fuelog-python

Official Python client library for the Fuelog API.

Covers the full REST API and the MCP (Model Context Protocol) server — with zero third-party dependencies, typed models, and a comprehensive test suite.

Installation

pip install fuelog

Requires Python 3.9+. No external dependencies.

Quick Start

REST Client

from fuelog import FuelogClient, CreateFuelLogRequest, CreateVehicleRequest, FuelType

client = FuelogClient(token="your_bearer_token")

# ── Logs ───────────────────────────────────────────────────────────────────

# List the 20 most recent fill-ups
logs = client.list_logs(limit=20)
for log in logs:
    print(f"{log.timestamp}  {log.brand}  €{log.cost}  {log.distance_km} km")

# Create a new fill-up
new_id = client.create_log(
    CreateFuelLogRequest(
        brand="Shell",
        cost=50.00,
        distance_km=400.0,
        fuel_amount_liters=35.0,
        vehicle_id="vehicle-abc",
    )
)
print(f"Created log: {new_id}")

# Update a log
from fuelog import UpdateFuelLogRequest
client.update_log(new_id, UpdateFuelLogRequest(brand="Shell Express", cost=51.00))

# Delete a log
client.delete_log(new_id)

# ── Vehicles ───────────────────────────────────────────────────────────────

vehicles = client.list_vehicles()

vehicle_id = client.create_vehicle(
    CreateVehicleRequest(
        name="My Corolla",
        make="Toyota",
        model="Corolla",
        year="2022",
        fuel_type=FuelType.PETROL,
        is_default=True,
    )
)

from fuelog import UpdateVehicleRequest
client.update_vehicle(vehicle_id, UpdateVehicleRequest(name="Daily Driver"))
client.delete_vehicle(vehicle_id)

# Include archived vehicles
all_vehicles = client.list_vehicles(include_archived=True)

# ── Analytics ──────────────────────────────────────────────────────────────

stats = client.get_analytics(
    vehicle_id="vehicle-abc",
    start_date="2025-01-01",
    end_date="2025-12-31",
)
print(f"Avg efficiency: {stats.efficiency} km/L")
print(f"Total spent:    {stats.home_currency} {stats.total_spent}")

MCP Client

The MCP client wraps all tools, resources, and prompts registered on the /api/mcp endpoint.

from fuelog import FuelogMCPClient, FuelType, TrendMetric, CostOptimizationPeriod

mcp = FuelogMCPClient(token="your_bearer_token")

# ── Tools ──────────────────────────────────────────────────────────────────

# List logs (with optional filters)
result = mcp.list_logs(limit=10, brand="Shell")
print(result.text)

# Log a fill-up
result = mcp.log_fuel(
    brand="BP",
    cost=45.0,
    distance_km=300.0,
    fuel_amount_liters=28.0,
    vehicle_id="v1",
)

# Edit and delete
mcp.edit_fuel_log("log-id", cost=46.0)
mcp.delete_fuel_log("log-id")  # confirm=True added automatically

# Vehicles
mcp.add_vehicle(name="EV", make="Tesla", model="Model 3", year="2024",
                fuel_type=FuelType.ELECTRIC, is_default=True)
mcp.update_vehicle("v1", is_archived=True)

# Analytics
result = mcp.get_analytics(vehicle_id="v1")
result = mcp.compare_vehicles(vehicle_ids=["v1", "v2"])

# ── Resources ──────────────────────────────────────────────────────────────

summary = mcp.get_analytics_summary_resource()
monthly = mcp.get_analytics_monthly_resource()
profile = mcp.get_profile_resource()

# By ID
log_data    = mcp.get_log_resource("log-123")
vehicle_data = mcp.get_vehicle_resource("v-abc")

# Raw URI access
data = mcp.read_resource("fuelog://analytics/vehicles")

# ── Prompts ────────────────────────────────────────────────────────────────

# Monthly report for March 2025
prompt = mcp.monthly_report(month="2025-03")
for msg in prompt.messages:
    print(f"[{msg.role}]", msg.content.get("text", ""))

# Trend analysis
prompt = mcp.trend_analysis(
    start_date="2025-01-01",
    end_date="2025-12-31",
    metric=TrendMetric.KML,
)

# Cost optimisation suggestions
prompt = mcp.cost_optimization(period=CostOptimizationPeriod.LAST_QUARTER)

Error Handling

All errors derive from FuelogError:

from fuelog import FuelogClient, FuelogNotFoundError, FuelogForbiddenError

client = FuelogClient(token="sk-...")

try:
    client.delete_log("non-existent-id")
except FuelogNotFoundError:
    print("Log not found")
except FuelogForbiddenError as e:
    print(f"Permission denied: {e}")

Configuration

client = FuelogClient(
    token="your_token",
    base_url="https://api.fuelog.app",  # default
    timeout=30,                          # seconds, default
)

Development

# Clone and install in editable mode with dev dependencies
git clone https://github.com/Irishsmurf/fuelog-python.git
cd fuelog-python
pip install -e ".[dev]"

# Run the test suite with coverage
pytest --cov=fuelog --cov-report=term-missing

# Lint
ruff check fuelog/ tests/

# Type check
mypy fuelog/

# Build for release
pip install hatch
hatch build

Publishing to PyPI

Releases are published automatically via GitHub Actions when a GitHub Release is created. The workflow uses OIDC Trusted Publishing — no PyPI API token is stored in repository secrets.

To set up a new repo:

1. Create a Trusted Publisher on pypi.org/manage/account/publishing/ pointing to this repository.
2. Tag a release: git tag vX.Y.Z && git push --tags.
3. Create a GitHub Release from the tag — the publish.yml workflow triggers automatically.

License

MIT