feat: link system ids to their files' section

A reviewed SYSTEM_PARENTS table leads variants (WonderSwan Color, Neo Geo MVS, MSX2) to the section holding their files, an Arcade heading that reads the name exactly is accepted, and system_search_aliases() gives the platforms' native ids for search. 260 of 636 ids linked before, 338 after.

scripts/siterender.py

@@ -223,14 +223,20 @@ def emulator_card_page(emulator: str, anchor: str) -> str:
 
 
 def system_section(sys_id: str) -> tuple[str, str] | None:
-    """The console section a system id is named after, as (manufacturer
-    slug, console): its own section, else the one section whose heading
-    reads system_name(). The folder a system's files sit in is no answer:
-    Emerson's Arcadia 2001 led to the Saturn and Sega's NAOMI to the
-    Dreamcast."""
+    """The console section a system id leads to, as (manufacturer slug,
+    console): its own section, else the parent SYSTEM_PARENTS names, else
+    the one section whose heading reads system_name(). The folder a
+    system's files sit in is no answer by itself: Emerson's Arcadia 2001
+    led to the Saturn and Sega's NAOMI to the Dreamcast.
+
+    A heading under Arcade counts when it reads the name exactly (Namco
+    System 246). One under Other never does, its headings name programs."""
     own = _own_section(sys_id)
     if own is not None:
         return _manufacturer_slug(own[0]), own[1]
+    parent = SYSTEM_PARENTS.get(sys_id)
+    if parent is not None and parent in _section_pairs:
+        return _manufacturer_slug(parent[0]), parent[1]
     try:
         name = system_name(sys_id)
     except KeyError:
@@ -239,7 +245,9 @@ def system_section(sys_id: str) -> tuple[str, str] | None:
     if len(places) != 1:
         return None
     mfr, console = next(iter(places))
-    if mfr in _PROGRAM_TOPS:
+    if mfr in _NAMELESS_TOPS or (
+        mfr in _PROGRAM_TOPS and console.casefold() in _program_names
+    ):
         return None
     return _manufacturer_slug(mfr), console
 
@@ -359,7 +367,12 @@ def _by_mode(value: str | dict, mode: str) -> str:
 # Folders that file copies by the emulator or board that loads them. Their
 # headings name a program, not a system.
 _PROGRAM_TOPS = frozenset({"Other", "Arcade"})
+# Of those, the top whose headings never name a system, even read whole.
+# An Arcade heading can (Namco System 246), an Other one cannot.
+_NAMELESS_TOPS = frozenset({"Other"})
 
+# Every (manufacturer, console) section the system pages carry.
+_section_pairs: set[tuple[str, str]] = set()
 # Normalized id -> the console sections spelled that way.
 _console_sections: dict[str, set[tuple[str, str]]] = {}
 # A console section by the name its heading reads (Bandai WonderSwan).
@@ -402,6 +415,8 @@ def set_console_sections(
     _console_sections.clear()
     _sections_by_name.clear()
     _console_anchors.clear()
+    _section_pairs.clear()
+    _section_pairs.update(pairs)
     by_maker: dict[str, list[str]] = {}
     for mfr, console in pairs:
         by_maker.setdefault(mfr, []).append(console)
@@ -927,3 +942,172 @@ SYSTEM_NAMES: dict[str, str] = {
     "zx-spectrum": "Sinclair ZX Spectrum",
     "zx-spectrum-128": "Sinclair ZX Spectrum 128",
 }
+
+
+# A system id with no section of its own -> the (manufacturer, console)
+# section holding its files: a model, a regional name or an add-on of the
+# machine that section covers, or a game whose data that section files.
+# Reviewed entry by entry against the files the profiles and platforms
+# declare for the id, never decided by a vote of the folders those files
+# sit in. A section that is not the machine stays out of this table even
+# when it holds a copy: NAOMI's naomi2.zip filed with the Dreamcast, the
+# PET's ROMs in VICE's tree under the C128, the Arcadia's under the Saturn.
+SYSTEM_PARENTS: dict[str, tuple[str, str]] = {
+    # Nintendo. The twin folders (DS and Nintendo DS, 64DD and Nintendo
+    # 64DD) gave each name two headings and no link.
+    "64dd": ("Nintendo", "Nintendo 64DD"),
+    "nintendo-64dd": ("Nintendo", "Nintendo 64DD"),
+    "nintendo-ds": ("Nintendo", "Nintendo DS"),
+    "nintendo-dsi": ("Nintendo", "Nintendo DS"),
+    "famicom": ("Nintendo", "NES"),
+    "super-famicom": ("Nintendo", "SNES"),
+    # Bandai: the Color boot ROMs, and those of Benesse's WonderSwan-based
+    # Pocket Challenge V2, sit with the WonderSwan.
+    "wonderswan-color": ("Bandai", "WonderSwan"),
+    "bandai-wswanc": ("Bandai", "WonderSwan"),
+    "pocket-challenge-v2": ("Bandai", "WonderSwan"),
+    # SNK
+    "ngpc": ("SNK", "Neo Geo Pocket"),
+    "snk-ngpc": ("SNK", "Neo Geo Pocket"),
+    "neo-geo-pocket-color": ("SNK", "Neo Geo Pocket"),
+    "snk-neogeo-aes": ("SNK", "Neo Geo"),
+    "snk-neogeo-mvs": ("SNK", "Neo Geo"),
+    # Sega. The VMU is the Dreamcast's memory card, its images sit there.
+    # Sega's arcade section holds the OutRun board ROMs Cannonball reads.
+    "sega-genesis": ("Sega", "Mega Drive"),
+    "mega-cd-32x": ("Sega", "Mega CD"),
+    "sega-vmu": ("Sega", "Dreamcast"),
+    "sega-outrun": ("Sega", "Arcade"),
+    # NEC and Pioneer: both LaserActive packs (LD-ROM2 and Mega LD) file
+    # their BIOS under the LaserActive.
+    "nec-supergrafx": ("NEC", "PC Engine"),
+    "supergrafx": ("NEC", "PC Engine"),
+    "supergrafx-cd": ("NEC", "PC Engine CD"),
+    "pc-engine-ld": ("Pioneer", "LaserActive"),
+    "mega-ld": ("Pioneer", "LaserActive"),
+    # Microsoft MSX generations
+    "microsoft-msx2": ("Microsoft", "MSX"),
+    "msx2": ("Microsoft", "MSX"),
+    "msx2plus": ("Microsoft", "MSX"),
+    "msx2+": ("Microsoft", "MSX"),
+    "msxturboR": ("Microsoft", "MSX"),
+    "msxturbor": ("Microsoft", "MSX"),
+    # Sinclair: 128-0.rom and 128-1.rom, and Timex's tc2048.rom.
+    "zx-spectrum-128": ("Sinclair", "ZX Spectrum"),
+    "timex-tc2048": ("Sinclair", "ZX Spectrum"),
+    # Acorn: os12.rom and basic2.rom for the Model B and B+, and the
+    # Master Compact's bbcmc.zip.
+    "bbc-micro-b": ("Acorn", "BBC Micro"),
+    "bbc-micro-bplus": ("Acorn", "BBC Micro"),
+    "bbc-master-compact": ("Acorn", "BBC Micro"),
+    # Atari: falcon.img, ste.img and tt.img sit with the ST TOS images.
+    "atari-ste": ("Atari", "ST"),
+    "atari-tt": ("Atari", "ST"),
+    "atari-falcon": ("Atari", "ST"),
+    # Home computers filed under their family
+    "apple-iie": ("Apple", "Apple II"),
+    "apf-m1000": ("APF", "Imagination Machine"),
+    "apfm1000": ("APF", "Imagination Machine"),
+    "bbk-longman-4980": ("BBK", "A4980"),
+    "bbk-longman-4988": ("BBK", "A4988"),
+    "coco3": ("Tandy", "CoCo"),
+    "commodore-cdtv": ("Commodore", "Amiga"),
+    "dragon32": ("Dragon", "Dragon"),
+    "dragon64": ("Dragon", "Dragon"),
+    "enterprise-64": ("Enterprise", "64-128"),
+    "enterprise-128": ("Enterprise", "64-128"),
+    "fm77av": ("Fujitsu", "FM-7"),
+    "gamepock": ("Epoch", "Game Pocket"),
+    "ibm-pcjr": ("IBM", "PC"),
+    "ibm-xt": ("IBM", "PC"),
+    "marty": ("Fujitsu", "FM Towns"),
+    "mz700": ("Sharp", "MZ"),
+    "mz80k": ("Sharp", "MZ"),
+    "oricatmos": ("Oric", "Oric"),
+    "palm-os": ("Palm", "Palm"),
+    "philips-vg5k": ("Philips", "VG5000"),
+    "vg5k": ("Philips", "VG5000"),
+    "videobrain": ("VideoBrain", "Family Computer"),
+    # Files kept at the maker's root, so each file is a heading of its own
+    # on a page that is the machine's: the first one leads there.
+    "galaksija": ("Galaksija", "CHRGEN.BIN"),
+    "vircon32": ("Vircon", "Vircon32Bios.v32"),
+    # Games, filed with the game whose folder holds their data
+    "dinothawr": ("Dinothawr", "Dinothawr"),
+    "doom": ("Id Software", "Doom"),
+    "ctf": ("Id Software", "Quake II"),
+    "teamarena": ("Id Software", "Quake III Arena"),
+    "enemyterritory": ("Id Software", "Wolfenstein Enemy Territory"),
+    "gearbox": ("Valve", "Half-Life"),
+    "dmc": ("Valve", "Half-Life"),
+    "ricochet": ("Valve", "Half-Life"),
+    "tfc": ("Valve", "Half-Life"),
+    "mohaa": ("Electronic Arts", "Medal of Honor Allied Assault"),
+    "mohaa-breakthrough": ("Electronic Arts", "Medal of Honor Allied Assault"),
+    "mohaa-spearhead": ("Electronic Arts", "Medal of Honor Allied Assault"),
+    "ur-quan-masters": ("Toys for Bob", "The Ur-Quan Masters"),
+    "star-control-2": ("Toys for Bob", "The Ur-Quan Masters"),
+    # The collection's general arcade section, where naomi.zip, neogeo.zip
+    # and the other board sets sit.
+    "arcade": ("Arcade", "Arcade"),
+}
+
+
+def system_ids(profiles: dict[str, dict], platforms: list[dict]) -> set[str]:
+    """Every system id a page can name: the platforms' and the profiles'."""
+    ids = {str(sys_id) for config in platforms for sys_id in config.get("systems", {})}
+    for profile in profiles.values():
+        ids.update(str(sys_id) for sys_id in profile.get("systems") or [])
+        for entry in profile.get("files") or []:
+            if not isinstance(entry, dict):
+                continue
+            if entry.get("system"):
+                ids.add(str(entry["system"]))
+            ids.update(str(sys_id) for sys_id in entry.get("systems") or [])
+    return ids
+
+
+def link_coverage(ids: set[str]) -> tuple[int, int]:
+    """How many of ``ids`` lead to a section, and how many there are."""
+    return sum(1 for sys_id in ids if system_section(sys_id)), len(ids)
+
+
+# A system id -> the names the platforms' own files give it, built once.
+_search_aliases: dict[str, list[str]] = {}
+_search_aliases_built = [False]
+# A native name as a person types it: no space, no comma. RetroArch's
+# native names are No-Intro titles ("Sony - PlayStation") and stay out.
+_ALIAS_SHAPE = re.compile(r"[A-Za-z0-9][A-Za-z0-9_.+-]*")
+
+
+def system_search_aliases(sys_id: str) -> list[str]:
+    """The native system ids the platform YAMLs file ``sys_id`` under
+    (psx, dreamcast, gba), sorted, for search. Read from each platform's
+    own entries by common.platform_system_names, never typed by hand."""
+    if not _search_aliases_built[0]:
+        from common import (
+            list_registered_platforms,
+            load_platform_config,
+            platform_system_names,
+        )
+
+        platforms_dir = str(_REPO_ROOT / "platforms")
+        index: dict[str, set[str]] = {}
+        for platform in list_registered_platforms(platforms_dir, include_archived=True):
+            config = load_platform_config(platform, platforms_dir)
+            names = platform_system_names(config, list(config.get("systems", {})))
+            for owner, spelled in names.items():
+                index.setdefault(owner, set()).update(
+                    name for name in spelled if _ALIAS_SHAPE.fullmatch(name)
+                )
+        # Search ignores case: one spelling per word, the lower-case one
+        # when a platform writes it (psx, not PSX).
+        _search_aliases.clear()
+        for owner, names in index.items():
+            spellings: dict[str, str] = {}
+            for name in sorted(names, key=lambda n: (n != n.lower(), n)):
+                spellings.setdefault(name.casefold(), name)
+            if spellings:
+                _search_aliases[owner] = sorted(spellings.values(), key=str.casefold)
+        _search_aliases_built[0] = True
+    return list(_search_aliases.get(str(sys_id), []))

tests/test_site_r3_p0.py

@@ -0,0 +1,293 @@
+"""Every system id with a section leads to it (round 3, package P0).
+
+260 of 636 system ids linked: a variant (WonderSwan Color, Neo Geo MVS)
+never matched a heading, twin folders (DS and Nintendo DS) made the exact
+name ambiguous, and every Arcade heading was refused. siterender now reads
+a reviewed SYSTEM_PARENTS table after the id's own section, and accepts an
+Arcade heading that reads the name exactly. An id whose files the
+collection holds and that stays plain is named below, with the reason.
+Reads the repository: skips when the database is absent.
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import sys
+import unittest
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).resolve().parents[1]
+sys.path.insert(0, str(REPO_ROOT / "scripts"))
+
+import siterender  # noqa: E402
+
+# Ids whose files the collection holds and that lead to no section.
+UNLINKED: dict[str, str] = {
+    # Files only under a program's folder (Other/<emulator>, Arcade/<core>,
+    # Game Engines/<engine>, a top named after the program). The heading
+    # names the program, not the system.
+    "atari-system1": "Other > fbneo",
+    "bbc-master-128": "Other > beebem",
+    "bbc-master-et": "Other > beebem",
+    "capcom-cps3": "Arcade > FBNeo",
+    "cave-story": "Other > NXEngine",
+    "cps3": "Arcade > Arcade holds one CPS-3 set among every board's",
+    "commanderx16": "Arcade > Arcade",
+    "dark-forces-ii": "Game Engines > OpenJKDF2",
+    "deco-cassette": "Arcade > FBNeo",
+    "dos": "DOSBox, Roland and Other folders",
+    "fpinball": "Other > future-pinball-fploader",
+    "fruitmachine": "Other > mfme",
+    "igs-pgm": "Arcade > FBNeo",
+    "konami-bubsys": "Arcade > FBNeo",
+    "microsoft-xbox-360": "Other > xenia-edge",
+    "midway-mcr": "Arcade > FBNeo",
+    "mysteries-of-the-sith": "Game Engines > OpenJKDF2",
+    "mz2000": "Arcade > Arcade, not the Sharp MZ section",
+    "mz2500": "Arcade > Arcade, not the Sharp MZ section",
+    "mz800": "Arcade > Arcade, not the Sharp MZ section",
+    "namco-na1": "Arcade > FBNeo",
+    "namco-nb1": "Arcade > FBNeo",
+    "nmk16": "Arcade > FBNeo",
+    "omicron": "Other > omicron",
+    "pegasus": "Arcade > MAME",
+    "rpgmaker": "RPG Maker > easyrpg",
+    "scummvm": "Other > ScummVM",
+    "sdlpal": "Other > sdlpal",
+    "sega-system16b": "Arcade > FBNeo",
+    "sinclair-zx-spectrum-next": "Other > cspect",
+    "smbx": "Game Engines > TheXTech",
+    "solarus": "Game Engines > Solarus",
+    "sunplus-spmp8000": "Other > spmp8000emu",
+    "super-kaneko-nova": "Arcade > FBNeo",
+    "taito-cchip": "Arcade > FBNeo",
+    "thomson-mo6": "Other > Clock Signal",
+    "tomb-raider-1": "Game Engines > TRX",
+    "vgmplay": "Arcade > Arcade and MAME",
+    "vm-labs-nuon": "Other > nuance",
+    "vpinball": "Arcade > PinMAME",
+    "xrick": "Other > xrick",
+    "zc210": "ZC > ZC210, the emulator's folder",
+    # Files filed with another machine: a link would name the wrong one.
+    "atom": "atom.zip is filed with the BBC Micro",
+    "commodore-c64-supercpu": "VICE's tree under the C128",
+    "commodore-cbm-ii": "VICE's tree under the C128",
+    "commodore-pet": "VICE's tree under the C128",
+    "emerson-arcadia-2001": "ar_bios.zip is filed with the Saturn",
+    "naomi2": "naomi2.zip is filed with the Dreamcast",
+    "neogeo64": "the platform declares the Neo Geo's neogeo.zip",
+    "sega-dreamcast-arcade": "naomi.zip and awbios.zip, Arcade and Dreamcast",
+    "spectravideo": "the SVI ROMs are filed with the MSX",
+    "vvvvvv": "SDL's controller database",
+    "ym2608": "an FBNeo sound ROM, also filed with the NES",
+    # Files of distinct machines or games, one section each.
+    "catacomb-adventure": "three games, three sections",
+    "sinclair-zx80-81": "ZX 80 and ZX 81 sections",
+}
+
+
+def _accepts(fe: dict, local: str | None, status: str) -> bool:
+    """A resolution that proves the entry's content or its full path."""
+    from common import resolution_is_hash_exact
+
+    if not local:
+        return False
+    if resolution_is_hash_exact(status) or status == "path_exact":
+        return True
+    path = fe.get("path")
+    return bool(path) and local.endswith("/" + str(path))
+
+
+@unittest.skipUnless((REPO_ROOT / "database.json").exists(), "database.json is not built")
+class SystemLinks(unittest.TestCase):
+
+    @classmethod
+    def setUpClass(cls) -> None:
+        previous = os.getcwd()
+        os.chdir(REPO_ROOT)
+        try:
+            from common import (
+                list_registered_platforms,
+                load_emulator_profiles,
+                load_platform_config,
+                resolve_local_file,
+            )
+            from generate_site import _copy_places, _group_by_manufacturer
+
+            with open("database.json", encoding="utf-8") as handle:
+                db = json.load(handle)
+            cls.manufacturers = _group_by_manufacturer(db)
+            profiles = load_emulator_profiles("emulators", skip_aliases=False)
+            siterender.set_console_sections(
+                cls.manufacturers, siterender.program_names(profiles))
+            platforms = [
+                load_platform_config(name, "platforms")
+                for name in list_registered_platforms("platforms", include_archived=True)
+            ]
+            cls.ids = siterender.system_ids(profiles, platforms)
+
+            entries: dict[str, list[dict]] = {}
+            for config in platforms:
+                for sys_id, system in config.get("systems", {}).items():
+                    entries.setdefault(sys_id, []).extend(system.get("files") or [])
+            for profile in profiles.values():
+                systems = [str(s) for s in profile.get("systems") or []]
+                for fe in profile.get("files") or []:
+                    if not isinstance(fe, dict):
+                        continue
+                    owners = {str(fe["system"])} if fe.get("system") else set()
+                    owners.update(str(s) for s in fe.get("systems") or [])
+                    if not owners and len(systems) == 1:
+                        owners = set(systems)
+                    for owner in owners:
+                        entries.setdefault(owner, []).append(fe)
+
+            places = _copy_places(db)
+            by_path = {entry.get("path"): sha1 for sha1, entry in db["files"].items()}
+            # Id -> the (manufacturer, console) sections holding its files.
+            cls.held: dict[str, set[tuple[str, str]]] = {}
+            for sys_id in cls.ids:
+                sections: set[tuple[str, str]] = set()
+                for fe in entries.get(sys_id, []):
+                    local, status = resolve_local_file(fe, db)
+                    if not _accepts(fe, local, status):
+                        continue
+                    sha1 = by_path.get(local)
+                    if sha1 is not None:
+                        sections.update((m, c) for m, c, _ in places.get(sha1, set()))
+                if sections:
+                    cls.held[sys_id] = sections
+        finally:
+            os.chdir(previous)
+
+    def test_every_id_with_files_links(self) -> None:
+        linked, total = siterender.link_coverage(self.ids)
+        with_files = set(self.held)
+        print(f"\n  {linked} of {total} system ids linked, "
+              f"{len(with_files - set(UNLINKED))} of {len(with_files)} with files")
+        plain = sorted(
+            sys_id for sys_id in with_files - set(UNLINKED)
+            if siterender.system_section(sys_id) is None
+        )
+        self.assertEqual(plain, [], "link these or name them in UNLINKED with the reason")
+
+    def test_unlinked_list_is_current(self) -> None:
+        stale = sorted(
+            sys_id for sys_id in UNLINKED
+            if sys_id not in self.held or siterender.system_section(sys_id) is not None
+        )
+        self.assertEqual(stale, [])
+
+    def test_a_matching_heading_is_linked(self) -> None:
+        """An id whose name reads one heading, outside Other, leads there."""
+        by_heading: dict[str, set[tuple[str, str]]] = {}
+        for mfr, consoles in self.manufacturers.items():
+            for console in consoles:
+                for name in {siterender._heading_name(mfr, console), console}:
+                    by_heading.setdefault(name.casefold(), set()).add((mfr, console))
+        missed = []
+        for sys_id in sorted(self.ids):
+            places = by_heading.get(siterender.system_name(sys_id).casefold(), set())
+            if len(places) != 1:
+                continue
+            mfr, console = next(iter(places))
+            if mfr == "Other" or (
+                mfr == "Arcade" and console.casefold() in siterender._program_names
+            ):
+                continue
+            if siterender.system_section(sys_id) is None:
+                missed.append(sys_id)
+        self.assertEqual(missed, [])
+
+    def test_parents_are_sections(self) -> None:
+        sections = {
+            (mfr, console)
+            for mfr, consoles in self.manufacturers.items() for console in consoles
+        }
+        self.assertEqual(
+            sorted(k for k, v in siterender.SYSTEM_PARENTS.items() if v not in sections), [])
+
+    def test_parents_hold_the_files(self) -> None:
+        """A parent holds a file of the id, or of an id of the same name.
+        An entry its own section shadows is never read and is not judged."""
+        by_name: dict[str, set[tuple[str, str]]] = {}
+        for sys_id, sections in self.held.items():
+            by_name.setdefault(siterender.system_name(sys_id), set()).update(sections)
+        wrong = []
+        for sys_id, parent in sorted(siterender.SYSTEM_PARENTS.items()):
+            if siterender._own_section(sys_id) is not None:
+                continue
+            held = by_name.get(siterender.system_name(sys_id))
+            if held is not None and parent not in held:
+                wrong.append((sys_id, parent))
+        self.assertEqual(wrong, [])
+
+    def test_variants_lead_to_their_family(self) -> None:
+        expect = {
+            "wonderswan-color": ("bandai", "WonderSwan"),
+            "bandai-wswanc": ("bandai", "WonderSwan"),
+            "ngpc": ("snk", "Neo Geo Pocket"),
+            "snk-neogeo-mvs": ("snk", "Neo Geo"),
+            "snk-neogeo-aes": ("snk", "Neo Geo"),
+            "nintendo-ds": ("nintendo", "Nintendo DS"),
+            "nintendo-64dd": ("nintendo", "Nintendo 64DD"),
+            "zx-spectrum-128": ("sinclair", "ZX Spectrum"),
+            "pc-engine-ld": ("pioneer", "LaserActive"),
+        }
+        for sys_id, section in expect.items():
+            with self.subTest(sys_id=sys_id):
+                self.assertEqual(siterender.system_section(sys_id), section)
+
+    def test_arcade_heading_by_exact_name(self) -> None:
+        self.assertEqual(
+            siterender.system_section("namco-system246"), ("arcade", "Namco System 246"))
+        # No heading reads NAOMI: its sets are cards of the general section.
+        self.assertIsNone(siterender.system_section("sega-naomi"))
+        # Never a folder of copies a core loads.
+        self.assertIsNone(siterender.system_section("igs-pgm"))
+
+    def test_folder_votes_stay_out(self) -> None:
+        for sys_id in ("emerson-arcadia-2001", "naomi2", "commodore-pet"):
+            with self.subTest(sys_id=sys_id):
+                self.assertIsNone(siterender.system_section(sys_id))
+                self.assertNotIn(sys_id, siterender.SYSTEM_PARENTS)
+
+    def test_emulator_rows_link_the_variant(self) -> None:
+        from generate_site import _named_system_link
+
+        self.assertEqual(
+            _named_system_link("wonderswan-color", "../"),
+            "[Bandai WonderSwan Color](../systems/bandai.md#wonderswan)")
+
+
+class SearchAliases(unittest.TestCase):
+
+    def test_platform_names(self) -> None:
+        self.assertIn("psx", siterender.system_search_aliases("sony-playstation"))
+        self.assertIn("dc", siterender.system_search_aliases("sega-dreamcast"))
+        self.assertIn("gba", siterender.system_search_aliases("nintendo-gba"))
+
+    def test_one_typed_word_per_name(self) -> None:
+        siterender.system_search_aliases("sony-playstation")
+        for sys_id, names in siterender._search_aliases.items():
+            with self.subTest(sys_id=sys_id):
+                self.assertNotIn(sys_id, names)
+                self.assertEqual(len({n.casefold() for n in names}), len(names))
+                for name in names:
+                    self.assertNotRegex(name, r"[\s,]")
+
+    def test_unknown_id(self) -> None:
+        self.assertEqual(siterender.system_search_aliases("no-such-system"), [])
+
+
+class BuildCount(unittest.TestCase):
+
+    def test_generator_prints_the_count(self) -> None:
+        source = (REPO_ROOT / "scripts" / "generate_site.py").read_text(encoding="utf-8")
+        self.assertIn("link_coverage(", source)
+        self.assertIn("system ids linked", source)
+
+
+if __name__ == "__main__":
+    unittest.main()

wiki/testing-guide.md

@@ -54,7 +54,7 @@ Every module in `tests/`, by the area it guards.
 | Installer and downloads | loopback server | `test_check_release`, `test_download`, `test_install`, `test_install_cores`, `test_install_experience`, `test_install_offline`, `test_install_selection` |
 | Handheld port | stubbed PortMaster dialogs, fake installer | `test_port` |
 | Truth and native export | synthetic + cached originals | `test_export_counts`, `test_export_native_layer`, `test_export_report`, `test_exporters`, `test_truth_data_dir_readers`, `test_truth_diff`, `test_truth_layout` |
-| Scrapers, upstream and provenance | inline sources, synthetic | `test_batocera_md5_list`, `test_check_freshness`, `test_contributing_text`, `test_coverage_cache`, `test_data_dir_refs`, `test_dataclass_defaults`, `test_dedup`, `test_doc_pages`, `test_doc_surfaces`, `test_elec_scrapers`, `test_fbneo_parser`, `test_find_pinned_revision`, `test_hash_merge`, `test_launchers`, `test_launchers_emudeck`, `test_launchers_recalbox`, `test_launchers_retrobat`, `test_launchers_retrodeck`, `test_launchers_retropie`, `test_launchers_rocknix`, `test_mame_devices`, `test_mame_parser`, `test_mame_ref_audit`, `test_optional_dependencies`, `test_profile_sync`, `test_prose_style`, `test_provenance`, `test_refresh_data_dirs`, `test_refresh_failures`, `test_refresh_stale`, `test_retrodeck_scraper_errors`, `test_romset_recipes`, `test_schema_semantic_gate`, `test_scraper_contract`, `test_site_availability`, `test_site_content_data`, `test_site_copy`, `test_site_cores`, `test_site_emulator_page`, `test_site_exports`, `test_site_funnel`, `test_site_images`, `test_site_picker`, `test_site_platform_hashes`, `test_site_r2_components`, `test_site_r2_css`, `test_site_r2_emulators`, `test_site_r2_entry`, `test_site_r2_gate`, `test_site_r2_nav`, `test_site_r2_p0`, `test_site_r2_reference`, `test_site_r2_systems`, `test_site_r2_wiki`, `test_site_render`, `test_site_system_links`, `test_site_validation`, `test_site_weight`, `test_site_write`, `test_upstream`, `test_upstream_index`, `test_validate_pr`, `test_validate_pr_flow`, `test_validate_pr_hashes`, `test_workflow_paths` |
+| Scrapers, upstream and provenance | inline sources, synthetic | `test_batocera_md5_list`, `test_check_freshness`, `test_contributing_text`, `test_coverage_cache`, `test_data_dir_refs`, `test_dataclass_defaults`, `test_dedup`, `test_doc_pages`, `test_doc_surfaces`, `test_elec_scrapers`, `test_fbneo_parser`, `test_find_pinned_revision`, `test_hash_merge`, `test_launchers`, `test_launchers_emudeck`, `test_launchers_recalbox`, `test_launchers_retrobat`, `test_launchers_retrodeck`, `test_launchers_retropie`, `test_launchers_rocknix`, `test_mame_devices`, `test_mame_parser`, `test_mame_ref_audit`, `test_optional_dependencies`, `test_profile_sync`, `test_prose_style`, `test_provenance`, `test_refresh_data_dirs`, `test_refresh_failures`, `test_refresh_stale`, `test_retrodeck_scraper_errors`, `test_romset_recipes`, `test_schema_semantic_gate`, `test_scraper_contract`, `test_site_availability`, `test_site_content_data`, `test_site_copy`, `test_site_cores`, `test_site_emulator_page`, `test_site_exports`, `test_site_funnel`, `test_site_images`, `test_site_picker`, `test_site_platform_hashes`, `test_site_r2_components`, `test_site_r2_css`, `test_site_r2_emulators`, `test_site_r2_entry`, `test_site_r2_gate`, `test_site_r2_nav`, `test_site_r2_p0`, `test_site_r2_reference`, `test_site_r2_systems`, `test_site_r2_wiki`, `test_site_r3_p0`, `test_site_render`, `test_site_system_links`, `test_site_validation`, `test_site_weight`, `test_site_write`, `test_upstream`, `test_upstream_index`, `test_validate_pr`, `test_validate_pr_flow`, `test_validate_pr_hashes`, `test_workflow_paths` |
 
 ## Test architecture