Since 1997 we systematically perform direct angle resolved photoemission spectroscopy (ARPES) on in-situ grown thin ($&lt; 30$ nm) cuprate films. Specifically, we probe low-energy electronic structure and properties of high-$T_{c}$ superconductors (HTSC) under different degrees of epitaxial (compressive vs. tensile) strain. In overdoped and underdoped in-plane compressed (the strain is induced by the choice of substrate) $\simeq 15$ nm thin La$_{2-x}$Sr$_{x}$CuO₄ (LSCO) films we almost double $T_{c}$ to 40 K, from 20 K and 24 K, respectively. Yet the Fermi surface (FS) remains essentially two-dimensional. In contrast, ARPES data under tensile strain exhibit the dispersion that is three-dimensional, yet $T_{c}$ drastically decreases. It seems that the in-plane compressive strain tends to push the apical oxygen far away from the CuO₂ plane, enhances the two-dimensional character of the dispersion and increases $T_{c}$, while the tensile strain acts in the opposite direction and the resulting dispersion is three-dimensional. We have established the shape of the FS for both cases, and all our data are consistent with other ongoing studies, like EXAFS. As the actual lattice of cuprates is like a `Napoleon-cake', i.e. rigid CuO<sub>2</sub> planes alternating with softer `reservoir', that distort differently under strain, our data rule out all oversimplified two-dimensional (rigid lattice) mean field models. The work is still in progress on optimized La-doped Bi-2201 ¯lms with enhanced $T_{c}$.