Shorten test IDs and generate histogram over all runs.

.github/workflows/test.yml

@@ -64,9 +64,17 @@ jobs:
       run: |
         compare_args=$(test -d .benchmarks && echo \
           --benchmark-compare \
-          --benchmark-compare-fail=mean:5% \
-          --benchmark-histogram='test-results/benchmark/histogram')
-        pytest test --benchmark-autosave $compare_args      
+          --benchmark-compare-fail=mean:5%) \
+        pytest test --benchmark-autosave $compare_args
+        
+    - name: Create histograms
+      shell: bash {0}
+      run: |
+        if [ -d .benchmarks ]; then
+          py.test-benchmark compare --histogram='test-results/benchmark/histogram'
+        else
+          echo "No previous benchmarks found."
+        fi
 
     - name: Post coverage report to Coveralls
       uses: coverallsapp/github-action@v2

test/test_trnsys_dck_parser/test_expression.py

@@ -17,6 +17,10 @@ class _ExpressionTestCase:
     parser_result: _pexpr.ParseResult
     benchmark: bool = True
 
+    @property
+    def id(self) -> str:
+        return f"{self.string[:3]},,{self.string[-3:]}"
+
 
 def _get_expression_test_cases() -> _tp.Iterable[_ExpressionTestCase]:  # pylint: disable=too-many-statements
     result: _pcom.ParseResult[_mexpr.Expression]
@@ -149,7 +153,7 @@ def _get_expression_test_cases() -> _tp.Iterable[_ExpressionTestCase]:  # pylint
     yield _ExpressionTestCase(string, result, benchmark=False)
 
 
-@_pt.mark.parametrize("test_case", _get_expression_test_cases(), ids=lambda etc: etc.string)
+@_pt.mark.parametrize("test_case", _get_expression_test_cases(), ids=lambda etc: etc.id)
 def test_expression(test_case: _ExpressionTestCase, benchmark) -> None:
     if test_case.benchmark:
         actual_expression = benchmark(_build.parse_expression, test_case.string)