Merge pull request #47 from AlexandrovLab/u_input_handling

U input handling

setup.py

@@ -14,7 +14,7 @@ def readme():
         return f.read()
 
 
-VERSION = "1.3.21"
+VERSION = "1.3.22"
 
 
 def write_version_py(filename="sigProfilerPlotting/version.py"):
@@ -23,7 +23,7 @@ def write_version_py(filename="sigProfilerPlotting/version.py"):
 # THIS FILE IS GENERATED FROM SIGPROFILERPLOTTING SETUP.PY
 short_version = '%(version)s'
 version = '%(version)s'
-update = 'Upgrade v1.3.21: Add CLI and container updates'
+update = 'Upgrade v1.3.22: Update input_processing index.name handling to be more robust.'
     
     """
     fh = open(filename, "w")

sigProfilerPlotting/sigProfilerPlotting.py

@@ -39,6 +39,7 @@
 matplotlib.use("Agg")
 
 MUTTYPE = "MutationType"
+INDEX_VALS = ["MutationType", "index", "Mutation Types", "classification"]
 SPP_PATH = spplt.__path__[0]
 SPP_TEMPLATES = os.path.join(SPP_PATH, "templates/")
 SPP_FONTS = os.path.join(SPP_PATH, "fonts/")
@@ -168,17 +169,26 @@ def process_input(matrix_path, plot_type):
     if isinstance(matrix_path, pd.DataFrame):
         # copy dataframe with deepcopy
         data = matrix_path.copy()
-        # Index is not MutationType
-        if MUTTYPE != data.index.name:
+        # index is a non-standard value
+        if data.index.name not in INDEX_VALS:
             if MUTTYPE in data.columns:
                 data = data.set_index(MUTTYPE, drop=True)
+            # the first column is non-MUTTYPE and non-integer
+            elif not data.iloc[:, 0].apply(lambda x: isinstance(x, int)).all():
+                    data.rename(columns={data.columns[0]: MUTTYPE}, inplace=True)
+                    data = data.set_index(data.columns[0], drop=True)
             else:
+                data = data.reset_index()
                 data.rename(columns={data.columns[0]: MUTTYPE}, inplace=True)
                 data = data.set_index(MUTTYPE, drop=True)
-    # input data is a path to a file
+        else:
+            # Note: set the index to MUTTYPE for consistency with the rest of the code
+            data.index.name = MUTTYPE
+    # input data is a file path
     elif isinstance(matrix_path, str):
         data = pd.read_csv(matrix_path, sep="\t", index_col=0)
         data = data.dropna(axis=1, how="all")
+        data.index.name = MUTTYPE
     # input data is a numpy array
     elif isinstance(matrix_path, np.ndarray):
         # Note: ndarray does not have index or column names and is not recommended

sigProfilerPlotting/version.py

@@ -1,7 +1,7 @@
 
 # THIS FILE IS GENERATED FROM SIGPROFILERPLOTTING SETUP.PY
-short_version = '1.3.20'
-version = '1.3.20'
-update = 'Upgrade v1.3.20: Add np.ndarray to process_input'
+short_version = '1.3.22'
+version = '1.3.22'
+update = 'Upgrade v1.3.22: Update input_processing index.name handling to be more robust.'
 
 

tests/test_plotting.py

@@ -22,7 +22,7 @@ def image_difference(img1_path, img2_path):
         diff = ImageChops.difference(img1, img2)
         total_difference = sum(abs(p) for p in diff.getdata())
         max_difference = img1.size[0] * img1.size[1] * 255
-        if total_difference > 1e-4:
+        if (total_difference / max_difference) > 1e-4:
             diff.show()
         return total_difference / max_difference