Finding similar weather patterns or analogs in historical datasets can be useful for data assimilation, forecasting, postprocessing and for attribution and impact studies. However, searching in those datasets is computationally expensive. A distance-regularized Siamese twin autoencoder (DIRESA) architecture is designed to preserve distance in latent space while capturing the nonlinearities in the datasets. The search can then be executed in latent space and the latent components thus obtained provide physical insight into the dominant modes of variability in the system. DIRESA is provided as an open-source Python package based on Tensorflow, using a CNN/Dense or custom encoder and decoder.