{\sc precision}: A fast python pipeline for high-contrast imaging -- application to SPHERE observations of the red supergiant VX Sagitariae
The search for extrasolar planets has driven rapid advances in instrumentation, resulting in cameras such as SPHERE at the VLT, GPI at Gemini South and SCExAO at Subaru, capable of achieving very high contrast ($\sim10^{6}$) around bright stars with small inner working angles ($\sim 0\farcs{1}$). The optimal exploitation of data from these instruments depends on the availability of easy-to-use software to process and analyse their data products. We present a pure-python pipeline, {\sc precision}, which provides fast, memory-efficient reduction of data from the SPHERE/IRDIS near-infrared imager, and can be readily extended to other instruments. We apply {\sc precision} to observations of the extreme red supergiant VX~Sgr, the inner outflow of which is revealed to host complex, asymmetric structure in the near-IR. In addition, optical polarimetric imaging reveals clear extended polarised emission on $\sim0.5^{\prime\prime}$ scales which varies significantly with azimuth, confirming the asymmetry. While not conclusive, this could suggest that the ejecta are confined to a disc or torus, which we are viewing nearly face on, although other non-spherical or clumpy configurations remain possible. VX~Sgr has no known companions, making such a geometry difficult to explain, as there is no obvious source of angular momentum in the system.