FiQCI Resource Estimator

Simple web app to estimate how many QPU seconds a job takes based on a model.

Quick Start

Web Application

The web application uses React, TailwindCSS and Vite. You can build the static files with:

npm run build

For local development:

npm run dev

Data collection and model generation

To collect data and generate models, a python environment is required and the python code in this project to be installed.

# 1. Create and activate virtual environment (using uv)
uv venv --python 3.11
source .venv/bin/activate

# 2. Install package in editable mode with all dependencies
uv pip install -e .

# 3. Generate data from quantum hardware
resource-estimator-generate --server-url <IQM_URL> --output data_analysis/data/device.csv

# Optional: Use checkpoint for long-running data collection (auto-resumes if interrupted)
resource-estimator-generate --server-url <IQM_URL> --output data.csv --checkpoint data.checkpoint.csv

# 4. Build the model
resource-estimator-build --data data_analysis/data/device.csv --device helmi

# 5. Validate and visualize
resource-estimator-validate --data data_analysis/data/device.csv --device helmi --output plots/

# 6. Update frontend with new model parameters (copy JS output from step 4)
# 7. Rebuild and redeploy
npm run build

Development

# Install package with all dependencies (including dev)
uv sync --group dev

# Run tests with coverage
uv run pytest tests/ -v --cov=src/resource_estimator --cov-report=term-missing

# Install pre-commit hooks for automatic checks
uv run pre-commit install

Data Availability

Data used in the resource estimation can be found in data_analysis/data/. It contains comma-separated value data of varying qubits, circuit depth, number of circuits in a batch, and shots with the real QPU seconds runtime calculated from IQM Timestamps as execution_end-execution_start.

Model Details

The resource estimator uses a polynomial ridge regression models using data from running on the hardware. Each device has its own model parameters stored in src/utils/ResourceEstimatorModel.js.

Model features:

• Number of qubits
• Circuit depth
• Number of circuits (batches)
• Number of shots (k_shots = shots/1000)

Updating the Model

When you need to update the model (e.g., after software or hardware changes):

1. Generate fresh data using resource-estimator-generate (with --checkpoint for reliability)
2. Build new model using resource-estimator-build
3. Validate model using resource-estimator-validate
4. Update frontend by copying the JavaScript output into src/utils/ResourceEstimatorModel.js
5. Rebuild the frontend with npm run build
6. Redeploy (automatic via GitHub Actions)

Documentation

The user-facing documentation on the webpage can be changed by editing public/documentation.md.