Migrate pysec importer to aboutcode pipeline

Signed-off-by: Keshav Priyadarshi <[email protected]>

Author: keshav-space

vulnerabilities/importers/__init__.py

@@ -26,7 +26,6 @@
 from vulnerabilities.importers import oss_fuzz
 from vulnerabilities.importers import postgresql
 from vulnerabilities.importers import project_kb_msr2019
-from vulnerabilities.importers import pysec
 from vulnerabilities.importers import redhat
 from vulnerabilities.importers import retiredotnet
 from vulnerabilities.importers import ruby
@@ -42,9 +41,9 @@
 from vulnerabilities.pipelines import npm_importer
 from vulnerabilities.pipelines import nvd_importer
 from vulnerabilities.pipelines import pypa_importer
+from vulnerabilities.pipelines import pysec_importer
 
 IMPORTERS_REGISTRY = [
-    pysec.PyPIImporter,
     alpine_linux.AlpineImporter,
     openssl.OpensslImporter,
     redhat.RedhatImporter,
@@ -78,6 +77,7 @@
     gitlab_importer.GitLabImporterPipeline,
     github_importer.GitHubAPIImporterPipeline,
     nvd_importer.NVDImporterPipeline,
+    pysec_importer.PyPIImporterPipeline,
 ]
 
 IMPORTERS_REGISTRY = {

vulnerabilities/importers/pysec.py

@@ -6,7 +6,7 @@
 # See https://github.com/aboutcode-org/vulnerablecode for support or download.
 # See https://aboutcode.org for more information about nexB OSS projects.
 #
-import logging
+
 from pathlib import Path
 from typing import Iterable
 

vulnerabilities/pipelines/pypa_importer.py

@@ -0,0 +1,65 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# VulnerableCode is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0
+# See http://www.apache.org/licenses/LICENSE-2.0 for the license text.
+# See https://github.com/aboutcode-org/vulnerablecode for support or download.
+# See https://aboutcode.org for more information about nexB OSS projects.
+#
+import json
+import logging
+from io import BytesIO
+from typing import Iterable
+from zipfile import ZipFile
+
+import requests
+
+from vulnerabilities.importer import AdvisoryData
+from vulnerabilities.importers.osv import parse_advisory_data
+from vulnerabilities.pipelines import VulnerableCodeBaseImporterPipeline
+
+
+class PyPIImporterPipeline(VulnerableCodeBaseImporterPipeline):
+    """Collect advisories from PyPI."""
+
+    pipeline_id = "pysec_importer"
+
+    license_url = "https://github.com/pypa/advisory-database/blob/main/LICENSE"
+    url = "https://osv-vulnerabilities.storage.googleapis.com/PyPI/all.zip"
+    spdx_license_expression = "CC-BY-4.0"
+    importer_name = "PyPI Importer"
+
+    @classmethod
+    def steps(cls):
+        return (
+            cls.fetch_zip,
+            cls.collect_and_store_advisories,
+            cls.import_new_advisories,
+        )
+
+    def fetch_zip(self):
+        self.log(f"Fetching `{self.url}`")
+        self.advisory_zip = requests.get(self.url).content
+
+    def advisories_count(self) -> int:
+        with ZipFile(BytesIO(self.advisory_zip)) as zip:
+            advisory_count = sum(1 for file in zip.namelist() if file.startswith("PYSEC-"))
+        return advisory_count
+
+    def collect_advisories(self) -> Iterable[AdvisoryData]:
+        """Yield AdvisoryData using a zipped data dump of OSV data"""
+        with ZipFile(BytesIO(self.advisory_zip)) as zip_file:
+            for file_name in zip_file.namelist():
+                if not file_name.startswith("PYSEC-"):
+                    self.log(
+                        f"Unsupported PyPI advisory data file: {file_name}",
+                        level=logging.ERROR,
+                    )
+                    continue
+                with zip_file.open(file_name) as f:
+                    vul_info = json.load(f)
+                    yield parse_advisory_data(
+                        raw_data=vul_info,
+                        supported_ecosystems=["pypi"],
+                        advisory_url=self.url,
+                    )

vulnerabilities/pipelines/pysec_importer.py

@@ -7,52 +7,51 @@
 # See https://aboutcode.org for more information about nexB OSS projects.
 #
 import json
-import os
+from pathlib import Path
 from unittest import TestCase
 
 from vulnerabilities.importers.osv import parse_advisory_data
 from vulnerabilities.tests.util_tests import VULNERABLECODE_REGEN_TEST_FIXTURES as REGEN
 from vulnerabilities.tests.util_tests import check_results_against_json
 
-BASE_DIR = os.path.dirname(os.path.abspath(__file__))
-TEST_DATA = os.path.join(BASE_DIR, "test_data/pysec")
+TEST_DATA = Path(__file__).parent.parent / "test_data" / "pysec"
 
 
 class TestPyPIImporter(TestCase):
     def test_to_advisories_with_summary(self):
-        with open(os.path.join(TEST_DATA, "pysec-advisories_with_summary.json")) as f:
+        with open(TEST_DATA / "pysec-advisories_with_summary.json") as f:
             mock_response = json.load(f)
         results = parse_advisory_data(mock_response, ["pypi"], "https://test.com").to_dict()
 
-        expected_file = os.path.join(TEST_DATA, "pysec-advisories_with_summary-expected.json")
+        expected_file = TEST_DATA / "pysec-advisories_with_summary-expected.json"
         check_results_against_json(
             results=results,
             expected_file=expected_file,
             regen=REGEN,
         )
 
     def test_to_advisories_without_summary(self):
-        with open(os.path.join(TEST_DATA, "pysec-advisories_without_summary.json")) as f:
+        with open(TEST_DATA / "pysec-advisories_without_summary.json") as f:
             mock_response = json.load(f)
 
         results = parse_advisory_data(mock_response, ["pypi"], "https://test.com").to_dict()
 
-        expected_file = os.path.join(TEST_DATA, "pysec-advisories_without_summary-expected.json")
+        expected_file = TEST_DATA / "pysec-advisories_without_summary-expected.json"
         check_results_against_json(
             results=results,
             expected_file=expected_file,
             regen=REGEN,
         )
 
     def test_to_advisories_with_cwe(self):
-        with open(os.path.join(TEST_DATA, "pysec-advisory_with_cwe.json")) as f:
+        with open(TEST_DATA / "pysec-advisory_with_cwe.json") as f:
             mock_response = json.load(f)
 
         results = parse_advisory_data(
             raw_data=mock_response, supported_ecosystems=["pypi"], advisory_url="https://tes.com"
         ).to_dict()
 
-        expected_file = os.path.join(TEST_DATA, "pysec-advisories_with_cwe-expected.json")
+        expected_file = TEST_DATA / "pysec-advisories_with_cwe-expected.json"
         check_results_against_json(
             results=results,
             expected_file=expected_file,