Return NaN instead of crashing when every EMD stratum is too small (#131)

* return nan instead of crashing when every emd stratum is too small

* Guard the pd.concat() calls in both the horizontal and vertical path
* Warn as soon as one cell type is missing from a split, not two
* Fix the newline in that warning

* update changelog

* simplify changes

* Update CHANGELOG.md

---------

Co-authored-by: Givanna Putri <givanna.h@gmail.com>
Co-authored-by: Luca Leomazzi <76624347+LuLeom@users.noreply.github.com>

Author: 3 people

CHANGELOG.md

@@ -252,3 +252,6 @@
 
 * Fix bug in `average_batch_r2` where the R2 was computed on all cell types of a donor at once
   instead of on each cell type separately (PR #127).
+
+
+* Fix check cell_type_in_both to > 0 (instead of 1) and some spelling mistakes in print.

src/metrics/emd/helper.py

@@ -224,11 +224,11 @@ def calculate_horizontal_emd(
             i_split1_donor.obs["cell_type"].unique(),
             i_split2_donor.obs["cell_type"].unique(),
         )
-        if len(cell_type_not_in_both) > 1:
+        if len(cell_type_not_in_both) > 0:
             print(
                 f"In donor {donor}: some cell types are in split 1"
                 f" but not in split 2.\n"
-                f"Cell types missing: {''.join(cell_type_not_in_both)}]n"
+                f"Cell types missing: {', '.join(cell_type_not_in_both)}\n"
                 f"Computing cell type EMD using just cell types common in both."
             )