mechanoChem: The mechanoChem library developed by our group over years is an open source code consisting of three sub-libraries: These are for (a) finite element (mechanoChemFEM) and (b) isogeometric analysis-based (mechanoChemIGA) direct numerical simulation of partial differential equations in materials physics and biophysics, supported by (c) machine learning algorithms, objects and workflows (mechanoChemML) that are motivated by these PDE-based problems in a number of ways. mechanoChem is in the public domain, available on Github, and follows the LGPL 3.0 license. Special to mechanoChemFEM and mecchanoChemIGA is the ability to couple reaction-transport phenomena, nonlinear elasticity, plasticity, and anelasticity (such as growth phenomena) heat transport, and electrochemistry. In turn, mechanoChemML has machine learning code to enhance the PDE solutions in various ways, or to carry out data-driven modelling with the generated direct numerical simulations. mechanoChemFEM and mechanoChemIGA sub-libraries leverage the deal.ii, FEniCS, Trilinos, PETSc and PETIGA libraries, and use algorithmic differentiation via Sacado for the computation of exact Jacobians. mechanoChemML uses TensorFlow, Keras, PyTorch and scikit-learn, and additionally has Python interfaces with the machine learning and PDE solver platforms.