Merge pull request #600 from Hritik14/migration/importer_yielder

Dump importer_yielder in favor of IMPORTER_REGISTRY and drop Etags
aboutcode-org · Feb 7, 2022 · 4ebaa48 · 4ebaa48
2 parents 14850ea + 45cdaf5
commit 4ebaa48
Show file tree

Hide file tree

Showing 79 changed files with 664 additions and 1,407 deletions.
diff --git a/vulnerabilities/admin.py b/vulnerabilities/admin.py
@@ -25,7 +25,6 @@
 
 from vulnerabilities.models import (
     PackageRelatedVulnerability,
-    Importer,
     Package,
     Vulnerability,
     VulnerabilityReference,
@@ -55,11 +54,6 @@ class PackageRelatedVulnerabilityAdmin(admin.ModelAdmin):
     search_fields = ["vulnerability__vulnerability_id", "package__name"]
 
 
-@admin.register(Importer)
-class ImporterAdmin(admin.ModelAdmin):
-    pass
-
-
 @admin.register(VulnerabilitySeverity)
 class VulnerabilitySeverityAdmin(admin.ModelAdmin):
     pass
diff --git a/vulnerabilities/fixtures/github.json b/vulnerabilities/fixtures/github.json
@@ -96,7 +96,7 @@
             "name": "github",
             "license": "",
             "last_run": "2021-03-06T09:09:01.523Z",
-            "data_source": "GitHubAPIDataSource",
+            "data_source": "GitHubAPIImporter",
             "data_source_cfg": {
                 "endpoint": "https://api.github.com/graphql",
                 "ecosystems": [
@@ -115,4 +115,4 @@
             "reference": 136
         }
     }
-]
+]
diff --git a/vulnerabilities/fixtures/openssl.json b/vulnerabilities/fixtures/openssl.json
@@ -40538,7 +40538,7 @@
             "name": "rust",
             "license": "cc0-1.0",
             "last_run": "2020-09-14T12:47:17.451Z",
-            "data_source": "RustDataSource",
+            "data_source": "RustImporter",
             "data_source_cfg": {
                 "branch": null,
                 "repository_url": "https://github.com/RustSec/advisory-db",
@@ -40555,7 +40555,7 @@
             "name": "alpine",
             "license": "",
             "last_run": null,
-            "data_source": "AlpineDataSource",
+            "data_source": "AlpineImporter",
             "data_source_cfg": {
                 "branch": null,
                 "repository_url": "https://gitlab.alpinelinux.org/alpine/infra/alpine-secdb"
@@ -40569,7 +40569,7 @@
             "name": "archlinux",
             "license": "mit",
             "last_run": null,
-            "data_source": "ArchlinuxDataSource",
+            "data_source": "ArchlinuxImporter",
             "data_source_cfg": {
                 "archlinux_tracker_url": "https://security.archlinux.org/json"
             }
@@ -40582,7 +40582,7 @@
             "name": "debian",
             "license": "mit",
             "last_run": null,
-            "data_source": "DebianDataSource",
+            "data_source": "DebianImporter",
             "data_source_cfg": {
                 "debian_tracker_url": "https://security-tracker.debian.org/tracker/data/json"
             }
@@ -40595,7 +40595,7 @@
             "name": "safetydb",
             "license": "cc-by-nc-4.0",
             "last_run": null,
-            "data_source": "SafetyDbDataSource",
+            "data_source": "SafetyDbImporter",
             "data_source_cfg": {
                 "url": "https://raw.githubusercontent.com/pyupio/safety-db/master/data/insecure_full.json",
                 "etags": {}
@@ -40609,7 +40609,7 @@
             "name": "npm",
             "license": "mit",
             "last_run": null,
-            "data_source": "NpmDataSource",
+            "data_source": "NpmImporter",
             "data_source_cfg": {
                 "repository_url": "https://github.com/nodejs/security-wg.git"
             }
@@ -40622,7 +40622,7 @@
             "name": "ruby",
             "license": "",
             "last_run": null,
-            "data_source": "RubyDataSource",
+            "data_source": "RubyImporter",
             "data_source_cfg": {
                 "repository_url": "https://github.com/rubysec/ruby-advisory-db.git"
             }
@@ -40635,7 +40635,7 @@
             "name": "ubuntu",
             "license": "gpl-2.0",
             "last_run": null,
-            "data_source": "UbuntuDataSource",
+            "data_source": "UbuntuImporter",
             "data_source_cfg": {
                 "etags": {},
                 "releases": [
@@ -40655,7 +40655,7 @@
             "name": "retiredotnet",
             "license": "mit",
             "last_run": null,
-            "data_source": "RetireDotnetDataSource",
+            "data_source": "RetireDotnetImporter",
             "data_source_cfg": {
                 "repository_url": "https://github.com/RetireNet/Packages.git"
             }
@@ -40668,7 +40668,7 @@
             "name": "suse_backports",
             "license": "",
             "last_run": null,
-            "data_source": "SUSEBackportsDataSource",
+            "data_source": "SUSEBackportsImporter",
             "data_source_cfg": {
                 "url": "http://ftp.suse.com/pub/projects/security/yaml/",
                 "etags": {}
@@ -40682,7 +40682,7 @@
             "name": "debian_oval",
             "license": "",
             "last_run": null,
-            "data_source": "DebianOvalDataSource",
+            "data_source": "DebianOvalImporter",
             "data_source_cfg": {
                 "etags": {},
                 "releases": [
@@ -40701,7 +40701,7 @@
             "name": "redhat",
             "license": "cc-by-4.0",
             "last_run": null,
-            "data_source": "RedhatDataSource",
+            "data_source": "RedhatImporter",
             "data_source_cfg": {}
         }
     },
@@ -40712,7 +40712,7 @@
             "name": "gentoo",
             "license": "",
             "last_run": null,
-            "data_source": "GentooDataSource",
+            "data_source": "GentooImporter",
             "data_source_cfg": {
                 "repository_url": "https://anongit.gentoo.org/git/data/glsa.git"
             }
@@ -40725,7 +40725,7 @@
             "name": "openssl",
             "license": "",
             "last_run": "2020-09-14T12:52:58.762Z",
-            "data_source": "OpenSSLDataSource",
+            "data_source": "OpenSSLImporter",
             "data_source_cfg": {
                 "etags": {
                     "https://www.openssl.org/news/vulnerabilities.xml": "\"513bd-5aee0a1c716f0-gzip\""
@@ -40740,7 +40740,7 @@
             "name": "ubuntu_usn",
             "license": "gpl-2.0",
             "last_run": null,
-            "data_source": "UbuntuUSNDataSource",
+            "data_source": "UbuntuUSNImporter",
             "data_source_cfg": {
                 "etags": {},
                 "db_url": "https://usn.ubuntu.com/usn-db/database-all.json.bz2"
@@ -40754,7 +40754,7 @@
             "name": "github",
             "license": "",
             "last_run": null,
-            "data_source": "GitHubAPIDataSource",
+            "data_source": "GitHubAPIImporter",
             "data_source_cfg": {
                 "endpoint": "https://api.github.com/graphql",
                 "ecosystems": [
@@ -40765,4 +40765,4 @@
             }
         }
     }
-]
+]
diff --git a/vulnerabilities/helpers.py b/vulnerabilities/helpers.py
@@ -27,6 +27,7 @@
 from typing import List
 from typing import Optional
 from typing import Tuple
+from unittest.mock import MagicMock
 
 import requests
 import saneyaml
@@ -67,32 +68,8 @@ def fetch_yaml(url):
     return saneyaml.load(response.content)
 
 
-# FIXME: this is NOT how etags work .
-# We should instead send the proper HTTP header
-# https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/If-None-Match
-# and integrate this finely in the processing as this typically needs to use
-# streaming=True requests, and proper handling of the HTTP return code
-# In all cases this ends up being a single request, not a HEADD followed
-# by another real request
-def create_etag(data_src, url, etag_key):
-    """
-    Etags are like hashes of web responses. For a data source `data_src`,
-    we maintain (url, etag) mappings in the DB.  `create_etag`  creates
-    (`url`, etag) pair. If a (`url`, etag) already exists then the code
-    skips processing the response further to avoid duplicate work.
-
-    `etag_key` is the name of header which contains the etag for the url.
-    """
-    etag = requests.head(url).headers.get(etag_key)
-    if not etag:
-        return True
-
-    elif url in data_src.config.etags:
-        if data_src.config.etags[url] == etag:
-            return False
-
-    data_src.config.etags[url] = etag
-    return True
+# FIXME: Remove this entirely after complete importer-improver migration
+create_etag = MagicMock()
 
 
 def contains_alpha(string):
@@ -190,3 +167,12 @@ def split_markdown_front_matter(text: str) -> Tuple[str, str]:
         return frontmatter, markdown
 
     return "", text
+
+
+# TODO: Replace this with combination of @classmethod and @property after upgrading to python 3.9
+class classproperty(object):
+    def __init__(self, fget):
+        self.fget = fget
+
+    def __get__(self, owner_self, owner_cls):
+        return self.fget(owner_cls)
diff --git a/vulnerabilities/import_runner.py b/vulnerabilities/import_runner.py
@@ -25,79 +25,50 @@
 import datetime
 import json
 import logging
-from typing import Set
+from typing import List
 from typing import Iterable
 
 
 from vulnerabilities import models
 from vulnerabilities.models import Advisory
-from vulnerabilities.data_source import AdvisoryData
+from vulnerabilities.importer import AdvisoryData
+from vulnerabilities.importer import Importer
 
 logger = logging.getLogger(__name__)
 
 
 class ImportRunner:
     """
     The ImportRunner is responsible for inserting and updating data about vulnerabilities and
-    affected/unaffected/fixed packages in the database. The two main goals for the implementation
-    are correctness and efficiency.
+    affected/unaffected/fixed packages in the database. The main goal for the implementation
+    is correctness
 
     Correctness:
         - There must be no duplicates in the database (should be enforced by the schema).
         - No valid data from the data source must be skipped or truncated.
-
-    Efficiency:
-        - Bulk inserts should be used whenever possible.
-        - Checking whether a record already exists should be kept to a minimum
-        (the data source should know this instead).
-        - All update and select operations must use indexed columns.
     """
 
-    def __init__(self, importer: models.Importer):
+    def __init__(self, importer: Importer):
         self.importer = importer
 
-    def run(self, cutoff_date: datetime.datetime = None) -> None:
+    def run(self) -> None:
         """
         Create a data source for the given importer and store the data retrieved in the database.
-
-        cutoff_date - optional timestamp of the oldest data to include in the import
-
-        NB: Data sources provide two kinds of records; vulnerabilities and packages. Vulnerabilities
-        are potentially shared across many packages, from the same data source and from different
-        data sources. For example, a vulnerability in the Linux kernel is mentioned by advisories
-        from all Linux distributions that package this kernel version.
         """
-        logger.info(f"Starting import for {self.importer.name}.")
-        data_source = self.importer.make_data_source(cutoff_date=cutoff_date)
-        with data_source:
-            advisory_data = data_source.advisory_data()
-            importer_name = data_source.qualified_name()
-            process_advisories(advisory_datas=advisory_data, importer_name=importer_name)
-        self.importer.last_run = datetime.datetime.now(tz=datetime.timezone.utc)
-        self.importer.data_source_cfg = dataclasses.asdict(data_source.config)
-        self.importer.save()
-
-        logger.info(f"Finished import for {self.importer.name}.")
-
+        importer_name = self.importer.qualified_name
+        importer_class = self.importer
+        logger.info(f"Starting import for {importer_name}")
+        advisory_datas = importer_class().advisory_data()
+        count = process_advisories(advisory_datas=advisory_datas, importer_name=importer_name)
+        logger.info(f"Finished import for {importer_name}. Imported {count} advisories.")
 
-def vuln_ref_exists(vulnerability, url, reference_id):
-    return models.VulnerabilityReference.objects.filter(
-        vulnerability=vulnerability, reference_id=reference_id, url=url
-    ).exists()
 
-
-def get_vuln_pkg_refs(vulnerability, package):
-    return models.PackageRelatedVulnerability.objects.filter(
-        vulnerability=vulnerability,
-        package=package,
-    )
-
-
-def process_advisories(advisory_datas: Iterable[AdvisoryData], importer_name: str) -> None:
+def process_advisories(advisory_datas: Iterable[AdvisoryData], importer_name: str) -> List:
     """
     Insert advisories into the database
+    Return the number of inserted advisories.
     """
-
+    count = 0
     for data in advisory_datas:
         obj, created = Advisory.objects.get_or_create(
             aliases=data.aliases,
@@ -114,5 +85,8 @@ def process_advisories(advisory_datas: Iterable[AdvisoryData], importer_name: st
             logger.info(
                 f"[*] New Advisory with aliases: {obj.aliases!r}, created_by: {obj.created_by}"
             )
+            count += 1
         else:
             logger.debug(f"Advisory with aliases: {obj.aliases!r} already exists. Skipped.")
+
+    return count