[GSoC] 1: Insights dashboard

insights/__init__.py

@@ -0,0 +1,8 @@
+#
+# 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.
+#

insights/apps.py

@@ -0,0 +1,14 @@
+#
+# 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.
+#
+
+from django.apps import AppConfig
+
+
+class InsightsConfig(AppConfig):
+    name = "insights"

insights/graphs/__init__.py

@@ -0,0 +1,89 @@
+from functools import partial
+
+from insights.models import GraphDefinition
+
+from .importer_panel import _get_snapshot_data
+from .importer_panel import build_importer_exploit_columns
+from .importer_panel import build_importer_package_columns
+from .importer_panel import collect_importers
+from .package_panel import collect_cwes
+from .package_panel import collect_ecosystem_dist
+from .package_panel import collect_packages
+from .package_panel import format_ecosystem_dist
+from .package_panel import format_top_cwes
+from .package_panel import format_top_packages
+from .severity_panel import collect_severities
+from .severity_panel import get_severity_snapshot_data
+
+GRAPHS = {
+    "pkg-dist-donut": GraphDefinition(
+        id="pkg-dist-donut",
+        title="Ecosystem Distribution",
+        panel="package_panel",
+        chart_type="donut",
+        data_fn=format_ecosystem_dist,
+        collect_fn=collect_ecosystem_dist,
+    ),
+    "pkg-name-donut": GraphDefinition(
+        id="pkg-name-donut",
+        title="Top 10 Packages",
+        panel="package_panel",
+        chart_type="donut",
+        data_fn=format_top_packages,
+        collect_fn=collect_packages,
+        is_per_package=True,
+    ),
+    "pkg-cwe-bar": GraphDefinition(
+        id="pkg-cwe-bar",
+        title="Top 10 CWE Distribution",
+        panel="package_panel",
+        chart_type="colored_bar",
+        data_fn=format_top_cwes,
+        collect_fn=collect_cwes,
+        is_per_package=False,
+        has_search=True,
+    ),
+    "severity-scatter-plot": GraphDefinition(
+        id="severity-scatter-plot",
+        title="Severity Distribution across Packages",
+        panel="severity_panel",
+        chart_type="scatter",
+        data_fn=get_severity_snapshot_data,
+        collect_fn=collect_severities,
+        has_search=True,
+    ),
+    "importer-empty-pkg-bar": GraphDefinition(
+        id="importer-empty-pkg-bar",
+        title="Affected Package Coverage across Importers",
+        panel="importer_panel",
+        chart_type="importer_bar",
+        data_fn=partial(_get_snapshot_data, build_columns_fn=build_importer_package_columns),
+        collect_fn=collect_importers,
+    ),
+    "importer-exploit-bar": GraphDefinition(
+        id="importer-exploit-bar",
+        title="Exploit Coverage across Importers",
+        panel="importer_panel",
+        chart_type="importer_bar",
+        data_fn=partial(_get_snapshot_data, build_columns_fn=build_importer_exploit_columns),
+        collect_fn=collect_importers,
+    ),
+}
+
+PANELS = [
+    "overview_panel",
+    "package_panel",
+    "severity_panel",
+    "importer_panel",
+    "data_quality_panel",
+]
+PANEL_LABELS = {
+    "overview_panel": "Overview",
+    "package_panel": "Package Analytics",
+    "severity_panel": "Severity Analytics",
+    "importer_panel": "Importer Analytics",
+    "data_quality_panel": "Data Quality Analytics",
+}
+PANEL_LAYOUTS = {
+    "package_panel": "split_top",
+}

insights/graphs/importer_panel.py

@@ -0,0 +1,181 @@
+#
+# 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.
+#
+from typing import Any
+from typing import Dict
+
+from django.db.models import Count
+from django.db.models import Exists
+from django.db.models import OuterRef
+from django.db.models import Q
+
+from insights.models import ImporterInsight
+from insights.utils import format_importer_name
+from vulnerabilities.models import AdvisoryExploit
+from vulnerabilities.models import AdvisoryV2
+
+# Ignore Phantom Importers that don't collect Affected Packages
+IGNORED_IMPORTERS = {
+    "epss_importer_v2",
+    "epss",
+    "vulnrichment_importer_v2",
+    "suse_importer_v2",
+}
+
+
+def collect_importers(pipeline):
+    """Calculates and stores importer insights."""
+
+    # Two graphs use this function, so we avoid rerunning the query twice
+    if pipeline.importer_insights:
+        return
+
+    kev_exploits = AdvisoryExploit.objects.filter(advisory=OuterRef("pk"), data_source="KEV")
+    metasploit_exploits = AdvisoryExploit.objects.filter(
+        advisory=OuterRef("pk"), data_source="Metasploit"
+    )
+    exploitdb_exploits = AdvisoryExploit.objects.filter(
+        advisory=OuterRef("pk"), data_source="Exploit-DB"
+    )
+
+    importer_statistics = (
+        AdvisoryV2.objects.filter(is_latest=True)
+        .exclude(datasource_id__in=IGNORED_IMPORTERS)
+        .annotate(
+            has_kev=Exists(kev_exploits),
+            has_metasploit=Exists(metasploit_exploits),
+            has_exploitdb=Exists(exploitdb_exploits),
+        )
+        .values("datasource_id")
+        .annotate(
+            total_advisories=Count("avid", distinct=True),
+            advisories_with_packages=Count(
+                "avid", filter=Q(impacted_packages__affecting_packages__isnull=False), distinct=True
+            ),
+            advisories_with_ghost_packages=Count(
+                "avid", filter=Q(impacted_packages__affecting_packages__is_ghost=True), distinct=True
+            ),
+            advisories_with_kev=Count("avid", filter=Q(has_kev=True), distinct=True),
+            advisories_with_metasploit=Count(
+                "avid", filter=Q(has_kev=False, has_metasploit=True), distinct=True
+            ),
+            advisories_with_exploitdb=Count(
+                "avid",
+                filter=Q(has_kev=False, has_metasploit=False, has_exploitdb=True),
+                distinct=True,
+            ),
+        )
+        .order_by("-total_advisories")
+        .iterator()
+    )
+
+    for record in importer_statistics:
+        pipeline.importer_insights.append(
+            ImporterInsight(
+                importer=record["datasource_id"],
+                total_advisories=record["total_advisories"],
+                advisories_with_packages=record["advisories_with_packages"],
+                advisories_with_ghost_packages=record["advisories_with_ghost_packages"],
+                advisories_with_kev=record["advisories_with_kev"],
+                advisories_with_metasploit=record["advisories_with_metasploit"],
+                advisories_with_exploitdb=record["advisories_with_exploitdb"],
+            )
+        )
+
+
+def _get_snapshot_data(snapshot: Any, build_columns_fn) -> Dict[str, Any]:
+    """Helper to format importer statistics using a provided column builder."""
+    all_importers = list(snapshot.importer_insights.order_by("-total_advisories"))
+    data = {}
+
+    if all_importers:
+        # Top 5 by default for global view
+        data["global"] = build_columns_fn(all_importers[:5])
+
+        # Individual data for each importer
+        for importer_insight in all_importers:
+            formatted_name = format_importer_name(importer_insight.importer)
+            data[formatted_name] = build_columns_fn([importer_insight])
+
+    return data
+
+
+def build_importer_package_columns(importer_list) -> Dict[str, Any]:
+    """Helper to build columns array for affected package horizontal bar charts."""
+    importers = []
+    total_advisories = []
+    advisories_with_packages = []
+    advisories_without_packages = []
+    advisories_with_ghost_packages = []
+    advisories_without_ghost_packages = []
+
+    for importer_insight in importer_list:
+        importers.append(format_importer_name(importer_insight.importer))
+        total_advisories.append(importer_insight.total_advisories)
+        advisories_with_packages.append(importer_insight.advisories_with_packages)
+        advisories_without_packages.append(
+            importer_insight.total_advisories - importer_insight.advisories_with_packages
+        )
+        advisories_with_ghost_packages.append(importer_insight.advisories_with_ghost_packages)
+        advisories_without_ghost_packages.append(
+            importer_insight.advisories_with_packages
+            - importer_insight.advisories_with_ghost_packages
+        )
+
+    return {
+        "x_categories": importers,
+        "columns": [
+            ["total_advisories"] + total_advisories,
+            ["advisories_without_packages"] + advisories_without_packages,
+            ["advisories_with_packages"] + advisories_with_packages,
+            ["advisories_without_ghost_packages"] + advisories_without_ghost_packages,
+            ["advisories_with_ghost_packages"] + advisories_with_ghost_packages,
+        ],
+        "groups": [
+            ["advisories_without_packages", "advisories_with_packages"],
+            ["advisories_without_ghost_packages", "advisories_with_ghost_packages"],
+        ],
+    }
+
+
+def build_importer_exploit_columns(importer_list) -> Dict[str, Any]:
+    """Helper to build columns array for exploit horizontal bar charts."""
+    importers = []
+    total_advisories = []
+    advisories_with_kev = []
+    advisories_with_metasploit = []
+    advisories_with_exploitdb = []
+    advisories_without_exploits = []
+
+    for importer_insight in importer_list:
+        importers.append(format_importer_name(importer_insight.importer))
+        total_advisories.append(importer_insight.total_advisories)
+        advisories_with_kev.append(importer_insight.advisories_with_kev)
+        advisories_with_metasploit.append(importer_insight.advisories_with_metasploit)
+        advisories_with_exploitdb.append(importer_insight.advisories_with_exploitdb)
+
+        total_with_exploits = (
+            importer_insight.advisories_with_kev
+            + importer_insight.advisories_with_metasploit
+            + importer_insight.advisories_with_exploitdb
+        )
+        advisories_without_exploits.append(importer_insight.total_advisories - total_with_exploits)
+
+    return {
+        "x_categories": importers,
+        "columns": [
+            ["total_advisories"] + total_advisories,
+            ["advisories_without_exploits"] + advisories_without_exploits,
+            ["advisories_with_exploitdb"] + advisories_with_exploitdb,
+            ["advisories_with_metasploit"] + advisories_with_metasploit,
+            ["advisories_with_kev"] + advisories_with_kev,
+        ],
+        "groups": [
+            ["advisories_with_exploitdb", "advisories_with_metasploit", "advisories_with_kev"]
+        ],
+    }