feat: verify whether the reported repository can be linked back to th…

…e artifact (#873)

Signed-off-by: Mohammad Abdollahpour <[email protected]>

src/macaron/artifact/maven.py

@@ -2,7 +2,7 @@
 # Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/.
 
 """This module declares types and utilities for Maven artifacts."""
-
+import re
 from collections.abc import Sequence
 
 from packageurl import PackageURL
@@ -140,3 +140,21 @@ def create_maven_purl_from_artifact_filename(
             )
 
     return None
+
+
+def is_valid_maven_group_id(group_id: str) -> bool:
+    """Check if the provided string is a valid maven group id.
+
+    Parameters
+    ----------
+    group_id : str
+        The group id to check.
+
+    Returns
+    -------
+    bool
+        True if the group id is valid, False otherwise
+    """
+    # Should match strings like org.example.foo, org.example-2.foo.bar_1.
+    pattern = r"^[a-zA-Z][a-zA-Z0-9-]*\.([a-zA-Z][a-zA-Z0-9-]*\.)*[a-zA-Z][a-zA-Z0-9-]*[a-zA-Z0-9]$"
+    return re.match(pattern, group_id) is not None

src/macaron/repo_finder/repo_finder_deps_dev.py

@@ -5,13 +5,15 @@
 import json
 import logging
 from enum import StrEnum
+from typing import Any
 from urllib.parse import quote as encode
 
 from packageurl import PackageURL
 
 from macaron.json_tools import json_extract
 from macaron.repo_finder.repo_finder_base import BaseRepoFinder
 from macaron.repo_finder.repo_validator import find_valid_repository_url
+from macaron.slsa_analyzer.git_url import clean_url
 from macaron.util import send_get_http_raw
 
 logger: logging.Logger = logging.getLogger(__name__)
@@ -71,6 +73,41 @@ def find_repo(self, purl: PackageURL) -> str:
 
         return ""
 
+    @staticmethod
+    def get_project_info(project_url: str) -> dict[str, Any] | None:
+        """Retrieve project information from deps.dev.
+
+        Parameters
+        ----------
+        project_url : str
+            The URL of the project.
+
+        Returns
+        -------
+        dict[str, Any] | None
+            The project information or None if the information could not be retrieved.
+        """
+        clean_repo_url = clean_url(project_url)
+        if clean_repo_url is None or clean_repo_url.hostname is None:
+            logger.debug("Invalid project url format: %s", project_url)
+            return None
+
+        project_key = clean_repo_url.hostname + clean_repo_url.path
+
+        request_url = f"https://api.deps.dev/v3alpha/projects/{encode(project_key, safe='')}"
+        response = send_get_http_raw(request_url)
+        if not (response and response.text):
+            logger.debug("Failed to retrieve additional repo info for: %s", project_url)
+            return None
+
+        try:
+            response_json: dict = json.loads(response.text)
+        except ValueError as error:
+            logger.debug("Failed to parse response from deps.dev: %s", error)
+            return None
+
+        return response_json
+
     def _create_urls(self, purl: PackageURL) -> list[str]:
         """
         Create the urls to search for the metadata relating to the passed artifact.

src/macaron/repo_verifier/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (c) 2024 - 2024, Oracle and/or its affiliates. All rights reserved.
+# Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/.
+
+"""This package contains classes for repository verification."""

src/macaron/repo_verifier/repo_verifier.py

@@ -0,0 +1,75 @@
+# Copyright (c) 2024 - 2024, Oracle and/or its affiliates. All rights reserved.
+# Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/.
+
+"""This module contains code to verify whether a reported repository can be linked back to the artifact."""
+import logging
+
+from macaron.repo_verifier.repo_verifier_base import (
+    RepositoryVerificationResult,
+    RepositoryVerificationStatus,
+    RepoVerifierBase,
+)
+from macaron.repo_verifier.repo_verifier_gradle import RepoVerifierGradle
+from macaron.repo_verifier.repo_verifier_maven import RepoVerifierMaven
+from macaron.slsa_analyzer.build_tool import BaseBuildTool, Gradle, Maven
+
+logger = logging.getLogger(__name__)
+
+
+def verify_repo(
+    namespace: str | None,
+    name: str,
+    version: str,
+    reported_repo_url: str,
+    reported_repo_fs: str,
+    build_tool: BaseBuildTool,
+) -> RepositoryVerificationResult:
+    """Verify whether the repository links back to the artifact.
+
+    Parameters
+    ----------
+    namespace : str | None
+        The namespace of the artifact.
+    name : str
+        The name of the artifact.
+    version : str
+        The version of the artifact.
+    reported_repo_url : str
+        The reported repository URL.
+    reported_repo_fs : str
+        The reported repository filesystem path.
+    build_tool : BaseBuildTool
+        The build tool used to build the package.
+
+    Returns
+    -------
+    RepositoryVerificationResult
+        The result of the repository verification
+    """
+    # TODO: Add support for other build tools.
+    verifier_map: dict[type[BaseBuildTool], type[RepoVerifierBase]] = {
+        Maven: RepoVerifierMaven,
+        Gradle: RepoVerifierGradle,
+        # Poetry(): RepoVerifierPoetry,
+        # Pip(): RepoVerifierPip,
+        # Docker(): RepoVerifierDocker,
+        # NPM(): RepoVerifierNPM,
+        # Yarn(): RepoVerifierYarn,
+        # Go(): RepoVerifierGo,
+    }
+
+    verifier_cls = verifier_map.get(type(build_tool))
+    if not verifier_cls:
+        return RepositoryVerificationResult(
+            status=RepositoryVerificationStatus.UNKNOWN, reason="unsupported_type", build_tool=build_tool
+        )
+
+    verifier = verifier_cls(
+        namespace=namespace,
+        name=name,
+        version=version,
+        reported_repo_url=reported_repo_url,
+        reported_repo_fs=reported_repo_fs,
+    )
+
+    return verifier.verify_repo()

src/macaron/repo_verifier/repo_verifier_base.py

@@ -0,0 +1,139 @@
+# Copyright (c) 2024 - 2024, Oracle and/or its affiliates. All rights reserved.
+# Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/.
+
+"""This module contains the base class and core data models for repository verification."""
+import abc
+import logging
+import os
+from collections import deque
+from dataclasses import dataclass
+from enum import Enum
+from pathlib import Path
+
+from macaron.slsa_analyzer.build_tool import BaseBuildTool
+
+logger = logging.getLogger(__name__)
+
+
+def find_file_in_repo(root_dir: Path, filename: str) -> Path | None:
+    """Find the highest level file with a given name in a local repository.
+
+    This function ignores certain paths that are not under the main source code directories.
+
+    Parameters
+    ----------
+    root_dir : Path
+        The root directory of the repository.
+    filename : str
+        The name of the file to search for.
+
+    Returns
+    -------
+    Path | None
+        The path to the file if it exists, otherwise
+    """
+    # TODO: Consider using BaseBuildTool.get_build_dirs.
+    #   + Refactor 'get_build_dirs' to skip certain directories
+    #   that are most likely not part of the main codebase (e.g., sample).
+    #   + Need to find a way to look for other
+    #   files (e.g., gradle.properties) for the purpose of repo verification
+    #   without breaking the current logic of finding build directories.
+    #   + Add the capability to return the content/path of the file.
+    if not os.path.isdir(root_dir):
+        return None
+
+    queue: deque[Path] = deque()
+    queue.append(Path(root_dir))
+    while queue:
+        current_dir = queue.popleft()
+
+        # Don't look through non-main directories.
+        if any(
+            keyword in current_dir.name.lower()
+            for keyword in ["test", "example", "sample", "doc", "demo", "spec", "mock"]
+        ):
+            continue
+
+        if Path(current_dir, filename).exists():
+            return Path(current_dir, filename)
+
+        # Ignore symlinks to prevent potential infinite loop.
+        sub_dirs = [Path(it) for it in current_dir.iterdir() if it.is_dir() and not it.is_symlink()]
+        queue.extend(sub_dirs)
+
+    return None
+
+
+class RepositoryVerificationStatus(str, Enum):
+    """A class to store the status of the repo verification."""
+
+    #: We found evidence to prove that the repository can be linked back to the publisher of the artifact.
+    PASSED = "passed"
+
+    #: We found evidence showing that the repository is not the publisher of the artifact.
+    FAILED = "failed"
+
+    #: We could not find any evidence to prove or disprove that the repository can be linked back to the artifact.
+    UNKNOWN = "unknown"
+
+
+@dataclass(frozen=True)
+class RepositoryVerificationResult:
+    """A class to store the information about repository verification."""
+
+    #: The status of the repository verification.
+    status: RepositoryVerificationStatus
+
+    #: The reason for the verification result.
+    reason: str
+
+    #: The build tool used to build the package.
+    build_tool: BaseBuildTool
+
+
+class RepoVerifierBase(abc.ABC):
+    """The base class to verify whether a reported repository links back to the artifact."""
+
+    @property
+    @abc.abstractmethod
+    def build_tool(self) -> BaseBuildTool:
+        """Define the build tool used to build the package."""
+
+    def __init__(
+        self,
+        namespace: str | None,
+        name: str,
+        version: str,
+        reported_repo_url: str,
+        reported_repo_fs: str,
+    ):
+        """Instantiate the class.
+
+        Parameters
+        ----------
+        namespace : str
+            The namespace of the artifact.
+        name : str
+            The name of the artifact.
+        version : str
+            The version of the artifact.
+        reported_repo_url : str
+            The URL of the repository reported by the publisher.
+        reported_repo_fs : str
+            The file system path of the reported repository.
+        """
+        self.namespace = namespace
+        self.name = name
+        self.version = version
+        self.reported_repo_url = reported_repo_url
+        self.reported_repo_fs = reported_repo_fs
+
+    @abc.abstractmethod
+    def verify_repo(self) -> RepositoryVerificationResult:
+        """Verify whether the repository links back to the artifact.
+
+        Returns
+        -------
+        RepositoryVerificationResult
+            The result of the repository verification
+        """