Add more bugginess heuristics

heuristics/bugginess_files.py

@@ -0,0 +1,237 @@
+import re
+from typing import Optional
+
+from bohr.collection.artifacts import Commit
+from bohr.core import Heuristic
+from bohr.labeling.labelset import Labels
+from labels import CommitLabel
+
+
+@Heuristic(Commit)
+def no_files_have_modified_status(commit: Commit) -> Optional[Labels]:
+    if len(commit.commit_files) == 0 or commit.commit_files[0].status == "empty":
+        return None
+    for file in commit.commit_files:
+        if file.status == "modified":
+            return None
+    return CommitLabel.NonBugFix
+
+
+@Heuristic(Commit)
+def bug_if_only_changed_lines_in_one_code_file(commit: Commit) -> Optional[Labels]:
+    if (
+        len(commit.commit_files) == 1
+        and commit.commit_files[0].status == "modified"
+        and not isinstance(commit.commit_files[0].filename, float)
+        and commit.commit_files[0].filename.split(".")[-1]
+        in [*code_extensions, *passive_code_extensions]
+        and commit.commit_files[0].changes
+        and commit.commit_files[0].no_added_lines()
+        and commit.commit_files[0].no_removed_lines()
+    ):
+        return CommitLabel.BugFix
+    return None
+
+
+@Heuristic(Commit)
+def bugless_if_at_least_5_added_files(commit: Commit) -> Optional[Labels]:
+    added_count = 0
+    for file in commit.commit_files:
+        if file.status == "added":
+            added_count += 1
+    return CommitLabel.NonBugFix if added_count >= 5 else None
+
+
+@Heuristic(Commit)
+def bugless_if_one_added_file(commit: Commit) -> Optional[Labels]:
+    if len(commit.commit_files) == 1 and commit.commit_files[0].status == "added":
+        return CommitLabel.NonBugFix
+    return None
+
+
+@Heuristic(Commit)
+def bugless_if_at_least_2_removed_files(commit: Commit) -> Optional[Labels]:
+    removed_count = 0
+    for file in commit.commit_files:
+        if file.status == "removed":
+            removed_count += 1
+    return CommitLabel.NonBugFix if removed_count >= 2 else None
+
+
+@Heuristic(Commit)
+def bugless_if_one_removed_file(commit: Commit) -> Optional[Labels]:
+    if len(commit.commit_files) == 1 and commit.commit_files[0].status == "removed":
+        return CommitLabel.NonBugFix
+    return None
+
+
+@Heuristic(Commit)
+def refactoring_if_at_least_2_renamed(commit: Commit) -> Optional[Labels]:
+    renamed_count = 0
+    for file in commit.commit_files:
+        if file.status == "renamed":
+            renamed_count += 1
+    return CommitLabel.Refactoring if renamed_count >= 2 else None
+
+
+code_extensions = [
+    "js",
+    "java",
+    "py",
+    "php",
+    "cpp",
+    "h",
+    "rb",
+    "ts",
+    "c",
+    "go",
+    "css",
+    "cs",
+    "scss",
+    "jsx",
+    "m",
+    "less",
+    "sh",
+    "scala",
+    "cc",
+    "coffee",
+    "F90",
+    "hpp",
+    "inc",
+    "sql",
+    "erb",
+    "tsx",
+    "kt",
+    "Makefile",
+    "groovy",
+    "hbs",
+    "swift",
+    "hh",
+    "twig",
+    "haml",
+    "hs",
+    "scssc",
+]
+ignore_extensions = ["gitignore"]
+config_extensions = [
+    "yaml",
+    "yml",
+    "gradle",
+    "in",
+    "properties",
+    "conf",
+    "csproj",
+    "ini",
+    "config",
+]
+passive_code_extensions = [
+    *ignore_extensions,
+    *config_extensions,
+    "xml",
+    "html",
+    "json",
+    "htm",
+]
+binary_extensions = [
+    "png",
+    "po",
+    "jpg",
+    "gif",
+    "tgz",
+    "gz",
+    "jar",
+    "mo",
+    "slj",
+    "class",
+    "gem",
+    "map",
+    "pdf",
+    "ttf",
+    "aw",
+]
+generated_text_extensions = ["out"]
+doc_text_extensions = [
+    "markdown",
+    "md",
+    "rst",
+    "adoc",
+]
+non_code_extensions = [
+    *binary_extensions,
+    *generated_text_extensions,
+    *doc_text_extensions,
+    "txt",
+    "svg",
+    "lock",
+    "LICENSE",
+    "csv",
+]
+
+
+@Heuristic(Commit)
+def bugless_if_not_code_files(commit: Commit) -> Optional[Labels]:
+    file_found = False
+    for file in commit.commit_files:
+        if isinstance(file.filename, float):
+            continue  # TODO filename is NaN if it's not given <- needs to be fixed
+        if file.filename.split(".")[-1] in [*code_extensions, *passive_code_extensions]:
+            # TODO move filetype logic to CommitFile artifact?
+            return None
+        else:
+            file_found = True
+    return CommitLabel.NonBugFix if file_found else None
+
+
+@Heuristic(Commit)
+def buggless_if_many_lines_changed(commit: Commit) -> Optional[Labels]:
+    sum = 0
+    for file in commit.commit_files:
+        if file.changes:
+            sum += len(file.changes)
+    return CommitLabel.NonBugFix if sum > 5000 else None
+
+
+@Heuristic(Commit)
+def bugless_if_many_files_changes(commit: Commit) -> Optional[Labels]:
+    if len(commit.commit_files) > 15:
+        return CommitLabel.NonBugFix
+    else:
+        return None
+
+
+@Heuristic(Commit)
+def all_files_test_add(commit: Commit) -> Optional[Labels]:
+    TEST_FILE_REGEX = re.compile(r"test", flags=re.I)
+    if len(commit.commit_files) == 0:
+        return None
+
+    for file in commit.commit_files:
+
+        def only_additions():
+            return not (
+                not file.changes or "<re>" in file.changes or "<del>" in file.changes
+            )
+
+        if not (
+            TEST_FILE_REGEX.match(str(file.filename))
+            and (file.status == "added" or only_additions())
+        ):
+            return None
+    return CommitLabel.TestAdd
+
+
+@Heuristic(Commit)
+def all_files_test_fix(commit: Commit) -> Optional[Labels]:
+    TEST_FILE_REGEX = re.compile(r"test", flags=re.I)
+    if len(commit.commit_files) == 0:
+        return None
+
+    for file in commit.commit_files:
+        if not (
+            TEST_FILE_REGEX.match(str(file.filename))
+            and file.status == "modified"
+            and file.changes
+            and "<re>" in file.changes
+        ):
+            return None
+    return CommitLabel.TestFix

heuristics/commit_explorer.py

@@ -0,0 +1,67 @@
+from typing import Optional
+
+from commitexplorer.client import CommitExplorerClientException, CommitNotFoundException
+
+import labels as l
+from bohr.collection.artifacts import Commit
+from bohr.core import Heuristic
+from bohr.labeling.labelset import Labels
+
+
+@Heuristic(Commit)
+def commit_explorer_output_merge(commit: Commit) -> Optional[Labels]:
+    try:
+        if commit.commit_explorer_data is None:
+            return None
+
+        if (
+            "special_commit_finder/0_1" in commit.commit_explorer_data
+            and commit.commit_explorer_data["special_commit_finder/0_1"]["merge"]
+        ):
+            return l.CommitLabel.Merge
+    except (CommitExplorerClientException, CommitNotFoundException) as ex:
+        return None
+
+
+@Heuristic(Commit)
+def commit_explorer_output_init(commit: Commit) -> Optional[Labels]:
+    try:
+        if commit.commit_explorer_data is None:
+            return None
+
+        if (
+            "special_commit_finder/0_1" in commit.commit_explorer_data
+            and commit.commit_explorer_data["special_commit_finder/0_1"]["initial"]
+        ):
+            return l.CommitLabel.InitialCommit
+    except (CommitExplorerClientException, CommitNotFoundException) as ex:
+        return None
+
+
+# @Heuristic(Commit)
+# def commit_explorer_output_refactoring_miner(commit: Commit) -> Optional[Labels]:
+#     if commit.commit_explorer_data is None:
+#         return None
+#
+#     data = commit.commit_explorer_data
+#     if "refactoring_miner/2_1_0" in data:
+#         if data["refactoring_miner/2_1_0"]:
+#             if data["refactoring_miner/2_1_0"]['status'] == 'ok':
+#                 if len(data["refactoring_miner/2_1_0"]["refactorings"]) == 1:
+#                     if data["refactoring_miner/2_1_0"]["refactorings"][0]["type"] in ['Move Class']:
+#                         print("Contains refactoring")
+#                         return l.CommitLabel.Refactoring
+
+
+@Heuristic(Commit)
+def commit_explorer_output_sstubs(commit: Commit) -> Optional[Labels]:
+    try:
+        if commit.commit_explorer_data is None:
+            return None
+
+        data = commit.commit_explorer_data
+        if "mine_sstubs/head" in data and data["mine_sstubs/head"]:
+            print("Contains bug")
+            return l.CommitLabel.BugFix
+    except (CommitExplorerClientException, CommitNotFoundException) as ex:
+        return None

heuristics/conventional_commits.py

@@ -0,0 +1,77 @@
+import re
+from typing import Optional
+
+import labels as l
+from bohr.collection.artifacts import Commit
+from bohr.core import Heuristic
+from bohr.labeling.labelset import Labels
+
+example1 = """feat: allow provided config object to extend other configs
+
+BREAKING CHANGE: `extends` key in config file is now used for extending other config files
+"""
+
+example2 = """refactor!: drop support for Node 6
+"""
+
+example3 = """refactor!: drop support for Node 6
+
+BREAKING CHANGE: refactor to use JavaScript features not available in Node 6.
+"""
+
+example4 = """docs: correct spelling of CHANGELOG
+"""
+
+example5 = """feat(lang): add polish language
+"""
+
+example6 = """fix: correct minor typos in code
+
+see the issue for details
+
+on typos fixed.
+
+Reviewed-by: Z
+Refs #133
+"""
+
+REGEX = re.compile(
+    r"\A(((Initial commit)|(Merge [^\r\n]+)|"
+    r"((build|chore|ci|docs|feat|fix|perf|refactor|revert|style|test)(\(\w+\))?!?: [^\r\n]+"
+    r"((\r|\n|\r\n)((\r|\n|\r\n)[^\r\n]+)+)*"
+    r")"
+    r")(\r|\n|\r\n)?)"
+)
+
+
+@Heuristic(Commit)
+def conventional_commit_regex(commit: Commit) -> Optional[Labels]:
+    """
+    >>> conventional_commit_regex(Commit('a', 'a', '1df23', example1))
+    CommitLabel.Feature
+    >>> conventional_commit_regex(Commit('a', 'a', '1df23', example2))
+    CommitLabel.Refactoring
+    >>> conventional_commit_regex(Commit('a', 'a', '1df23', example3))
+    CommitLabel.Refactoring
+    >>> conventional_commit_regex(Commit('a', 'a', '1df23', example4))
+    CommitLabel.DocChange
+    >>> conventional_commit_regex(Commit('a', 'a', '1df23', example5))
+    CommitLabel.Feature
+    >>> conventional_commit_regex(Commit('a', 'a', '1df23', example6))
+    CommitLabel.BugFix
+    """
+    m = REGEX.match(commit.message.raw)
+    if m is None:
+        return None
+    type = m.groups()[5]
+    if type == "fix":
+        return l.CommitLabel.BugFix
+    elif type == "feat":
+        return l.CommitLabel.Feature
+    elif type == "refactor":
+        return l.CommitLabel.Refactoring
+    elif type == "docs":
+        return l.CommitLabel.DocChange
+    else:
+        return None
+    # TODO add more types