{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## **Quickstart Guide**\n",
"\n",
"This guide covers the standard usage pattern and basic functionality to help you get started with twinLab. In this jupyter notebook we will:\n",
"\n",
"1. Upload a dataset to twinLab.\n",
"2. Use `Emulator.train` to train a surrogate model.\n",
"3. Use the model to make a prediction with `Emulator.predict`.\n",
"4. Visualise the results and their uncertainty.\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# Third-party imports\n",
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import pandas as pd\n",
"\n",
"# Project imports\n",
"import twinlab as tl"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### **Your twinLab information**\n",
"\n",
"Confirm your twinLab version\n"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'cloud': '2.0.0',\n",
" 'modal': '0.2.0',\n",
" 'library': '1.2.0',\n",
" 'image': 'jasper-twinlab-deployment'}"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tl.versions()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"And view your user information, including how many credits you have.\n"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'username': 'jasper@digilab.co.uk', 'credits': 0}"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tl.user_information()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### **Upload a dataset**\n",
"\n",
"Datasets must be data presented as a `pandas.DataFrame` object, or a filepaths which points to a csv file that can be parsed to a `pandas.DataFrame` object. **Both must be formatted with clearly labelled columns.** Here, we will label the input (predictor) variable `x` and the output variable `y`. In `twinlab`, data is expected to be in column-feature format, meaning each row represents a single data sample, and each column represents a data feature.\n",
"\n",
"`twinLab` contains a `Dataset` class with attirbutes and methods to process, view and summarise the dataset. Datasets must be created with a `dataset_id` which is used to access them. The dataset can be uploaded using the `upload` method.\n"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" x \n",
" y \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 0.696469 \n",
" -0.817374 \n",
" \n",
" \n",
" 1 \n",
" 0.286139 \n",
" 0.887656 \n",
" \n",
" \n",
" 2 \n",
" 0.226851 \n",
" 0.921553 \n",
" \n",
" \n",
" 3 \n",
" 0.551315 \n",
" -0.326334 \n",
" \n",
" \n",
" 4 \n",
" 0.719469 \n",
" -0.832518 \n",
" \n",
" \n",
" 5 \n",
" 0.423106 \n",
" 0.400669 \n",
" \n",
" \n",
" 6 \n",
" 0.980764 \n",
" -0.164966 \n",
" \n",
" \n",
" 7 \n",
" 0.684830 \n",
" -0.960764 \n",
" \n",
" \n",
" 8 \n",
" 0.480932 \n",
" 0.340115 \n",
" \n",
" \n",
" 9 \n",
" 0.392118 \n",
" 0.845795 \n",
" \n",
" \n",
"
\n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" x \n",
" \n",
" \n",
" \n",
" \n",
" 0 \n",
" 0.000000 \n",
" \n",
" \n",
" 1 \n",
" 0.007874 \n",
" \n",
" \n",
" 2 \n",
" 0.015748 \n",
" \n",
" \n",
" 3 \n",
" 0.023622 \n",
" \n",
" \n",
" 4 \n",
" 0.031496 \n",
" \n",
" \n",
" ... \n",
" ... \n",
" \n",
" \n",
" 123 \n",
" 0.968504 \n",
" \n",
" \n",
" 124 \n",
" 0.976378 \n",
" \n",
" \n",
" 125 \n",
" 0.984252 \n",
" \n",
" \n",
" 126 \n",
" 0.992126 \n",
" \n",
" \n",
" 127 \n",
" 1.000000 \n",
" \n",
" \n",
"
\n",
"
128 rows × 1 columns
\n",
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Plot parameters\n",
"nsigs = [1, 2]\n",
"# nsigs = [0.674, 1.960, 2.576]\n",
"color = \"blue\"\n",
"alpha = 0.5\n",
"plot_training_data = True\n",
"plot_model_mean = True\n",
"plot_model_bands = True\n",
"\n",
"# Plot results\n",
"grid = df_eval[\"x\"]\n",
"mean = df_mean\n",
"err = df_stdev\n",
"if plot_model_bands:\n",
" label = r\"Model prediction\"\n",
" plt.fill_between(grid, np.nan, np.nan, lw=0, color=color, alpha=alpha, label=label)\n",
" for isig, nsig in enumerate(nsigs):\n",
" plt.fill_between(\n",
" grid,\n",
" mean - nsig * err,\n",
" mean + nsig * err,\n",
" lw=0,\n",
" color=color,\n",
" alpha=alpha / (isig + 1),\n",
" )\n",
"if plot_model_mean:\n",
" label = r\"Model prediction\" if not plot_model_bands else None\n",
" plt.plot(grid, mean, color=color, alpha=alpha, label=label)\n",
"if plot_training_data:\n",
" plt.plot(df[\"x\"], df[\"y\"], \".\", color=\"black\", label=\"Training data\")\n",
"plt.xlim((0.0, 1.0))\n",
"plt.xlabel(r\"$X$\")\n",
"plt.ylabel(r\"$y$\")\n",
"plt.legend()\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### **Deleteing datasets and campaigns**\n",
"\n",
"To keep your cloud storage tidy you should delete your datasets and emulators when you are finished with them. `Emulator.delete` and `Dataset.delete` deletes the emulators and the datasets from the cloud storage respectively.\n"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"# Delete dataset\n",
"dataset.delete()\n",
"\n",
"# Delete campaign\n",
"emulator.delete()"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.0"
}
},
"nbformat": 4,
"nbformat_minor": 4
}