Fix --expand mode for empty attributes

If an attribute appears in sample_attribute but has no value
associated with it, it is as good as non-existent, we
discard this attribute. This only effects it on a record level
and not the entire run table. If any other sample in the same
run table has that attribute, the DataFrame will still have
all the relevant columns with the record with missing value
for an attribute as an NaN.

Closes #11

pysradb/filter_attrs.py

@@ -8,19 +8,31 @@ def _get_sample_attr_keys(sample_attribute):
     if sample_attribute is None:
         return None, None
     sample_attribute_splitted = sample_attribute.split("||")
-    split_by_colon = [str(attr).split(": ") for attr in sample_attribute_splitted]
+    split_by_colon = [
+        str(attr).strip().split(": ") for attr in sample_attribute_splitted
+    ]
+
     # Iterate once more to consider first one as the key
     # and remaining as the value
     # This is because of bad annotations like in this example
     # Example: isolate: not applicable || organism: Mus musculus || cell_line: 17-Cl1 ||\
     # infect: MHV-A59 || time point: 5: hour || compound: cycloheximide ||\
     # sequencing protocol: RiboSeq || biological repeat: long read sequencing
     # Notice the `time: 5: hour`
+    # sample_attribute: investigation type: metagenome || project name: Landsort Depth 20090415 transect ||
+    # sequencing method: 454 || collection date: 2009-04-15 || ammonium: 8.7: µM || chlorophyll: 0: µg/L ||
+    # dissolved oxygen: -1.33: µmol/kg || nitrate: 0.02: µM || nitrogen: 0: µM ||
+    # environmental package: water || geographic location (latitude): 58.6: DD ||
+    # geographic location (longitude): 18.2: DD || geographic location (country and/or sea,region): Baltic Sea ||
+    # environment (biome): 00002150 || environment (feature): 00002150 || environment (material): 00002150 ||
+    # depth: 400: m || Phosphate:  || Total phosphorous:  || Silicon:
+    # Handle empty cases as above
+    split_by_colon = [attr for attr in split_by_colon if len(attr) >= 2]
 
     for index, element in enumerate(split_by_colon):
         if len(element) > 2:
-            key = element[0]
-            value = ":".join(element[1:])
+            key = element[0].strip()
+            value = ":".join(element[1:]).strip()
             split_by_colon[index] = [key, value]
 
     try: