wattnet-storage 1.1.0

Extensible plugin-based client for unified wattnet time-series metric storage across multiple backends.

Storage Backend and Python Client Interface

wattnet-storage provides two things:

• Storage backend — a Docker Compose stack with ClickHouse for time-series metric storage and Grafana for visualization.
• Python client library — a plugin-based interface to read and write energy metrics (generation, load, carbon footprint, etc.) against the storage backend.

Purpose

Multiple Wattnet containers compute and expose energy data through their own APIs, each with its own domain model and internal representation. wattnet-storage acts as the shared persistence layer between them: it provides a single, uniform interface so that any Wattnet module can persist and retrieve metrics without coupling to a specific storage technology or to the internal data structures of other modules.

To achieve this, metrics are always stored in a common format regardless of how they are represented in the originating domain model:

Field	Type	Description
name	str	Metric identifier (MetricType value, e.g. zone_generation)
value	float	Numeric measurement
timestamp	datetime	Point in time the measurement was taken
metadata	dict	Arbitrary key-value labels (e.g. zone, source, unit)

Any Wattnet module that needs to persist data translates its domain objects into this format before writing, and reconstructs its own representation after reading. The storage backend in use (ClickHouse, or any future backend) is fully transparent to the caller.

Architecture

For the full system architecture see the wattnet-architecture repository.

The client layer is plugin-based (via stevedore). Additional storage backends can be added by implementing BaseStorageClient and registering a wattnet.storage.clients entry point.

Requirements

• Python ≥ 3.10
• Docker and Docker Compose (for the storage backend)

Backend Setup

Start the ClickHouse and Grafana services:

docker compose up -d

Service	Default address
ClickHouse	http://localhost:8123
Grafana	http://localhost:3000

Grafana's default credentials are admin / admin. The ClickHouse datasource and dashboards are provisioned automatically.

Python Client

Installation

pip install wattnet-storage

Or with Poetry:

poetry add wattnet-storage

Configuration

wattnet-storage is configured by the calling application through a StorageConfig object. Create the config in your service and pass it to MetricsRepository.

from wattnet.storage import StorageConfig

config = StorageConfig(
    timeseries_step_minutes=15,
    storage_clients=["clickhouse"],
    plugin_configs={
        "clickhouse": {
            "host": "localhost",
            "port": 8123,
            "user": "default",
            "password": "",
            "database": "wattnet",
            "connect_retries": 5,
            "connect_retry_delay": 3,
        }
    },
)

StorageConfig is immutable: all fields are frozen at construction time. Attempting to reassign a field raises FrozenInstanceError. Validation runs automatically on construction:

• timeseries_step_minutes must be a positive integer — ValueError is raised otherwise.
• storage_clients must not contain duplicate entries — ValueError is raised otherwise.

wattnet-storage does not load or manage a project-specific .env file. If your application uses a .env, it should load it before creating the repository.

Configuration precedence for ClickHouse is:

1. Process environment variables (CLICKHOUSE_*) — highest priority
2. .env file values (CLICKHOUSE_*) — loaded by the consuming application
3. Code defaults in ClickHouseConfig — lowest priority

ClickHouseConfig does not read .env files directly. Consuming applications (wattnet-api, wattnet-core or wattnet-forecast) instantiate ClickHouseConfig(_env_file=".env") in their plugin_settings mechanism and pass the resolved values to StorageConfig.plugin_configs, achieving the priority order above automatically. This means a process-level environment variable (e.g. set in Docker Compose or Kubernetes) always wins over a .env file entry.

StorageConfig.plugin_configs["clickhouse"] is the highest-priority override but is intended for programmatic use only (e.g. tests or one-off scripts); in normal deployments the values come from the environment.

For ClickHouse, these process environment variables are supported:

Variable	Default	Description
CLICKHOUSE_HOST	localhost	ClickHouse hostname
CLICKHOUSE_PORT	8123	ClickHouse HTTP port
CLICKHOUSE_USER	default	ClickHouse username
CLICKHOUSE_PASSWORD	(empty)	ClickHouse password
CLICKHOUSE_DATABASE	wattnet	Target database name
CLICKHOUSE_CONNECT_RETRIES	5	Number of bootstrap attempts before giving up
CLICKHOUSE_CONNECT_RETRY_DELAY	3	Seconds to wait between bootstrap attempts

If ClickHouse is unreachable after all attempts, startup fails with a single RuntimeError message indicating the host, port, and number of attempts — no deep traceback from the HTTP layer. This also handles the typical docker-compose race condition where ClickHouse is not yet ready when the API container starts.

Usage

from datetime import datetime
from wattnet.storage import MetricsRepository, StorageConfig
from wattnet.storage.models import Metric, MetricType

config = StorageConfig(
    timeseries_step_minutes=15,
    storage_clients=["clickhouse"],
    plugin_configs={
        "clickhouse": {
            "host": "localhost",
            "port": 8123,
            "user": "default",
            "password": "",
            "database": "wattnet",
            "connect_retries": 5,
            "connect_retry_delay": 3,
        }
    },
)

repo = MetricsRepository(config)

# Write metrics
metrics = [
    Metric(
        metric_type=MetricType.ZONE_GENERATION,
        value=1500.0,
        timestamp=datetime.now(),
        metadata={"zone": "ES", "source": "solar"},
    )
]
repo.write_metrics(metrics)

# Query metrics
results = repo.query_metrics(
    metric_name=MetricType.ZONE_GENERATION.value,
    start=datetime(2025, 1, 1),
    end=datetime(2025, 1, 2),
    labels={"zone": "ES"},
)

Supported metric types

MetricType	Value	Description
ZONE_GENERATION	zone_generation	Electricity generation in a zone
ZONE_IMPORT	zone_import	Electricity imports
ZONE_EXPORT	zone_export	Electricity exports
ZONE_LOAD	zone_load	Electricity consumption / load
ZONE_MIX_GENERATION	zone_mix_generation	Generation mix per source
FACTOR	factor	Emission factor
LOCAL_FOOTPRINT	local_footprint	Location-based carbon footprint
GLOBAL_FOOTPRINT	global_footprint	Market-based carbon footprint
LOCAL_IMPACT	local_impact	Location-based carbon impact
GLOBAL_IMPACT	global_impact	Market-based carbon impact
LOCAL_SCORE	local_score	Location-based carbon score
GLOBAL_SCORE	global_score	Market-based carbon score
FLOW_SHARE	flow_share	Share of electricity flow between zones
MIX_SHARE	mix_share	Share of generation mix
FOOTPRINT_SHARE	footprint_share	Share attributed to carbon footprint
IMPACT_SHARE	impact_share	Share attributed to carbon impact

Logging

wattnet-storage follows the standard library logging recommendation for libraries: it adds a NullHandler to the wattnet.storage logger and never configures handlers itself. This means no log output appears by default, and the calling application remains fully in control.

To see storage logs, configure the wattnet.storage logger (or the parent wattnet logger) in your application:

import logging

# Minimal setup — output all wattnet.* logs to the console
logging.getLogger("wattnet").setLevel(logging.DEBUG)
logging.getLogger("wattnet").addHandler(logging.StreamHandler())

If you use wattnet-api, wattnet-core or wattnet-forecast, their setup_logging() call already covers wattnet.storage.* logs automatically — no additional configuration is needed.

Related Projects

The following Wattnet components use wattnet-storage as their persistence layer:

• wattnet-api: RESTful API exposing real-time, historical, and forecasted electricity footprint data.
• wattnet-core: Core service that computes carbon and water footprints from electricity generation data.
• wattnet-forecast: Forecasting service for electricity carbon footprint across European zones.

Contributing

Contributions are welcome. See CONTRIBUTING.md for environment setup, code style, how to run the tests, and how to add a new storage backend.

License

This repository is licensed under the Apache License 2.0.

See the LICENSE file for more details.

Funding and Acknowledgments

This work is funded by the European Union's Horizon Europe research and innovation programme through the GreenDIGIT project, under grant agreement 101131207.

© 2026 Spanish National Research Council (CSIC). All rights reserved.