Add org.osbuild.dnf4.sbom.spdx stage

Add a new stage, which allows analyzing the installed packages in a
given filesystem tree using DNF4 API and generating an SPDX v2.3 SBOM
document for it.

One can provide the filesystem tree to be analyzed as a stage input. If
no input is provided, the stage will analyze the filesystem tree of the
current pipeline.

Add tests cases for both usage variants of the stage, as well as the
unit test for stage schema validation.

Signed-off-by: Tomáš Hozza <[email protected]>

stages/org.osbuild.dnf4.sbom.spdx

@@ -0,0 +1,52 @@
+#!/usr/bin/python3
+import json
+import sys
+import tempfile
+
+import dnf
+
+import osbuild
+from osbuild.util.bom.dnf import dnf_pkgset_to_bom_pkgset
+from osbuild.util.bom.spdx import bom_pkgset_to_spdx2_doc
+
+
+def get_installed_packages(tree):
+    with tempfile.TemporaryDirectory() as tempdir:
+        conf = dnf.conf.Conf()
+        conf.installroot = tree
+        conf.persistdir = f"{tempdir}{conf.persistdir}"
+        conf.cachedir = f"{tempdir}{conf.cachedir}"
+        conf.reposdir = [f"{tree}{d}" for d in conf.reposdir]
+        conf.pluginconfpath = [f"{tree}{d}" for d in conf.pluginconfpath]
+        conf.varsdir = [f"{tree}{d}" for d in conf.varsdir]
+        conf.prepend_installroot("config_file_path")
+
+        base = dnf.Base(conf)
+        base.read_all_repos()
+        base.fill_sack(load_available_repos=False)
+        return base.sack.query().installed()
+
+
+def main(inputs, tree, options):
+    config = options["config"]
+    doc_path = config["doc_path"]
+
+    tree_to_analyze = tree
+    if inputs:
+        tree_to_analyze = inputs["root-tree"]["path"]
+
+    installed = get_installed_packages(tree_to_analyze)
+    bom_pkgset = dnf_pkgset_to_bom_pkgset(installed)
+    spdx2_doc = bom_pkgset_to_spdx2_doc(bom_pkgset)
+    spdx2_json = spdx2_doc.to_dict()
+
+    with open(f"{tree}{doc_path}", "w", encoding="utf-8") as f:
+        json.dump(spdx2_json, f)
+
+    return 0
+
+
+if __name__ == '__main__':
+    args = osbuild.api.arguments()
+    r = main(args.get("inputs", {}), args["tree"], args["options"])
+    sys.exit(r)

stages/org.osbuild.dnf4.sbom.spdx.meta.json

@@ -0,0 +1,59 @@
+{
+  "summary": "Generate SPDX SBOM document for the installed packages.",
+  "description": [
+    "The stage generates a Software Bill of Materials (SBOM) document",
+    "in SPDX v2 format for the installed RPM packages. DNF4 API is used",
+    "to retrieve the installed packages and their metadata. The SBOM",
+    "document is saved in the specified path. If a tree is provided,",
+    "as an input, the stage will analyze the tree instead of the",
+    "current pipeline tree."
+  ],
+  "schema_2": {
+    "options": {
+      "additionalProperties": false,
+      "description": "Options for the SPDX SBOM generator.",
+      "required": [
+        "config"
+      ],
+      "properties": {
+        "config": {
+          "type": "object",
+          "description": "Configuration for the SPDX SBOM generator.",
+          "additionalProperties": false,
+          "required": [
+            "doc_path"
+          ],
+          "properties": {
+            "doc_path": {
+              "type": "string",
+              "pattern": "^\\/(?!\\.\\.)((?!\\/\\.\\.\\/).)+[\\w]{1,250}\\.spdx.json$",
+              "description": "Path used to save the SPDX SBOM document."
+            }
+          }
+        }
+      }
+    },
+    "inputs": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "root-tree"
+      ],
+      "properties": {
+        "root-tree": {
+          "type": "object",
+          "additionalProperties": true,
+          "description": "The tree containing the installed packages. If the input is not provided, the stage will analyze the tree of the current pipeline.",
+          "properties": {
+            "type": {
+              "type": "string",
+              "enum": [
+                "org.osbuild.tree"
+              ]
+            }
+          }
+        }
+      }
+    }
+  }
+}

stages/test/test_dnf4_sbom_spdx.py

@@ -0,0 +1,130 @@
+#!/usr/bin/python3
+
+import pytest
+
+import osbuild.testutil as testutil
+
+STAGE_NAME = "org.osbuild.dnf4.sbom.spdx"
+
+
+@pytest.mark.parametrize("test_data,expected_err", [
+    # good
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "/image.spdx.json",
+                }
+            }
+        },
+        "",
+    ),
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "/root/doc.spdx.json",
+                }
+            }
+        },
+        "",
+    ),
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "/image.spdx.json",
+                }
+            },
+            "inputs": {
+                "root-tree": {
+                    "type": "org.osbuild.tree",
+                    "origin": "org.osbuild.pipeline",
+                    "references": [
+                        "name:root-tree"
+                    ]
+                }
+            }
+        },
+        "",
+    ),
+    # bad
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "/image.spdx",
+                }
+            }
+        },
+        "'/image.spdx' does not match '^\\\\/(?!\\\\.\\\\.)((?!\\\\/\\\\.\\\\.\\\\/).)+[\\\\w]{1,250}\\\\.spdx.json$'",
+    ),
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "/image.json",
+                }
+            }
+        },
+        "'/image.json' does not match '^\\\\/(?!\\\\.\\\\.)((?!\\\\/\\\\.\\\\.\\\\/).)+[\\\\w]{1,250}\\\\.spdx.json$'",
+    ),
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "image.spdx.json",
+                }
+            }
+        },
+        "'image.spdx.json' does not match '^\\\\/(?!\\\\.\\\\.)((?!\\\\/\\\\.\\\\.\\\\/).)+[\\\\w]{1,250}\\\\.spdx.json$'",
+    ),
+    (
+        {
+            "options": {
+                "config": {}
+            }
+        },
+        "'doc_path' is a required property",
+    ),
+    (
+        {
+            "options": {}
+        },
+        "'config' is a required property",
+    ),
+    (
+        {
+            "options": {
+                "config": {
+                    "doc_path": "/image.spdx.json",
+                }
+            },
+            "inputs": {
+                "root-tree": {
+                    "type": "org.osbuild.file",
+                    "origin": "org.osbuild.pipeline",
+                    "references": [
+                        "name:root-tree"
+                    ]
+                }
+            }
+        },
+        "'org.osbuild.file' is not one of ['org.osbuild.tree']",
+    ),
+])
+@pytest.mark.parametrize("stage_schema", ["2"], indirect=True)
+def test_schema_validation(stage_schema, test_data, expected_err):
+    test_input = {
+        "type": STAGE_NAME,
+        "options": test_data["options"],
+    }
+    if "inputs" in test_data:
+        test_input["inputs"] = test_data["inputs"]
+
+    res = stage_schema.validate(test_input)
+    if expected_err == "":
+        assert res.valid is True, f"err: {[e.as_dict() for e in res.errors]}"
+    else:
+        assert res.valid is False
+        testutil.assert_jsonschema_error_contains(res, expected_err, expected_num_errs=1)