Merge branch 'vara-dev' into f-FixesPackagingBug

docs/source/conf.py

@@ -87,6 +87,10 @@
 # modules that require this module before setting the type checking flag.
 import scipy.stats  # isort:skip
 
+# Matplotlib >=3.8 has a type-checking-flag-guarded import of a symbol that does
+# not exist in the shipped version.
+import matplotlib.pyplot  # isort:skip
+
 # The autodocs typehints plugin does not resolve circular imports caused by type
 # annotations, so we have to manually break the circles.
 import rich.console  # isort:skip

docs/source/vara-ts-api/tools/vara-cs-gui.rst

@@ -10,6 +10,7 @@ The gui is started by::
 
 The gui provides 3 Strategies to generate case studies:
     - Manual revision selection: Select revision from the revision history of a project. Multiple revisions can be selected by holding `ctrl` and ranges by holding `shift`. Revisions which are blocked because of bugs in the compilation of the project are marked blue.
+
     .. figure:: vara-cs-gui-manual.png
 
     - Random Sampling: Sample a number of revisions using a random a Normal or HalfNormal Distribution.

tests/experiment/test_workload_util.py

@@ -52,6 +52,20 @@ def test_workload_commands_tags_selected(self) -> None:
         )
         self.assertEqual(len(commands), 1)
 
+    def test_workload_commands_requires(self) -> None:
+        revision = Revision(Xz, Variant(Xz.SOURCE[0], "c5c7ceb08a"))
+        project = Xz(revision=revision)
+        binary = Xz.binaries_for_revision(ShortCommitHash("c5c7ceb08a"))[0]
+
+        commands = wu.workload_commands(
+            project, binary, [wu.WorkloadCategory.EXAMPLE]
+        )
+        self.assertEqual(len(commands), 1)
+        commands = wu.workload_commands(
+            project, binary, [wu.WorkloadCategory.MEDIUM]
+        )
+        self.assertEqual(len(commands), 1)
+
 
 class TestWorkloadFilenames(unittest.TestCase):
 

tests/utils/test_git_util.py

@@ -2,7 +2,7 @@
 import unittest
 from pathlib import Path
 
-from benchbuild.utils.revision_ranges import RevisionRange
+from benchbuild.utils.revision_ranges import RevisionRange, SingleRevision
 
 from varats.project.project_util import (
     get_local_project_git,
@@ -568,6 +568,17 @@ def test_specification_validity_range_multiple_binaries(self) -> None:
         self.assertIn("SingleLocalMultipleRegions", self.rv_map)
         self.assertIn("SingleLocalSimple", self.rv_map)
 
+    def test_specification_single_revision(self) -> None:
+        """Check if we can add binaries that are only valid with a single
+        revision."""
+        self.rv_map.specify_binary(
+            "build/bin/SingleLocalMultipleRegions",
+            BinaryType.EXECUTABLE,
+            only_valid_in=SingleRevision("162db88346")
+        )
+
+        self.assertIn("SingleLocalMultipleRegions", self.rv_map)
+
     def test_specification_binaries_with_special_name(self) -> None:
         """Check if we can add binaries that have a special name."""
         self.rv_map.specify_binary(

varats-core/varats/experiment/workload_util.py

@@ -19,6 +19,7 @@
     Command,
 )
 
+from varats.experiment.experiment_util import get_extra_config_options
 from varats.project.project_util import ProjectBinaryWrapper
 from varats.project.varats_project import VProject
 from varats.report.report import KeyedReportAggregate, ReportTy
@@ -92,8 +93,33 @@ def workload_commands(
         )
     ]
 
+    # Filter commands that have required args set.
+    extra_options = set(get_extra_config_options(project))
+
+    def requires_any_filter(prj_cmd: ProjectCommand) -> bool:
+        if hasattr(
+            prj_cmd.command, "requires_any"
+        ) and prj_cmd.command.requires_any:
+            args = set(prj_cmd.command._args).union(extra_options)
+            return bool(args.intersection(prj_cmd.command.requires_any))
+        return True
+
+    def requires_all_filter(prj_cmd: ProjectCommand) -> bool:
+        if hasattr(
+            prj_cmd.command, "requires_all"
+        ) and prj_cmd.command.requires_all:
+            args = set(prj_cmd.command._args).union(extra_options)
+            return bool(prj_cmd.command.requires_all.issubset(args))
+        return True
+
+    available_cmds = filter(
+        requires_all_filter, filter(requires_any_filter, project_cmds)
+    )
+
     return list(
-        filter(lambda prj_cmd: prj_cmd.path.name == binary.name, project_cmds)
+        filter(
+            lambda prj_cmd: prj_cmd.path.name == binary.name, available_cmds
+        )
     )
 
 

varats-core/varats/project/project_util.py

@@ -7,6 +7,7 @@
 
 import benchbuild as bb
 import pygit2
+from benchbuild.command import Command
 from benchbuild.source import Git
 from benchbuild.utils.cmd import git
 from plumbum import local
@@ -382,3 +383,35 @@ def copy_renamed_git_to_dest(src_dir: Path, dest_dir: Path) -> None:
         for name in dirs:
             if name == ".gitted":
                 os.rename(os.path.join(root, name), os.path.join(root, ".git"))
+
+
+class VCommand(Command):  # type: ignore [misc]
+    """
+    Wrapper around benchbuild's Command class.
+
+    Attributes:
+    requires_any: sufficient args that must be available for successful execution.
+    requires_all: all args that must be available for successful execution.
+    """
+
+    _requires: tp.Set[str]
+
+    def __init__(
+        self,
+        *args: tp.Any,
+        requires_any: tp.Optional[tp.Set[str]] = None,
+        requires_all: tp.Optional[tp.Set[str]] = None,
+        **kwargs: tp.Union[str, tp.List[str]],
+    ) -> None:
+
+        super().__init__(*args, **kwargs)
+        self._requires_any = requires_any if requires_any else set()
+        self._requires_all = requires_all if requires_all else set()
+
+    @property
+    def requires_any(self) -> tp.Set[str]:
+        return self._requires_any
+
+    @property
+    def requires_all(self) -> tp.Set[str]:
+        return self._requires_all

varats-core/varats/provider/patch/patch_provider.py

@@ -18,6 +18,7 @@
 
 from varats.project.project_util import get_local_project_git_path
 from varats.provider.provider import Provider, ProviderType
+from varats.utils.filesystem_util import lock_file
 from varats.utils.git_commands import pull_current_branch, fetch_repository
 from varats.utils.git_util import (
     CommitHash,
@@ -237,12 +238,10 @@ class PatchProvider(Provider):
     def __init__(self, project: tp.Type[Project]):
         super().__init__(project)
 
-        # BB only performs a fetch so our repo might be out of date
-        pull_current_branch(self._get_patches_repository_path())
+        self._update_local_patches_repo()
+        repo_path = self._get_patches_repository_path()
 
-        patches_project_dir = Path(
-            self._get_patches_repository_path() / self.project.NAME
-        )
+        patches_project_dir = repo_path / self.project.NAME
 
         if not patches_project_dir.is_dir():
             warnings.warn(
@@ -316,6 +315,12 @@ def create_default_provider(
 
     @classmethod
     def _get_patches_repository_path(cls) -> Path:
-        cls.patches_source.fetch()
-
         return Path(target_prefix()) / cls.patches_source.local
+
+    @classmethod
+    def _update_local_patches_repo(cls):
+        lock_path = Path(target_prefix()) / "patch_provider.lock"
+
+        with lock_file(lock_path):
+            cls.patches_source.fetch()
+            pull_current_branch(cls._get_patches_repository_path())

varats-core/varats/utils/filesystem_util.py

@@ -1,6 +1,8 @@
 """Utility functions for handling filesystem related tasks."""
-
+import fcntl
+import os.path
 import typing as tp
+from contextlib import contextmanager
 from pathlib import Path
 
 
@@ -13,3 +15,15 @@ def __init__(self, folder: tp.Union[Path, str]) -> None:
             f"Folder: '{str(folder)}' should be created "
             "but was already present."
         )
+
+
+@contextmanager
+def lock_file(lock_path: Path, lock_mode: int = fcntl.LOCK_EX) -> tp.Generator:
+    open_mode = os.O_RDWR | os.O_CREAT | os.O_TRUNC
+    lock_fd = os.open(lock_path, open_mode)
+    try:
+        fcntl.flock(lock_fd, lock_mode)
+        yield
+    finally:
+        fcntl.flock(lock_fd, fcntl.LOCK_UN)
+        os.close(lock_fd)

varats-core/varats/utils/git_util.py

@@ -1063,7 +1063,9 @@ def specify_binary(
         override_entry_point = kwargs.get("override_entry_point", None)
         if override_entry_point:
             override_entry_point = Path(override_entry_point)
-        validity_range = kwargs.get("only_valid_in", None)
+        validity_range: AbstractRevisionRange = kwargs.get(
+            "only_valid_in", None
+        )
         valid_exit_codes = kwargs.get("valid_exit_codes", None)
 
         wrapped_binary = ProjectBinaryWrapper(
@@ -1072,6 +1074,7 @@ def specify_binary(
         )
 
         if validity_range:
+            validity_range.init_cache(self.__repo_location)
             self.__revision_specific_mappings[validity_range].append(
                 wrapped_binary
             )
@@ -1087,10 +1090,7 @@ def __getitem__(self,
 
         for validity_range, wrapped_binaries \
                 in self.__revision_specific_mappings.items():
-            if revision in get_all_revisions_between(
-                validity_range.id_start, validity_range.id_end, ShortCommitHash,
-                self.__repo_location
-            ):
+            if revision in map(ShortCommitHash, validity_range):
                 revision_specific_binaries.extend(wrapped_binaries)
 
         revision_specific_binaries.extend(self.__always_valid_mappings)

varats-core/varats/utils/settings.py

@@ -283,7 +283,10 @@ def bb_cfg() -> s.Configuration:
             bb_cfg_path = Path(bb_root) / ".benchbuild.yml"
             if bb_cfg_path.exists():
                 BB_CFG.load(local.path(bb_cfg_path))
-                BB_CFG.init_from_env()
+
+        # Environment should always override config files
+        BB_CFG.init_from_env()
+
         _BB_CFG = BB_CFG
         create_missing_bb_folders()
     return _BB_CFG

varats/varats/data/metrics.py

@@ -28,7 +28,7 @@ def gini_coefficient(distribution: pd.Series) -> float:
     Calculates the Gini coefficient of the data.
 
     For more information see online
-    `gini coefficient <https://en.wikipedia.org/wiki/Gini_coefficient>`_.
+    `Gini coefficient <https://en.wikipedia.org/wiki/Gini_coefficient>`_.
 
     Args:
         distribution: sorted series to calculate the Gini coefficient for
@@ -141,10 +141,12 @@ class ConfusionMatrix(tp.Generic[T]):
     """
     Helper class to automatically calculate classification results.
 
-                        |  Predicted Positive (PP)  |  Predicted Negative (PN)
-    --------------------|---------------------------|--------------------------
-    Actual Positive (P) |  True Positive      (TP)  |  False Negative     (FN)
-    Actual Negative (N) |  False Positive     (FP)  |  True Negative      (TN)
+    +---------------------+-------------------------+-------------------------+
+    |                     | Predicted Positive (PP) | Predicted Negative (PN) |
+    +---------------------+-------------------------+-------------------------+
+    | Actual Positive (P) | True Positive      (TP) | False Negative     (FN) |
+    | Actual Negative (N) | False Positive     (FP) | True Negative      (TN) |
+    +---------------------+-------------------------+-------------------------+
 
     Reference: https://en.wikipedia.org/wiki/Precision_and_recall
     """

varats/varats/projects/c_projects/xz.py

@@ -2,7 +2,7 @@
 import typing as tp
 
 import benchbuild as bb
-from benchbuild.command import Command, SourceRoot, WorkloadSet
+from benchbuild.command import SourceRoot, WorkloadSet
 from benchbuild.source import HTTPMultiple
 from benchbuild.utils.cmd import autoreconf, make
 from benchbuild.utils.revision_ranges import (
@@ -18,6 +18,7 @@
 from varats.paper.paper_config import PaperConfigSpecificGit
 from varats.project.project_domain import ProjectDomains
 from varats.project.project_util import (
+    VCommand,
     ProjectBinaryWrapper,
     get_local_project_git_path,
     BinaryType,
@@ -84,26 +85,33 @@ class Xz(VProject):
 
     WORKLOADS = {
         WorkloadSet(WorkloadCategory.EXAMPLE): [
-            Command(
+            VCommand(
                 SourceRoot("xz") / RSBinary("xz"),
                 "-k",
-                "geo-maps/countries-land-1km.geo.json",
+                # Use output_param to ensure input file
+                # gets appended after all arguments.
+                output_param=["{output}"],
+                output=SourceRoot("geo-maps/countries-land-250m.geo.json"),
                 label="countries-land-1km",
                 creates=["geo-maps/countries-land-1km.geo.json.xz"]
             )
         ],
         WorkloadSet(WorkloadCategory.MEDIUM): [
-            Command(
+            VCommand(
                 SourceRoot("xz") / RSBinary("xz"),
                 "-k",
                 "-9e",
                 "--compress",
                 "--threads=1",
                 "--format=xz",
                 "-vv",
-                "geo-maps/countries-land-250m.geo.json",
+                # Use output_param to ensure input file
+                # gets appended after all arguments.
+                output_param=["{output}"],
+                output=SourceRoot("geo-maps/countries-land-250m.geo.json"),
                 label="countries-land-250m",
-                creates=["geo-maps/countries-land-250m.geo.json.xz"]
+                creates=["geo-maps/countries-land-250m.geo.json.xz"],
+                requires_all={"--compress"},
             )
         ],
     }