API Reference¶

This section documents Falcon's Python API.

Package Structure¶

falcon/
├── core/              # Core framework
│   ├── graph           # Graph and Node definitions
│   ├── deployed_graph  # Runtime execution
│   └── base_estimator  # Estimator interface
├── estimators/        # Posterior estimation
│   ├── flow            # Flow-based posterior estimation
│   ├── gaussian        # Gaussian posterior estimation
│   └── flow_density    # Normalizing flow networks
├── priors/            # Prior distributions
│   ├── hypercube       # Hypercube mapping prior
│   └── product         # Product of marginals
└── embeddings/        # Observation embeddings
    ├── builder         # Declarative embedding pipelines
    ├── norms           # Online normalization utilities
    └── svd             # Streaming PCA

Core Classes¶

Class	Description
`Graph`	Container for computational graph nodes
`Node`	Single random variable in the graph
`DeployedGraph`	Runtime orchestration with Ray
`BaseEstimator`	Abstract interface for estimators

Estimators¶

Class	Description
`Flow`	Flow-based posterior estimation (normalizing flows)
`Gaussian`	Full covariance Gaussian posterior
`FlowDensity`	Normalizing flow `nn.Module` (internal)

Priors¶

Class	Description
`Hypercube`	Hypercube-to-target distribution mapping
`Product`	Product of independent marginals with latent space transforms

Embeddings¶

Class / Function	Description
`instantiate_embedding`	Declarative embedding pipeline builder
`LazyOnlineNorm`	Online normalization
`DiagonalWhitener`	Diagonal whitening
`PCAProjector`	Streaming PCA projector

Quick Import¶

import falcon

# Core
from falcon import Graph, Node, CompositeNode, DeployedGraph

# Estimators
from falcon.estimators import Flow, Gaussian

# Priors
from falcon.priors import Hypercube, Product

# Embeddings
from falcon.embeddings import instantiate_embedding, LazyOnlineNorm, PCAProjector

# Utilities
from falcon import read_run, load_run, read_samples