minor tweaks in reference notebook

1 files changed, 166 insertions, 80 deletions

diff --git a/notebooks/cc/references.ipynb b/notebooks/cc/references.ipynb
index 271dad5a..28e589ec 100644
--- a/notebooks/cc/references.ipynb
+++ b/notebooks/cc/references.ipynb
@@ -150,15 +150,8 @@
     "df_id_rich[\"has_outgoing_indirect_references\"] = df_id_rich.indirectly_referencing.notnull()\n",
     "df_id_rich[\"has_incoming_indirect_references\"] = df_id_rich.indirectly_referenced_by.notnull()\n",
     "\n",
-    "archived_cert_id_list = set(df_id_rich[df_id_rich.status == \"archived\"].cert_id)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
+    "archived_cert_id_list = set(df_id_rich[df_id_rich.status == \"archived\"].cert_id)\n",
+    "\n",
     "def len_if_exists(x) -> int:\n",
     "    return len(x) if pd.notnull(x) else 0\n",
     "\n",
@@ -240,7 +233,7 @@
     "df_reach_evolution_melted = df_reach_evolution.melt(id_vars=\"date\", var_name=\"certificate\", value_name=\"reach\")\n",
     "\n",
     "g = sns.lineplot(data=df_reach_evolution_melted, x=\"date\", y=\"reach\", hue=\"certificate\")\n",
-    "g.set(title=\"Certificate reach over time\", xlabel=\"Time\", ylabel=\"Certificate reach\")\n",
+    "g.set(title=\"Reach of top-10 certificates in time\", xlabel=\"Time\", ylabel=\"Certificate reach\")\n",
     "plt.savefig(RESULTS_DIR / \"lineplot_top_certificate_reach.pdf\", bbox_inches=\"tight\")\n",
     "plt.show()"
    ]
@@ -259,36 +252,66 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "def compute_avg_references(df: pd.DataFrame, smartcards: bool, variable: str) -> pd.DataFrame:\n",
+    "def compute_avg_references(df: pd.DataFrame, smartcards: bool, variable: str, date_range: pd.DateTimeIndex) -> dict:\n",
     "    \"\"\"\n",
-    "    Computes a series where index is date and value is average 'variable' value for certificates that were valid at that time,\n",
-    "    coming from either smardcards or other categories. E.g., variable can be `directly_referencing`.\n",
+    "    Computes a dictionary (from which series can be created) where index is date and value is average 'variable' value\n",
+    "    for certificates that were valid at that time, coming from either smardcards or other categories.\n",
+    "    E.g., variable can be `directly_referencing`.\n",
     "    \"\"\"\n",
     "    df_copy = df.copy()\n",
     "    df_copy[\"target_variable\"] = df[variable].map(lambda x: len(x) if pd.notnull(x) else 0)\n",
     "\n",
-    "    date_range = pd.date_range(df_copy.not_valid_before.min(), df_copy.not_valid_before.max())\n",
     "    dct = {}\n",
     "    for date in date_range:\n",
     "        if smartcards:\n",
-    "            dct[date] = df_copy.loc[(date >= df_copy.not_valid_before) & (date <= df_copy.not_valid_after) & (df_copy.category == \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")][\"target_variable\"].mean()\n",
+    "            dct[date] = df_copy.loc[\n",
+    "                (date >= df_copy.not_valid_before)\n",
+    "                & (date <= df_copy.not_valid_after)\n",
+    "                & (df_copy.category == \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "            ][\"target_variable\"].mean()\n",
     "        else:\n",
-    "            dct[date] = df_copy.loc[(date >= df_copy.not_valid_before) & (date <= df_copy.not_valid_after) & (df_copy.category != \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")][\"target_variable\"].mean()\n",
-    "    series_name = variable + \" smartcards\" if smartcards else variable + \" other categories\"\n",
-    "    return pd.Series(dct, name=series_name)   \n",
+    "            dct[date] = df_copy.loc[\n",
+    "                (date >= df_copy.not_valid_before)\n",
+    "                & (date <= df_copy.not_valid_after)\n",
+    "                & (df_copy.category != \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "            ][\"target_variable\"].mean()\n",
+    "    return dct\n",
     "\n",
-    "refs_smartcards = compute_avg_references(df_id_rich, True, \"directly_referencing\")\n",
-    "trans_refs_smartcards = compute_avg_references(df_id_rich, True, \"indirectly_referencing\")\n",
-    "refs_others = compute_avg_references(df_id_rich, False, \"directly_referencing\")\n",
-    "trans_refs_others = compute_avg_references(df_id_rich, False, \"indirectly_referencing\")\n",
-    "df_avg_num_refs = pd.concat([refs_smartcards, refs_others, trans_refs_smartcards, trans_refs_others], axis=1)\n",
+    "date_range =  pd.date_range(df_id_rich.not_valid_before.min(), df_id_rich.not_valid_before.max())\n",
+    "\n",
+    "refs_smartcards = compute_avg_references(df_id_rich, True, \"directly_referencing\", date_range)\n",
+    "trans_refs_smartcards = compute_avg_references(df_id_rich, True, \"indirectly_referencing\", date_range)\n",
+    "refs_others = compute_avg_references(df_id_rich, False, \"directly_referencing\", date_range)\n",
+    "trans_refs_others = compute_avg_references(df_id_rich, False, \"indirectly_referencing\", date_range)\n",
+    "df_avg_num_refs = pd.concat(\n",
+    "    [\n",
+    "        pd.Series(refs_smartcards, name=\"smartcard references\"),\n",
+    "        pd.Series(refs_others, name=\"other references\"),\n",
+    "        pd.Series(trans_refs_smartcards, name=\"smartcard transitive references\"),\n",
+    "        pd.Series(trans_refs_others, name=\"other transitive references\"),\n",
+    "    ],\n",
+    "    axis=1,\n",
+    ")\n",
     "df_avg_num_refs.index.name = \"date\"\n",
     "df_avg_num_refs = df_avg_num_refs.reset_index()\n",
     "df_avg_num_refs_melted = df_avg_num_refs.melt(id_vars=[\"date\"], var_name=\"category\", value_name=\"n_references\")\n",
     "\n",
-    "reach_smartcards = compute_avg_references(df_id_rich, True, \"indirectly_referenced_by\")\n",
-    "reach_others = compute_avg_references(df_id_rich, False, \"indirectly_referenced_by\")\n",
-    "df_avg_reach = pd.concat([reach_smartcards, reach_others], axis=1)\n",
+    "reach_smartcards = compute_avg_references(df_id_rich, True, \"indirectly_referenced_by\", date_range)\n",
+    "reach_others = compute_avg_references(df_id_rich, False, \"indirectly_referenced_by\", date_range)\n",
+    "\n",
+    "df_vulnerable = df_id_rich.loc[df_id_rich.related_cves.notnull()]\n",
+    "reach_vuln_smartcards = compute_avg_references(df_vulnerable, True, \"indirectly_referenced_by\", date_range)\n",
+    "reach_vuln_others = compute_avg_references(df_vulnerable, False, \"indirectly_referenced_by\", date_range)\n",
+    "\n",
+    "df_avg_reach = pd.concat(\n",
+    "    [\n",
+    "        pd.Series(reach_smartcards, name=\"transitively referenced smartcards\"),\n",
+    "        pd.Series(reach_others, name=\"transitively referenced by others\"),\n",
+    "        pd.Series(reach_vuln_smartcards, name=\"transitively referenced vulnerable smartcard\"),\n",
+    "        pd.Series(reach_vuln_others, name=\"transitively referenced vulnerable other cert.\"),\n",
+    "    ],\n",
+    "    axis=1,\n",
+    ")\n",
     "df_avg_reach.index.name = \"date\"\n",
     "df_avg_reach = df_avg_reach.reset_index()\n",
     "df_avg_reach_melted = df_avg_reach.melt(id_vars=[\"date\"], var_name=\"category\", value_name=\"certificate reach\")\n",
@@ -298,10 +321,10 @@
     "plt.savefig(RESULTS_DIR / \"lineplot_avg_n_references.pdf\", bbox_inches=\"tight\")\n",
     "plt.show()\n",
     "\n",
-    "g = sns.lineplot(data=df_avg_reach_melted, x=\"date\", y=\"certificate reach\", hue=\"category\")\n",
+    "g = sns.lineplot(data=df_avg_reach_melted, x=\"date\", y=\"certificate reach\", hue=\"category\", errorbar=None)\n",
     "g.set(title=\"Average reach of a certificate in time\", xlabel=\"Time\", ylabel=\"Reach\")\n",
     "plt.savefig(RESULTS_DIR / \"lineplot_avg_reach.pdf\", bbox_inches=\"tight\")\n",
-    "plt.show()"
+    "plt.show()\n"
    ]
   },
   {
@@ -375,30 +398,130 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# For every date range, I first get a subset of all certificates that are valid and have some references in that date.\n",
-    "#    - I also get a list of all cert_ids that are valid at that date\n",
-    "#    - Then I take a dataframe subset of certificates that have non-empty outgoing references, but the intersection of their refs with all cert_ids valid at that date (previous row) is empty => they reference an archived cert. \n",
-    "\n",
     "date_range = pd.date_range(df_id_rich.not_valid_before.min(), df_id_rich.not_valid_before.max())\n",
-    "dct_direct = {}\n",
-    "dct_transitive = {}\n",
+    "dct_direct_others = {}\n",
+    "dct_direct_smartcards = {}\n",
+    "dct_transitive_others = {}\n",
+    "dct_transitive_smartcards = {}\n",
     "for date in tqdm(date_range):\n",
     "    active_certs = df_id_rich.loc[(date >= df_id_rich.not_valid_before) & (date <= df_id_rich.not_valid_after)].copy()\n",
     "    active_certs_cert_ids = set(active_certs[\"cert_id\"].tolist())\n",
-    "    active_certs[\"no_intersection\"] = active_certs.directly_referencing.map(lambda x: False if pd.isnull(x) else not x.intersection(active_certs_cert_ids))\n",
-    "    active_certs[\"no_transitive_intersection\"] = active_certs.indirectly_referencing.map(lambda x: False if pd.isnull(x) else not x.intersection(active_certs_cert_ids))\n",
-    "    dct_direct[date] = active_certs.loc[active_certs.no_intersection].shape[0]\n",
-    "    dct_transitive[date] = active_certs.loc[active_certs.no_transitive_intersection].shape[0]\n",
+    "    active_certs[\"no_intersection\"] = active_certs.directly_referencing.map(\n",
+    "        lambda x: False if pd.isnull(x) else not x.intersection(active_certs_cert_ids)\n",
+    "    )\n",
+    "    active_certs[\"no_transitive_intersection\"] = active_certs.indirectly_referencing.map(\n",
+    "        lambda x: False if pd.isnull(x) else not x.intersection(active_certs_cert_ids)\n",
+    "    )\n",
+    "    dct_direct_others[date] = active_certs.loc[\n",
+    "        (active_certs.no_intersection)\n",
+    "        & (active_certs.category != \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "    dct_transitive_others[date] = active_certs.loc[\n",
+    "        (active_certs.no_transitive_intersection)\n",
+    "        & (active_certs.category != \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "    dct_direct_smartcards[date] = active_certs.loc[\n",
+    "        (active_certs.no_intersection)\n",
+    "        & (active_certs.category == \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "    dct_transitive_smartcards[date] = active_certs.loc[\n",
+    "        (active_certs.no_transitive_intersection)\n",
+    "        & (active_certs.category == \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
     "\n",
-    "df_refs_to_archived = pd.concat([pd.Series(dct_direct, name=\"direct reference\"), pd.Series(dct_transitive, name=\"transitive reference\")], axis=1)\n",
+    "df_refs_to_archived = pd.concat(\n",
+    "    [\n",
+    "        pd.Series(dct_direct_others, name=\"direct reference others\"),\n",
+    "        pd.Series(dct_transitive_others, name=\"transitive reference others\"),\n",
+    "        pd.Series(dct_direct_smartcards, name=\"direct reference smartcards\"),\n",
+    "        pd.Series(dct_transitive_smartcards, name=\"transitive reference smartcards\"),\n",
+    "    ],\n",
+    "    axis=1,\n",
+    ")\n",
     "df_refs_to_archived.index.name = \"date\"\n",
     "df_refs_to_archived = df_refs_to_archived.reset_index()\n",
-    "df_refs_to_archived_melted = df_refs_to_archived.melt(id_vars=[\"date\"], var_name=\"reference type\", value_name=\"number of certificates\")\n",
+    "df_refs_to_archived_melted = df_refs_to_archived.melt(\n",
+    "    id_vars=[\"date\"], var_name=\"reference type\", value_name=\"number of certificates\"\n",
+    ")\n",
     "\n",
     "g = sns.lineplot(data=df_refs_to_archived_melted, x=\"date\", y=\"number of certificates\", hue=\"reference type\")\n",
-    "g.set(title=\"Number of active certificates that reference some archived certificate\", xlabel=\"Time\", ylabel=\"Number of certificates\")\n",
+    "g.set(\n",
+    "    title=\"Number of active certificates that reference some archived certificate\",\n",
+    "    xlabel=\"Time\",\n",
+    "    ylabel=\"Number of certificates\",\n",
+    ")\n",
     "plt.savefig(RESULTS_DIR / \"lienplot_active_certs_referencing_archived.pdf\", bbox_inches=\"tight\")\n",
-    "plt.show()"
+    "plt.show()\n"
+   ]
+  },
+  {
+   "attachments": {},
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Certificates referencing vulnerable certificates in time"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "date_range = pd.date_range(df_id_rich.not_valid_before.min(), df_id_rich.not_valid_before.max())\n",
+    "vulnerable_cert_ids = set(df_id_rich.loc[df_id_rich.related_cves.notnull()].cert_id.tolist())\n",
+    "dct_direct_others = {}\n",
+    "dct_transitive_others = {}\n",
+    "dct_direct_smartcards = {}\n",
+    "dct_transitive_smartcards = {}\n",
+    "for date in tqdm(date_range):\n",
+    "    active_certs = df_id_rich.loc[(date >= df_id_rich.not_valid_before) & (date <= df_id_rich.not_valid_after)].copy()\n",
+    "    active_certs[\"directly_references_vulnerable_cert\"] = active_certs.directly_referencing.map(\n",
+    "        lambda x: False if pd.isnull(x) else bool(x.intersection(vulnerable_cert_ids))\n",
+    "    )\n",
+    "    active_certs[\"transitively_references_vulnerable_cert\"] = active_certs.indirectly_referencing.map(\n",
+    "        lambda x: False if pd.isnull(x) else bool(x.intersection(vulnerable_cert_ids))\n",
+    "    )\n",
+    "    dct_direct_others[date] = active_certs.loc[\n",
+    "        (active_certs.directly_references_vulnerable_cert)\n",
+    "        & (active_certs.category != \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "    dct_transitive_others[date] = active_certs.loc[\n",
+    "        (active_certs.transitively_references_vulnerable_cert)\n",
+    "        & (active_certs.category != \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "    dct_direct_smartcards[date] = active_certs.loc[\n",
+    "        (active_certs.directly_references_vulnerable_cert)\n",
+    "        & (active_certs.category == \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "    dct_transitive_smartcards[date] = active_certs.loc[\n",
+    "        (active_certs.transitively_references_vulnerable_cert)\n",
+    "        & (active_certs.category == \"ICs, Smart Cards and Smart Card-Related Devices and Systems\")\n",
+    "    ].shape[0]\n",
+    "\n",
+    "df_references_vuln = pd.concat(\n",
+    "    [\n",
+    "        pd.Series(dct_direct_others, name=\"direct references others\"),\n",
+    "        pd.Series(dct_transitive_others, name=\"transitive references others\"),\n",
+    "        pd.Series(dct_direct_smartcards, name=\"direct references smartcards\"),\n",
+    "        pd.Series(dct_transitive_smartcards, name=\"transitive references smartcards\"),\n",
+    "    ],\n",
+    "    axis=1,\n",
+    ")\n",
+    "df_references_vuln.index.name = \"date\"\n",
+    "df_references_vuln = df_references_vuln.reset_index()\n",
+    "df_references_vuln_melted = df_references_vuln.melt(\n",
+    "    id_vars=\"date\", var_name=\"reference type\", value_name=\"number of certificates\"\n",
+    ")\n",
+    "\n",
+    "g = sns.lineplot(data=df_references_vuln_melted, x=\"date\", y=\"number of certificates\", hue=\"reference type\")\n",
+    "g.set(\n",
+    "    title=\"Number of active certificates that reference some vulnerable certificate\",\n",
+    "    xlabel=\"Time\",\n",
+    "    ylabel=\"Number of certificates\",\n",
+    ")\n",
+    "plt.savefig(RESULTS_DIR / \"lienplot_active_certs_referencing_vulnerable.pdf\", bbox_inches=\"tight\")\n",
+    "plt.show()\n"
    ]
   },
   {
@@ -604,6 +727,7 @@
     "\n",
     "figure.savefig(str(RESULTS_DIR / \"scheme_references.pdf\"), bbox_inches=\"tight\")\n",
     "figure.savefig(str(RESULTS_DIR / \"scheme_references.pgf\"), bbox_inches=\"tight\")\n",
+    "plt.show()\n",
     "plt.close(figure)"
    ]
   },
@@ -612,43 +736,6 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Temporal evolution of references\n",
-    "\n",
-    "Shows plot with relative number of certificates for a given year that reference some other certificate"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "pycharm": {
-     "name": "#%%\n"
-    }
-   },
-   "outputs": [],
-   "source": [
-    "# TODO: Again, this plot is neither shown nor saved as a figure\n",
-    "df_temporal = df.loc[df.year_from < 2022].groupby([\"year_from\"])[\"directly_referencing\"].count().reset_index().set_index(\"year_from\")\n",
-    "n_issued_certs = df.groupby(\"year_from\").name.count().reset_index().rename(columns={\"name\": \"n_certs\"}).set_index(\"year_from\")\n",
-    "df_temporal.directly_referencing = 100 * df_temporal.directly_referencing / n_issued_certs.n_certs\n",
-    "\n",
-    "line = sns.lineplot(data=df_temporal, x=\"year_from\", y=\"directly_referencing\")\n",
-    "line.yaxis.set_major_formatter(mtick.PercentFormatter())"
-   ]
-  },
-  {
-   "attachments": {},
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Cross references"
-   ]
-  },
-  {
-   "attachments": {},
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
     "## Reference network visualization"
    ]
   },
@@ -820,7 +907,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# TODO: Not sure what this thing does\n",
     "nx.draw(view, pos=nx.planar_layout(view), with_labels=True)"
    ]
   },