Added analysis with OpenCV package

analysis-opencv.yml

@@ -0,0 +1,58 @@
+# Analysis environment with the rather large OpenCV package for image analysis
+
+name: analysis-stable
+channels:
+   - conda-forge
+dependencies:
+# essentials
+   - python=3.13   # keep at 3.13 beause of bitshuffle
+   - pip
+   - numpy>=1.26,<3 # also because of bitshuffle (perhaps not necessary and be deleted in the future)
+   - scipy
+   - pandas
+   - numba
+   - matplotlib
+   - ipython
+   - jupyterlab
+   - ipympl
+   - joblib
+   - tqdm # sfdata dependency but here given explicitly
+   - lmfit
+   - pytest
+   - bitshuffle=0.5.2  # 0.5.2 is compatible with Python 3.12–3.13 on conda-forge;
+                       # Python >=3.14 requires new builds, not a new version (currently not well specified enough to upgrade)
+   
+   - opencv # OpenCV package for more elaborate image analysis
+
+# useful development packages
+   - black
+   - pytest  # for testing the cristallina package
+   - h5py # sfdata dependency but here given explicitly
+   - pint
+   - line_profiler
+   - loguru
+   - pylint
+   - ipytest      # used to be taken from pip, but here we test the conda-forge version
+   - sqlalchemy   # X-ray transmission calculations and similar
+   - xraydb       # X-ray transmission calculations and similar
+   - scikit-image # Image processing
+   - partialjson  # For processing unfinished runs
+# and extra control parts
+   - fabric
+# Jupyterlab extensions
+   - nb_conda_kernels # for discovery of other kernels
+#   - jupyterlab-drawio
+   - jupyterlab_code_formatter
+   # - jupytext # needs to be tested manually
+# PSI specific modules
+   - paulscherrerinstitute::sfdata
+   - paulscherrerinstitute::jungfrau_utils # should be a dependency of sfdata, but not listed there explicitly
+   - paulscherrerinstitute::data_api
+   - pip:
+      - -e /sf/cristallina/applications/cristallina     # Creates a "local package" of cristallina .This shows a deprication warning and
+                                                        # should be changed to using .toml in the future.
+
+   # If wanted, jupyterlab_hdf can be added for nicer tables
+   # pip install jupyterlab_hdf
+   # jupyter labextension install @jupyterlab/hdf5
+