{ "cells": [ { "cell_type": "markdown", "id": "2d17761e", "metadata": {}, "source": [ "# df038_NumbaDeclare\n", "This tutorial illustrates how PyROOT supports declaring C++ callables from\n", "Python callables making them, for example, usable with RDataFrame. The feature\n", "uses the numba Python package for just-in-time compilation of the Python callable\n", "and supports fundamental types and ROOT::RVec thereof.\n", "\n", "\n", "\n", "\n", "**Author:** Stefan Wunsch \n", "This notebook tutorial was automatically generated with ROOTBOOK-izer from the macro found in the ROOT repository on Tuesday, May 19, 2026 at 08:10 PM." ] }, { "cell_type": "code", "execution_count": 1, "id": "5e790f77", "metadata": { "collapsed": false, "execution": { "iopub.execute_input": "2026-05-19T20:10:22.042096Z", "iopub.status.busy": "2026-05-19T20:10:22.041927Z", "iopub.status.idle": "2026-05-19T20:10:23.106432Z", "shell.execute_reply": "2026-05-19T20:10:23.105590Z" } }, "outputs": [], "source": [ "import ROOT" ] }, { "cell_type": "markdown", "id": "4e594791", "metadata": {}, "source": [ "To mark a Python callable to be used from C++, you have to use the decorator\n", "provided by PyROOT passing the C++ types of the input arguments and the return\n", "value." ] }, { "cell_type": "code", "execution_count": 2, "id": "b306674f", "metadata": { "collapsed": false, "execution": { "iopub.execute_input": "2026-05-19T20:10:23.119477Z", "iopub.status.busy": "2026-05-19T20:10:23.119322Z", "iopub.status.idle": "2026-05-19T20:10:23.918315Z", "shell.execute_reply": "2026-05-19T20:10:23.896396Z" } }, "outputs": [], "source": [ "@ROOT.Numba.Declare(['float', 'int'], 'float')\n", "def pypow(x, y):\n", " return x**y" ] }, { "cell_type": "markdown", "id": "7cb91cfc", "metadata": {}, "source": [ "The Python callable is now available from C++ in the Numba namespace.\n", "For example, we can use it from the interpreter." ] }, { "cell_type": "code", "execution_count": 3, "id": "7eeb7262", "metadata": { "collapsed": false, "execution": { "iopub.execute_input": "2026-05-19T20:10:23.930548Z", "iopub.status.busy": "2026-05-19T20:10:23.930294Z", "iopub.status.idle": "2026-05-19T20:10:24.046099Z", "shell.execute_reply": "2026-05-19T20:10:24.045723Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2^3 = 8\n" ] }, { "data": { "text/plain": [ "139983781803136" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ROOT.gInterpreter.ProcessLine('cout << \"2^3 = \" << Numba::pypow(2, 3) << endl;')" ] }, { "cell_type": "markdown", "id": "251ab8c6", "metadata": {}, "source": [ "Or we can use the callable as well within a RDataFrame workflow." ] }, { "cell_type": "code", "execution_count": 4, "id": "05b5cb6e", "metadata": { "collapsed": false, "execution": { "iopub.execute_input": "2026-05-19T20:10:24.063241Z", "iopub.status.busy": "2026-05-19T20:10:24.063096Z", "iopub.status.idle": "2026-05-19T20:10:25.707298Z", "shell.execute_reply": "2026-05-19T20:10:25.706645Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "pypow([0. 1. 2. 3.], 3) = [ 0. 1. 8. 27.]\n" ] } ], "source": [ "data = ROOT.RDataFrame(4).Define('x', '(float)rdfentry_')\\\n", " .Define('x_pow3', 'Numba::pypow(x, 3)')\\\n", " .AsNumpy()\n", "\n", "print('pypow({}, 3) = {}'.format(data['x'], data['x_pow3']))" ] }, { "cell_type": "markdown", "id": "e35429fb", "metadata": {}, "source": [ "ROOT uses the numba Python package to create C++ functions from python ones.\n", "We support as input and return types of the callable fundamental types and\n", "ROOT::RVec thereof. See the following callable computing the power of the\n", "elements in an array." ] }, { "cell_type": "code", "execution_count": 5, "id": "ce9ffbc0", "metadata": { "collapsed": false, "execution": { "iopub.execute_input": "2026-05-19T20:10:25.710687Z", "iopub.status.busy": "2026-05-19T20:10:25.710457Z", "iopub.status.idle": "2026-05-19T20:10:27.694434Z", "shell.execute_reply": "2026-05-19T20:10:27.693228Z" } }, "outputs": [ { "data": { "text/plain": [ "139983781803136" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" }, { "name": "stdout", "output_type": "stream", "text": [ "pypowarray({ 0, 1, 2, 3 }, 3) = { 0, 1, 8, 27 }\n" ] } ], "source": [ "@ROOT.Numba.Declare(['RVecF', 'int'], 'RVecF')\n", "def pypowarray(x, y):\n", " return x**y\n", "\n", "ROOT.gInterpreter.ProcessLine('''\n", "ROOT::RVecF x = {0, 1, 2, 3};\n", "cout << \"pypowarray(\" << x << \", 3) = \" << Numba::pypowarray(x, 3) << endl;\n", "''')" ] }, { "cell_type": "markdown", "id": "6d3bdaf3", "metadata": {}, "source": [ "and now with RDataFrame" ] }, { "cell_type": "code", "execution_count": 6, "id": "182283d1", "metadata": { "collapsed": false, "execution": { "iopub.execute_input": "2026-05-19T20:10:27.713161Z", "iopub.status.busy": "2026-05-19T20:10:27.712970Z", "iopub.status.idle": "2026-05-19T20:10:28.533931Z", "shell.execute_reply": "2026-05-19T20:10:28.528650Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "sum(pypowarray({ 1, 2, 3 }, 2)) = " ] }, { "name": "stdout", "output_type": "stream", "text": [ " 14.0\n" ] } ], "source": [ "s = ROOT.RDataFrame(1).Define('x', 'ROOT::RVecF{1,2,3}')\\\n", " .Define('x2', 'Numba::pypowarray(x, 2)')\\\n", " .Sum('x2') # 1 + 4 + 9 == 14\n", "print('sum(pypowarray({ 1, 2, 3 }, 2)) = ', s.GetValue())" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "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": "3.12.12" } }, "nbformat": 4, "nbformat_minor": 5 }