Files
libretro/scripts/scraper/rocknix_scraper.py
T
Abdessamad Derraz a7e757598f fix: list the libretro cores rocknix installs
ROCKNIX listed retroarch, which stood for every libretro profile (351
resolved), while its devices build 124 libretro packages. The platform
list now names the cores those packages install on at least one device.

The targets scraper read a package's name as its core: vice-lr installs
vice_x128, vice_x64, vice_xpet, vice_xplus4 and vice_xvic, which reached
no device. It now reads each package's makeinstall_target, the file
copied to ${INSTALL}/usr/lib/libretro under its target name, through
TARGET_ARCH and ENABLE_32BIT (pcsx_rearmed-lr, desmume-lr). 155
profiles resolve, each built by a device, and no device loses one.
2026-11-09 21:27:51 +01:00

337 lines
13 KiB
Python

#!/usr/bin/env python3
"""Scraper for ROCKNIX BIOS requirements.
Source: projects/ROCKNIX/packages/rocknix/sources/scripts/rocknix-systems
Format: Python source, a `systems` mapping of per-system BIOS declarations
Hash: MD5
ROCKNIX verification logic (rocknix-systems, checkBios):
- md5sum() of the file on disk against the declared md5
- checkInsideZip() when zippedFile is set, with an optional altmd5
- an empty md5 means the check is presence only
EmulationStation surfaces the result through ApiSystem::getBiosInformations,
which shells out to `rocknix-systems` (es-app/src/ApiSystem.cpp:486-516).
"""
from __future__ import annotations
import re
try:
from .base_scraper import (
BaseScraper,
BiosRequirement,
fetch_github_latest_version,
md5_field,
requirement_entry,
)
from .rocknix_launchers import extract_launchers
from .shell_launchers import RepoTree
except ImportError:
from base_scraper import (
BaseScraper,
BiosRequirement,
fetch_github_latest_version,
md5_field,
requirement_entry,
)
from rocknix_launchers import extract_launchers
from shell_launchers import RepoTree
PLATFORM_NAME = "rocknix"
GITHUB_REPO = "ROCKNIX/distribution"
SOURCE_TEMPLATE = (
"https://raw.githubusercontent.com/ROCKNIX/distribution/{ref}/"
"projects/ROCKNIX/packages/rocknix/sources/scripts/rocknix-systems"
)
# The branch a test or a manual run can still name. A scrape reads the tag
# of the release its YAML is labelled with.
SOURCE_URL = SOURCE_TEMPLATE.format(ref="next")
EMULATORS_MK_URL = (
"https://raw.githubusercontent.com/ROCKNIX/distribution/{ref}"
"/projects/ROCKNIX/packages/virtual/emulators/package.mk"
)
STANDALONE_TREE_URL = (
"https://api.github.com/repos/ROCKNIX/distribution/contents/"
"projects/ROCKNIX/packages/emulators/standalone?ref={ref}"
)
def launcher_revision(url: str) -> str:
"""The tag or commit a rocknix-systems URL is read at."""
found = re.search(r"/ROCKNIX/distribution/([^/]+)/projects/", url)
if found is None:
raise ValueError(f"{url} names no ROCKNIX/distribution revision")
return found.group(1)
def standalone_emulators(package_names: list[str]) -> list[str]:
"""The standalone emulators ROCKNIX packages, by the names profiles use.
A package is named after its emulator with an ``-sa`` suffix
(``duckstation-sa``), once without its hyphen (``hatarisa``), and both
spellings are kept so a profile listing either is found.
"""
names: set[str] = set()
for package in package_names:
names.add(package)
stripped = _emulator_of(package)
if stripped != package:
names.add(stripped)
return sorted(names)
def _emulator_of(package: str) -> str:
return re.sub(r"-?sa$", "", package) or package
def dual_mode_emulators(standalone: list[str], package_mk: str) -> list[str]:
"""The standalone emulators RetroArch also runs as a core.
package.mk adds each launch option with `add_emu_core <system> <emulator>
<core> <default>`. Only an emulator also added as `retroarch <core>` is
launched both ways: ARMSX2 is added as armsx2-sa alone.
"""
retroarch = set(re.findall(r"^\s*add_emu_core\s+\S+\s+retroarch\s+(\S+)", package_mk, re.M))
return sorted(name for name in standalone if _emulator_of(name) in retroarch)
# ROCKNIX system key -> retrobios system ID
SYSTEM_SLUG_MAP = {
"xbox": "microsoft-xbox",
"gamecube": "nintendo-gamecube",
"nds": "nintendo-ds",
"gb": "nintendo-gb",
"gba": "nintendo-gba",
"gbc": "nintendo-gbc",
"dreamcast": "sega-dreamcast",
"gamegear": "sega-game-gear",
"mastersystem": "sega-master-system",
"segacd": "sega-mega-cd",
"saturn": "sega-saturn",
"psx": "sony-playstation",
"ps2": "sony-playstation-2",
"ps3": "sony-playstation-3",
"psp": "sony-psp",
"3do": "3do",
"amiga": "commodore-amiga",
"atari5200": "atari-5200",
"atari7800": "atari-7800",
"atarilynx": "atari-lynx",
"colecovision": "coleco-colecovision",
"intellivision": "mattel-intellivision",
"msx": "microsoft-msx",
"neogeo": "snk-neogeo",
"neogeocd": "snk-neogeo-cd",
"pcengine": "nec-pc-engine",
"pcenginecd": "nec-pc-engine-cd",
"x68000": "sharp-x68000",
"zxspectrum": "sinclair-zx-spectrum",
}
_SYSTEM_HEADER = re.compile(
r'"(?P<key>[\w\-]+)"\s*:\s*\{\s*"name"\s*:\s*"(?P<name>[^"]+)"\s*,'
r'\s*"biosFiles"\s*:\s*\[',
)
_FIELD = re.compile(r'"(\w+)"\s*:\s*"([^"]*)"')
# A field written as a list of strings (`"md5": [ "a", "b" ]`), the form
# batocera.linux 6b08b873 gives the md5 of a file several dumps satisfy.
_LIST_FIELD = re.compile(r'"(\w+)"\s*:\s*\[([^\[\]]*)\]')
_LIST_ITEM = re.compile(r'"([^"]*)"')
def _slice_bios_block(source: str, start: int) -> tuple[str, int]:
"""Return the biosFiles list body starting at an opening bracket."""
depth = 0
for i in range(start, len(source)):
char = source[i]
if char == "[":
depth += 1
elif char == "]":
depth -= 1
if depth == 0:
return source[start + 1:i], i
raise ValueError("unterminated biosFiles list")
def parse_systems(source: str) -> dict[str, dict]:
"""Extract per-system BIOS declarations from the rocknix-systems source.
Parsed textually rather than evaluated: the file is upstream code, and
the declarations are plain string literals.
"""
source = re.sub(r"#[^\n]*", "", source)
systems: dict[str, dict] = {}
for header in _SYSTEM_HEADER.finditer(source):
body, _ = _slice_bios_block(source, header.end() - 1)
files = []
for raw_entry in re.findall(r"\{[^{}]*\}", body):
fields: dict[str, str | list[str]] = dict(_FIELD.findall(raw_entry))
for key, items in _LIST_FIELD.findall(raw_entry):
fields[key] = _LIST_ITEM.findall(items)
if fields.get("file"):
files.append(fields)
systems[header.group("key")] = {
"name": header.group("name"),
"biosFiles": files,
}
if not systems:
raise ValueError("no system declarations found in rocknix-systems")
return systems
class Scraper(BaseScraper):
"""Scraper for the ROCKNIX rocknix-systems script."""
def __init__(self, url: str | None = None) -> None:
self.tag = None if url else fetch_github_latest_version(GITHUB_REPO)
super().__init__(url=url or SOURCE_TEMPLATE.format(ref=self.tag))
def _built_cores(self, standalone: list[str]) -> tuple[list[str], list[str]]:
"""(cores, standalone cores) at least one device image carries.
RetroArch runs the libretro cores the packages of its emulators
metapackage install per device
(projects/ROCKNIX/packages/virtual/emulators/package.mk), read by
the targets scraper, where every libretro profile counted before. A
standalone package no device carries is not run either.
"""
try:
from .targets.rocknix_targets_scraper import Scraper as DeviceTargets
except ImportError:
from scraper.targets.rocknix_targets_scraper import Scraper as DeviceTargets
builds = DeviceTargets().device_builds(self.tag or launcher_revision(self.url))
libretro = sorted({core for _arch, lr, _sa in builds.values() for core in lr})
carried = set(standalone_emulators(sorted({p for *_rest, sa in builds.values() for p in sa})))
standalone = [name for name in standalone if name in carried]
return sorted({*libretro, *standalone}), standalone
def _fetch_standalone(self) -> list[str]:
"""The packages of projects/ROCKNIX/packages/emulators/standalone."""
import json
import urllib.error
import urllib.request
from .base_scraper import github_headers
url = STANDALONE_TREE_URL.format(ref=self.tag or "next")
req = urllib.request.Request(url, headers=github_headers())
try:
with urllib.request.urlopen(req, timeout=30) as resp:
listing = json.loads(resp.read())
except (urllib.error.URLError, json.JSONDecodeError) as e:
raise ConnectionError(f"Failed to list {url}: {e}") from e
packages = [item["name"] for item in listing if item.get("type") == "dir"]
if not packages:
raise RuntimeError(f"no standalone emulator package at {url}")
return standalone_emulators(packages)
def _fetch_emulators_mk(self) -> str:
"""The package.mk that adds each emulator and core to es_systems."""
import urllib.error
import urllib.request
url = EMULATORS_MK_URL.format(ref=self.tag or "next")
try:
with urllib.request.urlopen(url, timeout=30) as resp:
text = resp.read().decode("utf-8")
except urllib.error.URLError as e:
raise ConnectionError(f"Failed to fetch {url}: {e}") from e
if "add_emu_core" not in text:
raise RuntimeError(f"no add_emu_core line in {url}")
return text
def validate_format(self, raw_data: str) -> bool:
"""Validate the rocknix-systems format."""
has_dict = re.search(r"^systems\s*=\s*\{", raw_data, re.MULTILINE) is not None
has_check = "def checkBios" in raw_data
return has_dict and has_check and '"biosFiles"' in raw_data and '"md5"' in raw_data
def fetch_requirements(self) -> list[BiosRequirement]:
"""Parse rocknix-systems and return BIOS requirements."""
if not self.validate_format(self._fetch_raw()):
raise ValueError("rocknix-systems format validation failed")
requirements = []
for key, entry in parse_systems(self._fetch_raw()).items():
system_slug = SYSTEM_SLUG_MAP.get(key, key)
for bios in entry["biosFiles"]:
path = bios["file"]
# Declared paths are relative to the storage root: strip the
# leading bios/ so destinations match base_destination
destination = path[len("bios/"):] if path.startswith("bios/") else path
values, as_list = md5_field(bios.get("md5"))
native: dict[str, object] = {"native_name": entry["name"]}
if as_list:
native["md5_list"] = True
req = BiosRequirement(
name=destination.rsplit("/", 1)[-1],
system=system_slug,
md5=",".join(values) or None,
alt_md5=bios.get("altmd5", "") or None,
destination=destination,
required=True,
native_id=key,
native=native,
)
if bios.get("zippedFile"):
req.zipped_file = bios["zippedFile"]
requirements.append(req)
return requirements
def generate_platform_yaml(self) -> dict:
"""Build the ROCKNIX platform configuration."""
raw = parse_systems(self._fetch_raw())
names = {key: entry["name"] for key, entry in raw.items()}
systems: dict[str, dict] = {}
for req in self.fetch_requirements():
entry = systems.setdefault(
req.system,
{
"files": [],
"native_id": req.native_id,
"name": names.get(req.native_id, req.system),
},
)
entry["files"].append(requirement_entry(req))
cores, standalone = self._built_cores(self._fetch_standalone())
return {
"platform": "ROCKNIX",
"version": self.tag or fetch_github_latest_version(GITHUB_REPO),
"homepage": "https://rocknix.org",
"source": self.url,
"base_destination": "bios",
"hash_type": "md5",
"verification_mode": "md5",
"cores": cores,
"standalone_cores": standalone,
# Launched both ways only where package.mk also adds the
# emulator as a RetroArch core.
"dual_mode_cores": dual_mode_emulators(standalone, self._fetch_emulators_mk()),
"systems": systems,
# Read at the revision the BIOS list comes from.
"launchers": extract_launchers(
RepoTree.github(GITHUB_REPO, launcher_revision(self.url))
),
}
def main() -> None:
"""Entry point."""
try:
from .base_scraper import scraper_cli
except ImportError:
from base_scraper import scraper_cli
scraper_cli(Scraper, "Scrape ROCKNIX BIOS requirements")
if __name__ == "__main__":
main()