[kdiers]

This pull request refactors and improves the quality control (QC) image generation in the pipeline, replacing the previous Plotly/Kaleido-based approach with a new implementation using Matplotlib and PyVista. It also updates Docker and Singularity documentation and dependencies to reflect these changes, and adds support for new QC stages. The most important changes are grouped below:

QC image generation refactor and enhancements:

• Refactored qcPlots in hipsta/utils/qc_plots.py to use Matplotlib for 2D profile plots and PyVista for 3D mesh screenshots, replacing Plotly/Kaleido and the old lapy.plot backend. Added a new _renderTriaMesh helper for rendering meshes, and introduced support for a new "mid-surface" QC stage. Improved docstrings and code structure throughout.
• Updated the main pipeline in hipsta/hipsta.py to call qcPlots for the new "mid-surface" stage.

Docker and dependency updates:

• Removed installation of Chrome/Chromium and related dependencies from the Docker image, as they are no longer required for QC rendering; added libosmesa6 for offscreen rendering with PyVista.
• Updated Docker documentation to remove references to Chrome/Kaleido for QC rendering, reflecting the new backend.

Singularity and documentation improvements:

• Updated Singularity documentation to remove the --no-qc recommendation, as QC rendering is now supported in headless environments. Also added --no-mount home,cwd -e to recommended invocation for improved reproducibility.