Component Overview¶

A concise map of the repository's top-level components and their relationships.

Component map¶

gls-snowflake-workshop/
│
├── hub/                    # Platform infrastructure
│   ├── src/hub/core/       #   Session, config, credentials, environment
│   ├── src/hub/infra/      #   Bootstrap, schema creation, SQL grants
│   ├── scripts/            #   deploy_infra.py
│   ├── config/             #   infrastructure.yaml
│   └── Makefile            #   deploy-infra
│
├── mock_data/              # Data simulation
│   ├── src/pudo_data/      #   Generators, CLI, config models
│   ├── scripts/            #   seed_shared_data.py
│   ├── config/             #   data_generation.yaml, infrastructure.yaml
│   └── Makefile            #   seed-shared-data, add-morning/evening-data, etc.
│
└── projects/pudo/          # Reference ML project
    ├── src/pudo/
    │   ├── core/           #   Session, config, SQL utils, registry helpers
    │   ├── feature_view/   #   Entities, feature views, feature store setup
    │   ├── training/       #   DAG definition, modeling, training job, ops
    │   └── inference/      #   DAG definition, ops, CLI, utils
    ├── scripts/            #   deploy_*.py, run_*.py
    ├── config/             #   YAML configs with environment overlays
    └── Makefile            #   deploy-*, run-*, evaluate, alerts, summary

Dependency direction¶

graph TD
    mock_data["mock_data"] -->|"seeds data into"| hub["hub"]
    projects_pudo["projects/pudo"] -->|"reads from"| hub
    projects_pudo -->|"reads from"| mock_data

hub creates the platform that other components depend on.
mock_data seeds data into SHARED_DATA (created by hub).
projects/pudo reads from SHARED_DATA, uses feature stores and model registries (created by hub).

Per-component Python environments¶

Each component has its own pyproject.toml and uv.lock:

Component	Python version	Key dependencies
`hub`	3.10	snowflake-snowpark-python, pyyaml
`mock_data`	3.10	snowflake-snowpark-python, pyyaml, click
`projects/pudo`	3.10	snowflake-snowpark-python, snowflake-ml, xgboost, pyyaml, click

Per-component configuration¶

Each component has YAML configuration with environment overlays:

Component	Config directory	Key configs
`hub`	`hub/config/`	`infrastructure.yaml`
`mock_data`	`mock_data/config/`	`data_generation.yaml`, `infrastructure.yaml`
`projects/pudo`	`projects/pudo/config/`	`feature_view/`, `training/`, `inference/`, `infrastructure.yaml`

Per-component entry points¶

Each component exposes operations through its Makefile and scripts/ directory. See the Command Reference for the full list of available targets.