Merge pull request #6 from biowdl/extrapropsoverride

Add extra feature that makes it easier to add properties for samples in csv format.

CHANGELOG.rst

@@ -9,6 +9,8 @@ Changelog
 
 0.2.0-dev
 ---------------
++ Make sure only one line of additional properties per sample is need in a
+  csv file.
 + Fix a bug where an empty field for an additional property in a csv
   samplesheet would be defined as ``""`` instead of ``None``.
 

docs/index.rst

@@ -80,6 +80,25 @@ Additional properties at the sample level can be set using additional columns:
     "s1","lib1","rg1","r1_1.fq",,"r1_2.fq",,"yes","pizza"
     "s2","lib1","rg1","r2_1.fq",,"r2_2.fq",,"no","broccoli"
 
+Additional properties for the same sample only have to be defined in one line.
+This saves a lot of duplication for samples with a high readgroup or library
+count an makes it easier to read the file.
+
+.. code-block:: text
+
+    "sample","library","readgroup","R1","R1_md5","R2","R2_md5","HiSeq4000","other_property"
+    "s1","lib1","rg1","r1_1.fq",,"r1_2.fq",,"yes","pizza"
+    "s1","lib1","rg2","r1_1.fq",,"r1_2.fq",,,
+    "s1","lib2","rg1","r1_1.fq",,"r1_2.fq",,,
+    "s2","lib1","rg1","r2_1.fq",,"r2_2.fq",,"no","broccoli"
+    "s2","lib1","rg2","r2_1.fq",,"r2_2.fq",,,
+    "s2","lib1","rg3","r2_1.fq",,"r2_2.fq",,,
+
+If an additional column is filled with two conflicting values for the same
+sample an error will be thrown.
+
+Creating comma-delimited files
+------------------------------
 These files can be easily generated using a spreadsheet program (such as 
 Microsoft Excel or LibreOffice Calc). 
 

src/biowdl_input_converter/input_conversions.py

@@ -102,8 +102,22 @@ def samplesheet_csv_to_samplegroup(samplesheet_file: Path) -> SampleGroup:
         }
         # Add all remaining properties to additional properties at the
         # sample level
-        samples[sample]["additional_properties"] = {
-            key: value if value != "" else None
-            for key, value in row_dict.items()
-        }
+        if "additional_properties" not in samples[sample].keys():
+            samples[sample]["additional_properties"] = {}
+        for key, value in row_dict.items():
+            existing_value = samples[sample][
+                "additional_properties"].get(key, None)
+
+            updated_value = value if value != "" else None
+
+            if existing_value is None:
+                samples[sample]["additional_properties"][key] = updated_value
+            else:
+                if (updated_value is not None
+                        and existing_value != updated_value):
+                    raise ValueError(
+                        f"Conflicting fields in column '{key}' for sample "
+                        f"'{sample}'!"
+                    )
+
     return SampleGroup.from_dict_of_dicts(samples)

tests/files/conflicting_properties.csv

@@ -0,0 +1,3 @@
+"sample","library","readgroup","R1","R1_md5","R2","R2_md5","extra_field1"
+"s1","lib1","rg1","r1.fq","hello","r2.fq","hey","xf1"
+"s1","lib2","rg1","r1.fq","aa","r2.fq","bb","xfI"

tests/files/mixed_empty_filled_addprops.csv

@@ -0,0 +1,3 @@
+"sample","library","readgroup","R1","R1_md5","R2","R2_md5","extra_field1","extra_field2"
+"s1","lib1","rg1","r1.fq","hello","r2.fq","hey","xf1","xf2"
+"s1","lib2","rg1","r1.fq","aa","r2.fq","bb","xf1",

tests/test_import_samplesheet_csv.py

@@ -70,6 +70,21 @@ def test_extra_field():
     assert samplesheet[1].additional_properties["extra_field2"] is None
 
 
+def test_mixed_empty_and_filled_additional_properties():
+    samplesheet = samplesheet_csv_to_samplegroup(
+        FILESDIR / Path("mixed_empty_filled_addprops.csv"))
+    assert len(samplesheet.samples) == 1
+    assert samplesheet[0].additional_properties["extra_field2"] == "xf2"
+
+
+def test_conflicting_properties():
+    with pytest.raises(ValueError) as error:
+        samplesheet_csv_to_samplegroup(
+            FILESDIR / Path("conflicting_properties.csv")
+        )
+    error.match("Conflicting fields in column 'extra_field1' for sample 's1'")
+
+
 def test_duplicate_readgroup():
     with pytest.raises(ValueError) as error:
         samplesheet_csv_to_samplegroup(