pyprimed: a python library to manage PrimedIO

Create a personalized web application that is unique and relevant for each and every user with Primed.io.

Installation

pip install pyprimed

Quickstart

Import the SDK and initiate the connection

from pyprimed.pio import Pio

pio = Pio(uri='http://<user>:<password>@<api_url>:<port>')

Create a Universe, and attach a few Targets

# create a new universe and attach a single target
pio\
  .universes\
  .create(name='myfirstuniverse')\
  .targets\
  .upsert([{'key':'ARTICLE-1', 'value':{'url': 'www.example.com/article-1'}}])

# retrieving the newly created universe
u = pio.universes.filter(name='myfirstuniverse').first

# list all targets currently associated with this universe
for target in u.targets.all():
  print(target.key, target.created_at)

# prepare a list of new targets
new_targets = [
  {'key': 'ARTICLE-2', 'value': {'url': 'www.example.com/article-2'}}, 
  {'key': 'ARTICLE-3', 'value': {'url': 'www.example.com/article-3'}}
]

# upsert the new targets
u.targets.upsert(new_targets)

# targets are upserted, which means that for a given key there
# can be only one instance in the database. Trying to create an
# instance with the same key will update the value of the record
# in the database
u.targets.upsert([{'key':'ARTICLE-1', 'value':{'url': 'THIS IS NEW!'}}])
u.targets.filter(key='ARTICLE-1').first.value 

Create a Model, and attach a few Signals

# create a new model and attach a single signal
pio\
  .models\
  .create(name='myfirstmodel')\
  .signals\
  .upsert([{'key':'ALICE'}])

# retrieving the created model
m = pio.models.filter(name='myfirstmodel').first

# list all signals currently associated with this model
for signal in m.signals.all():
  print(signal.key, signal.created_at)

# prepare a list of new signals
new_signals = [
  {'key': 'BOB'}, 
  {'key': 'CHRIS'}
]

# create the new signals
m.signals.upsert(new_signals)

# prepare a set of predictions (sk stand for signal.key, and tk for target.key)
# WARNING: `sk` and `tk` should always be a string!
predictions = [
  {'sk': 'ALICE', 'tk': 'ARTICLE-1', 'score': 0.35},
  {'sk': 'BOB', 'tk': 'ARTICLE-1', 'score': 0.75}, 
  {'sk': 'CHRIS', 'tk': 'ARTICLE-1', 'score': 0.15}
]

# create the new predictions 
u = pio.universes.filter(name='myfirstuniverse').first

pio\
    .predictions\
    .on(model=m, universe=u)\
    .upsert(predictions, asynchronous=False)
# the `asynchronous=False` settings waits for this operation to end before 
# continuing

Create a Campaign, Experiment and set up an AB test to start using the Predictions

from pyprimed.models.abvariant import CustomAbvariant, RandomControlAbvariant, NullControlAbvariant

# we create a custom abvariant that blends models m1 and m2 using a 60%/40% weight ratio
ab0 = CustomAbvariant(label='A', models={m1: 0.6, m2: 0.4})
ab1 = RandomControlAbvariant(label='B')
ab2 = NullControlAbvariant(label='C')

# we attach the abvariants to the experiment
# `ab0` will receive 80% of traffic, `ab1` and `ab2`
# receive 10% each
e.abvariants.create({ab0: 0.8, ab1: 0.1, ab2: 0.1})

res = c.personalise(
  pubkey='mypubkey',
  secretkey='mysecretkey',
  signals={'userid': 'BOB'},
  abvariant='A'
)  # abvariant with label 'A' will be returned

Update Experiment

from pyprimed.models.abvariant import CustomAbvariant, RandomControlAbvariant, NullControlAbvariant

# obtain existing experiment
e = c.experiments.filter(name="myexperiment").first()

# change property
e.salt = "new_salt"
e.update()

# update abvariants
ab1 = CustomAbvariant(label="NEWLABEL", {m1: 0.33, m2: 0.77})
ab2 = RandomControlAbvariant(label="NEWRANDOMCONTROL")

e.abvariants.update({ab1: 0.5, ab2: 0.5})

Developer

Build the documentation:

cd docs && pydocmd build

Preview documentation on http://localhost:8000

cd docs && pydocmd serve