, 8]
Changing a Memory
When you change a memory in pensieve, all successors get notified and marked as stale but not updated immediately. As soon as a successor of a changed memory is needed it will be updated based on its relationship with its precursor memories.
# changing one memory affects all successors
pensieve.store(key='integers', content=list(range(16)))
pensieve['integers']
# output: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]
pensieve['even_integers']
# output: [0, 2, 4, 6, 8, 10, 12, 14]
Save and Load
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.