Use dashes in benchmark script filenames

Rename the benchmark scripts to match the dashed naming of the other dev
scripts (e.g. make-release-notes.py):
  benchmark_grounding.py -> benchmark-grounding.py
  benchmark_search.py    -> benchmark-search.py
  compare_benchmarks.py  -> compare-benchmarks.py
and update the cross-references in their docstrings.

Author: jendrikseipp

dev/benchmark_grounding.py dev/benchmark-grounding.pydev/benchmark_grounding.py renamed to dev/benchmark-grounding.py

@@ -4,15 +4,15 @@
 Each task is grounded in a separate subprocess so that a per-task timeout can
 be enforced (grounding a large task can take a long time or a lot of memory).
 The results are written as JSON so that two revisions can be compared with
-``compare_benchmarks.py``.
+``compare-benchmarks.py``.
 
 The ``--src`` option selects which pyperplan source tree to benchmark; it is put
 on ``PYTHONPATH`` for the worker subprocesses, so the script itself can live
 outside the revision under test.
 
 Examples:
-    dev/benchmark_grounding.py --out grounding-new.json
-    dev/benchmark_grounding.py --src /tmp/pyperplan-base --out grounding-base.json
+    dev/benchmark-grounding.py --out grounding-new.json
+    dev/benchmark-grounding.py --src /tmp/pyperplan-base --out grounding-base.json
 """
 
 import argparse

dev/benchmark_search.py dev/benchmark-search.pydev/benchmark_search.py renamed to dev/benchmark-search.py

@@ -4,7 +4,7 @@
 For each configuration and task, ``python -m pyperplan`` is invoked as a
 subprocess with a per-run timeout, and the planner's reported "Overall time" is
 recorded. The results are written as JSON so that two revisions can be compared
-with ``compare_benchmarks.py``.
+with ``compare-benchmarks.py``.
 
 The configurations below cover every search algorithm and every heuristic at
 least once. By default the first task of each benchmark domain is used; pass
@@ -14,8 +14,8 @@
 on ``PYTHONPATH`` for the subprocesses.
 
 Examples:
-    dev/benchmark_search.py --out search-new.json
-    dev/benchmark_search.py --src /tmp/pyperplan-base --out search-base.json
+    dev/benchmark-search.py --out search-new.json
+    dev/benchmark-search.py --src /tmp/pyperplan-base --out search-base.json
 """
 
 import argparse

dev/compare_benchmarks.py dev/compare-benchmarks.pydev/compare_benchmarks.py renamed to dev/compare-benchmarks.py

@@ -1,16 +1,16 @@
 #! /usr/bin/env python3
 """Compare two benchmark result files and plot relative runtimes.
 
-Reads two JSON files produced by ``benchmark_grounding.py`` or
-``benchmark_search.py`` (an "old" and a "new" revision) and, for the tasks that
+Reads two JSON files produced by ``benchmark-grounding.py`` or
+``benchmark-search.py`` (an "old" and a "new" revision) and, for the tasks that
 both revisions completed successfully, prints a runtime summary and writes a
 log-log scatter plot of old vs. new runtime. Points below the diagonal were
 faster on the new revision.
 
 Examples:
-    dev/compare_benchmarks.py grounding-base.json grounding-new.json \
+    dev/compare-benchmarks.py grounding-base.json grounding-new.json \
         --out grounding.png --title "Grounding time"
-    dev/compare_benchmarks.py search-base.json search-new.json \
+    dev/compare-benchmarks.py search-base.json search-new.json \
         --out search.png --title "Search time" --success solved
 """