get_consumption_analysis
Returns the sample household's recorded consumption between two dates from the repository's data, for the model to analyse; the dates must fall inside that data's range.
LABSCO SUMMARY
A forecast you can interrogate, not just read.
The explainability tool is what separates this from a plain prediction endpoint: you can ask what drives the model's forecast for an appliance over a horizon, then line that against the actual price for the period. The forecasts describe the sample household the models were trained on, not yours, so read it as a worked example of the method. Note the licence before you build on it — the author restricts use to academic evaluation, reference and non-commercial educational purposes with attribution, and asks that commercial use be arranged directly.
What it is
A Python MCP server on top of a master's-thesis energy-prediction system: machine-learning models trained on one sample household (its data ships in the repository), SHAP values that explain them, and the ESIOS API for Spanish electricity prices.
What you get
predict_consumption()— the sample household's forecast consumption, in Wh, between a futureinit_dateandend_dateget_consumption_analysis()— the sample household's recorded consumption over a date range, broken down by applianceexplain_predictions()— SHAP values for a givenapplianceover ahorizon, so the forecast is not a black boxget_precio()— electricity prices for a date range and aprice_typeget_precio_inteligente()— price for atarget_datewith automatic fallback when the primary source has nothing
Requirements
Python 3.8 or higher, an ESIOS account with Red Eléctrica de España, and the token in ESIOS_API_TOKEN. The client launches it with uv from your clone of the repository, whose bundled data and models the forecast tools read; the ESIOS token is only needed by the two price tools.
Setup effort
Build from source
USE CASES
01
Setting the sample household's forecast against the price
Forecast the sample household's consumption for coming dates, then pull prices for those dates with your ESIOS token.
02
Breaking down the sample household's use by appliance
Run the historical analysis for dates inside the bundled data, split by appliance.
03
Checking whether to trust a forecast
Ask for the SHAP explanation for the appliance and horizon that drives it.
5 tools
Listed in project code.
predict_consumption
Predicts the sample household's consumption, in Wh, between two future dates with the project's trained models; for past dates use get_consumption_analysis.
get_precio
Gets electricity prices from the ESIOS API for a date range; price_type is actual or pvpc (PVPC prices), market (OMIE day-ahead), forecast, or marginal.
get_precio_inteligente
Gets the electricity price for one date: real PVPC prices for past dates, the ESIOS forecast for future ones, or an estimate from the last 7 days' average when neither is available.
explain_predictions
Returns the stored SHAP values for one appliance of the sample household (e.g. Fridge, Washing_Machine) over a horizon in days, so the model can say which inputs weigh most.
BEFORE YOU INSTALL
What to know first
Tools you get
5 tools registered in the project's code.
Requires
- uv on PATH
- Python 3.8
- ESIOS_API_TOKEN in your environment
Where it runs
The documented client runs the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the 5 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/Davad122/TFM.gitMCP Clients
Any MCP client
ANY MCP CLIENT · JSON
{
"mcpServers": {
"mcp-david-TFM": {
"command": "uv",
"args": [
"--directory",
"<YOUR_PATH_TO_PROJECT>",
"run",
"-m",
"davidElectric"
],
"env": {
"ESIOS_API_TOKEN": "<YOUR_ESIOS_API_TOKEN>"
}
}
}
}
