Merge pull request #167 from kbase/develop

Develop -> Master (1.3.3 release)

Author: bio-boris

RELEASE_NOTES.md

@@ -1,3 +1,7 @@
+### Version 1.3.3
+- Fixed a bug in the csv/tsv bulk specification parser that would include an empty entry for
+  each empty line in the file.
+
 ### Version 1.3.2
 - Add `write_bulk_specification` endpoint for writing bulk specifications
 - Add `import_filetypes` endpoint for getting datatype -> filetype -> extension mappings

staging_service/app.py

@@ -34,7 +34,7 @@
 
 logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)
 routes = web.RouteTableDef()
-VERSION = "1.3.2"
+VERSION = "1.3.3"
 
 _DATATYPE_MAPPINGS = None
 

staging_service/import_specifications/individual_parsers.py

@@ -165,16 +165,17 @@ def _parse_xsv(path: Path, sep: str) -> ParseResults:
             _csv_next(rdr, 3, columns, spcsrc, "Missing 3rd header line")
             results = []
             for i, row in enumerate(rdr, start=4):
-                if row and len(row) != columns:  # skip empty rows
-                    # could collect errors (first 10?) and throw an exception with a list
-                    # lets wait and see if that's really needed
-                    raise _ParseException(Error(
-                        ErrorType.INCORRECT_COLUMN_COUNT,
-                        f"Incorrect number of items in line {i}, "
-                        + f"expected {columns}, got {len(row)}",
-                        spcsrc))
-                results.append(frozendict(
-                    {param_ids[j]: _normalize_xsv(row[j]) for j in range(len(row))}))
+                if row:  # skip empty rows
+                    if len(row) != columns:
+                        # could collect errors (first 10?) and throw an exception with a list
+                        # lets wait and see if that's really needed
+                        raise _ParseException(Error(
+                            ErrorType.INCORRECT_COLUMN_COUNT,
+                            f"Incorrect number of items in line {i}, "
+                            + f"expected {columns}, got {len(row)}",
+                            spcsrc))
+                    results.append(frozendict(
+                        {param_ids[j]: _normalize_xsv(row[j]) for j in range(len(row))}))
         if not results:
             raise _ParseException(Error(
                 ErrorType.PARSE_FAIL, "No non-header data in file", spcsrc))

tests/import_specifications/test_individual_parsers.py

@@ -103,7 +103,7 @@ def _xsv_parse_success_with_numeric_headers(
     ))
 
 
-def _xsv_parse_success_with_internal_and_trailing_empty_lines(temp_dir: Path):
+def test_xsv_parse_success_with_internal_and_trailing_empty_lines(temp_dir: Path):
     """
     Test that leaving one or more empty lines in a csv/tsv file does not cause the
     parse to fail. This is easy to do accidentally and so will annoy users.
@@ -135,7 +135,7 @@ def _xsv_parse_success_with_internal_and_trailing_empty_lines(
     assert res == ParseResults(frozendict(
         {"other_type": ParseResult(SpecificationSource(input_),
             tuple([
-                frozendict({"spec1": "val3", "spec4": "val2", "spec3": 1, "spec4": 8.9}),
+                frozendict({"spec1": "val3", "spec2": "val4", "spec3": 1, "spec4": 8.9}),
                 frozendict({"spec1": "val1", "spec2": "val2", "spec3": 7, "spec4": 3.2}),
             ])
         )}