Skip to main content

​
Overview

The Cortex class provides methods for predicting future states of twins and aggregates using the Cortex prediction engine. It can return predictions either mapped to human-readable embedding inputs or as raw state vectors.

​
Methods

​
predictions(data: NextStateRequest)

Predict the next n states for a twin and get responses mapped back to embedding inputs.
​
data
NextStateRequest
required
The prediction request parameters
​
response
tuple
Returns a tuple of (NextStateResponse | Error, status_code)
Example: 
from avenieca.api.model import NextStateRequest

nsr = NextStateRequest(
    module_id="aggregate001",
    recall=20,
    range=20,
    n=1,
    status="e"
)

res, status = eca.cortex.predictions(data=nsr)
if status == 200:
    print("Current state:")
    for twin in res.current_state:
        print(f"  {twin.module_id}: {twin.state}")
    
    print("\nPredicted next state:")
    for twins in res.next_state:
        for twin in twins.list:
            print(f"  {twin.module_id}: {twin.state}")

​
predictions_raw(data: NextStateRequest)

Predict the next n states for a twin and get responses as raw aggregate state vectors.
​
data
NextStateRequest
required
The prediction request parameters (same as predictions method)
​
response
tuple
Returns a tuple of (NextStateResponseRaw | Error, status_code)
Example: 
from avenieca.api.model import NextStateRequest

nsr = NextStateRequest(
    module_id="aggregate001",
    recall=20,
    range=20,
    n=3,  # Predict next 3 states
    status="e"
)

res, status = eca.cortex.predictions_raw(data=nsr)
if status == 200:
    print("Current state (raw):")
    for twin in res.current_state:
        print(f"  {twin.module_id}: {twin.state}")
    
    print(f"\nPredicted next {len(res.next_state)} states:")
    for i, twins in enumerate(res.next_state):
        print(f"\n  Step {i+1}:")
        for twin in twins.list:
            print(f"    {twin.module_id}: {twin.state}")

​
Understanding Predictions

The Cortex prediction engine uses historical sequences to predict future states:
  1. Range: Number of nearest neighbor sequences to consider
  2. Recall: How many historical states to look back
  3. N: Number of future states to predict
  4. Status: Filter sequences by their status value

​
Prediction Workflow

import os
from avenieca.api.eca import ECA
from avenieca.api.model import Config, NextStateRequest

# Initialize client
config = Config(
    uri="http://localhost:2580/v1",
    username=os.getenv("USERNAME"),
    password=os.getenv("PASSWORD")
)
eca = ECA(config)

# Make prediction request
request = NextStateRequest(
    module_id="air_conditioner",
    recall=10,      # Look back 10 states
    range=5,        # Consider 5 nearest sequences
    n=2,            # Predict next 2 states
    status="e",     # Only use 'existing' status sequences
    store_response=True,   # Save this prediction
    store_sequence=False   # Don't create new sequence
)

# Get human-readable predictions
res, status = eca.cortex.predictions(data=request)

if status == 200:
    # Display current state
    print("Current State:")
    for twin in res.current_state:
        print(f"  Module: {twin.module_id}")
        print(f"  ESS ID: {twin.ess_id}")
        print(f"  State: {twin.state}")
    
    # Display predictions
    print(f"\nPredicted States:")
    for step_num, twins_set in enumerate(res.next_state, 1):
        print(f"\n  Prediction {step_num}:")
        for twin in twins_set.list:
            print(f"    {twin.module_id}: {twin.state}")

​
When to Use Raw vs Mapped Predictions

Use predictions() when:
  • You need human-readable output
  • States are mapped to embedding inputs
  • Building user-facing features
Use predictions_raw() when:
  • You need numerical state vectors for computation
  • Performing further mathematical operations
  • States don’t have embedding mappings
  • Building analytics or ML pipelines

Build docs developers (and LLMs) love

Get started for freeTalk to us