Tweak authoring benchmarks section

README.md

@@ -44,7 +44,8 @@ repository, and the results are hosted on the
 
 - May 2021: https://ursalabs.org/blog/announcing-conbench/
 
-<hr>
+
+<br>
 
 
 ## Index
@@ -321,7 +322,10 @@ class ExternalBenchmark(conbench.runner.Benchmark):
             "time_unit": "s",
         }
 
-        return self.conbench.external(data, self.name, options=kwargs, output=data)
+        context = {"benchmark_language": "C++"}
+        return self.conbench.external(
+            data, self.name, context=context, options=kwargs, output=data
+        )
 ```
 
 
@@ -334,10 +338,6 @@ $ conbench external --help
       Run external benchmark.
 
     Options:
-      --iterations INTEGER   [default: 1]
-      --drop-caches BOOLEAN  [default: False]
-      --gc-collect BOOLEAN   [default: True]
-      --gc-disable BOOLEAN   [default: True]
       --show-result BOOLEAN  [default: True]
       --show-output BOOLEAN  [default: False]
       --run-id TEXT          Group executions together with a run id.

conbench/runner.py

@@ -118,20 +118,9 @@ def __init__(self):
         self._drop_caches_failed = False
         self._purge_failed = False
 
-    def run(
-        self,
-        f,
-        name,
-        tags=None,
-        context=None,
-        github=None,
-        options=None,
-    ):
-        tags = tags if tags is not None else {}
-        context = context if context is not None else {}
-        options = options if options is not None else {}
-        github = github if github is not None else self.get_github_info()
-
+    def run(self, f, name, **kwargs):
+        """Benchmark a function and publish the result."""
+        tags, context, github, options, _ = self._init(kwargs)
         benchmark, output = self.benchmark(
             f,
             name,
@@ -143,21 +132,9 @@ def run(
         self.publish(benchmark)
         return [(benchmark, output)]
 
-    def external(
-        self,
-        r,
-        name,
-        tags=None,
-        context=None,
-        github=None,
-        options=None,
-        output=None,
-    ):
-        tags = tags if tags is not None else {}
-        context = context if context is not None else {}
-        options = options if options is not None else {}
-        github = github if github is not None else self.get_github_info()
-
+    def external(self, r, name, **kwargs):
+        """Record and publish an external benchmark result."""
+        tags, context, github, options, output = self._init(kwargs)
         benchmark, output = self.record(
             r,
             name,
@@ -170,7 +147,16 @@ def external(
         self.publish(benchmark)
         return [(benchmark, output)]
 
+    def _init(self, kwargs):
+        tags = kwargs.get("tags", {})
+        context = kwargs.get("context", {})
+        github = kwargs.get("github", {})
+        options = kwargs.get("options", {})
+        github = github if github is not None else self.get_github_info()
+        return tags, context, github, options, kwargs.get("output")
+
     def benchmark(self, f, name, tags, context, github, options):
+        """Benchmark a function."""
         timing_options = self._get_timing_options(options)
         iterations = timing_options.pop("iterations")
         if iterations < 1:
@@ -189,6 +175,7 @@ def benchmark(self, f, name, tags, context, github, options):
         return benchmark, output
 
     def record(self, result, name, tags, context, github, options, output=None):
+        """Create a record for an external benchmark result."""
         tags["name"] = name
         timestamp = _now_formatted()
         run_id = options.get("run_id")