research: index complete IANA IPv4 range baseline

tests/test_sync_iana_ipv4.py

@@ -0,0 +1,121 @@
+from __future__ import annotations
+
+import csv
+import importlib.util
+import io
+import tempfile
+import unittest
+from ipaddress import IPv4Network
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[1]
+MODULE_PATH = ROOT / "scripts" / "sync_iana_ipv4.py"
+
+
+def load_module():
+    spec = importlib.util.spec_from_file_location("sync_iana_ipv4", MODULE_PATH)
+    if spec is None or spec.loader is None:
+        raise RuntimeError("cannot load sync_iana_ipv4")
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    return module
+
+
+class SyncIanaIPv4Tests(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.sync = load_module()
+
+    def test_top_level_prefix_is_normalized(self):
+        self.assertEqual(
+            self.sync.top_level_network("001/8"),
+            IPv4Network("1.0.0.0/8"),
+        )
+
+    def test_registry_slug_uses_rdap_authority(self):
+        self.assertEqual(
+            self.sync.registry_slug("APNIC", "https://rdap.apnic.net/"),
+            "apnic",
+        )
+        self.assertEqual(
+            self.sync.registry_slug(
+                "Legacy allocation",
+                "https://rdap.arin.net/registryhttp://rdap.arin.net/registry",
+            ),
+            "arin",
+        )
+        self.assertEqual(self.sync.registry_slug("IANA - Reserved", ""), "iana")
+
+    def test_rdap_url_concatenation_is_normalized(self):
+        self.assertEqual(
+            self.sync.clean_rdap("https://rdap.arin.net/registryhttp://rdap.arin.net/registry"),
+            "https://rdap.arin.net/registry",
+        )
+        self.assertEqual(self.sync.clean_rdap("https://rdap.apnic.net/"), "https://rdap.apnic.net/")
+
+    def test_filename_sorts_numerically_in_lexical_order(self):
+        self.assertEqual(
+            self.sync.range_filename(IPv4Network("1.2.3.0/24"), "apnic"),
+            "001-002-003-000--24--apnic.md",
+        )
+
+    def test_multiple_special_blocks_are_expanded(self):
+        self.assertEqual(
+            self.sync.expand_address_blocks("192.0.0.170/32, 192.0.0.171/32"),
+            [IPv4Network("192.0.0.170/32"), IPv4Network("192.0.0.171/32")],
+        )
+
+    def test_special_block_footnote_is_ignored_for_parsing(self):
+        self.assertEqual(
+            self.sync.expand_address_blocks("192.0.0.0/24 [2]"),
+            [IPv4Network("192.0.0.0/24")],
+        )
+
+    def test_source_text_whitespace_is_normalized(self):
+        self.assertEqual(self.sync.clean_name('"[RFC8190]\n        [RFC919], Section 7"'), "[RFC8190] [RFC919], Section 7")
+
+    def test_parse_top_level_requires_complete_ipv4_coverage(self):
+        output = io.StringIO()
+        writer = csv.DictWriter(
+            output,
+            fieldnames=["Prefix", "Designation", "Date", "WHOIS", "RDAP", "Status [1]", "Note"],
+        )
+        writer.writeheader()
+        for first in range(256):
+            writer.writerow({
+                "Prefix": f"{first:03d}/8",
+                "Designation": "Example",
+                "Date": "2026-01",
+                "WHOIS": "",
+                "RDAP": "",
+                "Status [1]": "RESERVED",
+                "Note": "",
+            })
+        rows = self.sync.parse_top_level_csv(output.getvalue())
+        self.assertEqual(len(rows), 256)
+        self.assertEqual(rows[0]["network"], IPv4Network("0.0.0.0/8"))
+        self.assertEqual(rows[-1]["network"], IPv4Network("255.0.0.0/8"))
+
+    def test_parse_top_level_rejects_missing_slash_eight(self):
+        sample = "Prefix,Designation,Date,WHOIS,RDAP,Status [1],Note\n000/8,IANA,1981-09,,,RESERVED,\n"
+        with self.assertRaisesRegex(ValueError, "256"):
+            self.sync.parse_top_level_csv(sample)
+
+    def test_manual_range_file_is_not_overwritten(self):
+        with tempfile.TemporaryDirectory() as temporary:
+            directory = Path(temporary)
+            path = directory / "001-000-000-000--08--apnic.md"
+            path.write_text("# hand researched\n", encoding="utf-8")
+            changed = self.sync.write_generated(path, "# generated\n")
+            self.assertFalse(changed)
+            self.assertEqual(path.read_text(encoding="utf-8"), "# hand researched\n")
+
+    def test_generated_csv_uses_lf_line_endings(self):
+        with tempfile.TemporaryDirectory() as temporary:
+            path = Path(temporary) / "ranges.csv"
+            self.sync.write_csv(path, ["prefix"], [{"prefix": "1.0.0.0/8"}])
+            self.assertNotIn(b"\r", path.read_bytes())
+
+
+if __name__ == "__main__":
+    unittest.main()