Files
sub-provider/app/services/conf_loader.py
2026-04-21 16:56:06 +08:00

361 lines
15 KiB
Python

from __future__ import annotations
import csv
import os
import re
from pathlib import Path
from typing import Any
from app.conf_models import (
ConfBaseConfig,
ConfBuiltin,
ConfGroup,
ConfInlineRule,
ConfModule,
ConfPolicyOverride,
ConfProfileConfig,
ConfSelector,
ConfSource,
)
_ENV_PATTERN = re.compile(r"\$\{([A-Z0-9_]+)\}")
_SCALAR_KEYS = {
"listen",
"output_dir",
"cache_dir",
"mode",
"allow_lan",
"log_level",
"ipv6",
"append_userinfo_header",
"userinfo_source_policy",
}
_BOOL_KEYS = {"allow_lan", "ipv6", "append_userinfo_header"}
_MULTI_KEYS = {"source", "selector", "group", "module", "builtin"}
_SOURCE_TYPES = {"url", "file", "inline", "base64"}
_PROFILE_SCALAR_KEYS = {
"name",
"enabled",
"sources",
"include_modules",
"exclude_modules",
"include_builtins",
}
_PROFILE_MULTI_KEYS = {"override_policy", "prepend_rule", "append_rule"}
class ConfConfigError(ValueError):
def __init__(self, path: Path, line_no: int, message: str) -> None:
self.path = path
self.line_no = line_no
self.message = message
super().__init__(f"{path}:{line_no}: {message}")
def _expand_env(value: str) -> str:
return _ENV_PATTERN.sub(lambda match: os.getenv(match.group(1), ""), value)
def _parse_bool(value: str, *, path: Path, line_no: int, key: str) -> bool:
normalized = value.strip().lower()
mapping = {
"true": True,
"yes": True,
"1": True,
"false": False,
"no": False,
"0": False,
}
if normalized not in mapping:
raise ConfConfigError(path, line_no, f"invalid boolean for {key}: {value}")
return mapping[normalized]
def _parse_csv_fields(value: str, *, path: Path, line_no: int) -> list[str]:
try:
row = next(csv.reader([value], skipinitialspace=True))
except Exception as exc: # noqa: BLE001
raise ConfConfigError(path, line_no, f"invalid csv payload: {exc}") from exc
return [field.strip() for field in row]
def _split_key_value(raw_line: str, *, path: Path, line_no: int) -> tuple[str, str]:
if "=" not in raw_line:
raise ConfConfigError(path, line_no, "expected key = value")
key, value = [part.strip() for part in raw_line.split("=", 1)]
if not key:
raise ConfConfigError(path, line_no, "key cannot be empty")
return key, value
def _iter_conf_lines(path: Path) -> list[tuple[int, str]]:
records: list[tuple[int, str]] = []
for line_no, raw_line in enumerate(path.read_text(encoding="utf-8").splitlines(), start=1):
line = raw_line.strip()
if not line or line.startswith("#") or line.startswith(";"):
continue
records.append((line_no, raw_line))
return records
def _ensure_named_records(records: list[Any], *, key: str, path: Path) -> None:
seen: dict[str, int] = {}
for record in records:
name = getattr(record, key)
line_no = getattr(record, "line_no")
previous_line = seen.get(name)
if previous_line is not None:
raise ConfConfigError(path, line_no, f"duplicate {record.__class__.__name__} key '{name}', first defined at line {previous_line}")
seen[name] = line_no
def _parse_source(value: str, *, path: Path, line_no: int) -> ConfSource:
fields = _parse_csv_fields(_expand_env(value), path=path, line_no=line_no)
if len(fields) < 3:
raise ConfConfigError(path, line_no, "source requires at least 3 fields: key,type,value")
key, source_type, source_value, *extras = fields
if not key:
raise ConfConfigError(path, line_no, "source key cannot be empty")
if source_type not in _SOURCE_TYPES:
raise ConfConfigError(path, line_no, f"unsupported source type: {source_type}")
options: dict[str, str] = {}
enabled = True
cache_ttl: int | None = None
for item in extras:
if "=" not in item:
raise ConfConfigError(path, line_no, f"source option must be k=v: {item}")
option_key, option_value = [part.strip() for part in item.split("=", 1)]
if not option_key:
raise ConfConfigError(path, line_no, "source option key cannot be empty")
option_value = _expand_env(option_value)
if option_key == "enabled":
enabled = _parse_bool(option_value, path=path, line_no=line_no, key="source.enabled")
elif option_key == "cache_ttl":
try:
cache_ttl = int(option_value)
except ValueError as exc:
raise ConfConfigError(path, line_no, f"invalid integer for source.cache_ttl: {option_value}") from exc
else:
options[option_key] = option_value
return ConfSource(
key=key,
source_type=source_type,
value=source_value,
enabled=enabled,
cache_ttl=cache_ttl,
options=options,
line_no=line_no,
)
def _parse_selector(value: str, *, path: Path, line_no: int) -> ConfSelector:
fields = _parse_csv_fields(value, path=path, line_no=line_no)
if len(fields) != 2:
raise ConfConfigError(path, line_no, "selector requires exactly 2 fields: key,regex")
key, regex = fields
if not key or not regex:
raise ConfConfigError(path, line_no, "selector key and regex cannot be empty")
return ConfSelector(key=key, regex=regex, line_no=line_no)
def _parse_group(value: str, *, path: Path, line_no: int) -> ConfGroup:
segments = [segment.strip() for segment in value.split("`")]
if len(segments) < 2:
raise ConfConfigError(path, line_no, "group requires at least name`type")
name, group_type, *tokens = segments
if not name or not group_type:
raise ConfConfigError(path, line_no, "group name and type cannot be empty")
return ConfGroup(name=name, group_type=group_type, tokens=tokens, raw=value.strip(), line_no=line_no)
def _parse_module(value: str, *, path: Path, line_no: int, config_dir: Path) -> ConfModule:
fields = _parse_csv_fields(value, path=path, line_no=line_no)
if len(fields) != 5:
raise ConfConfigError(path, line_no, "module requires exactly 5 fields: key,path,policy,order,enabled")
key, module_path, policy, order_text, enabled_text = fields
if not key or not module_path or not policy:
raise ConfConfigError(path, line_no, "module key, path, and policy cannot be empty")
try:
order = int(order_text)
except ValueError as exc:
raise ConfConfigError(path, line_no, f"invalid integer for module.order: {order_text}") from exc
enabled = _parse_bool(enabled_text, path=path, line_no=line_no, key="module.enabled")
resolved_path = (config_dir / module_path).resolve()
if not resolved_path.is_file():
raise ConfConfigError(path, line_no, f"module path does not exist: {module_path}")
return ConfModule(key=key, path=module_path, policy=policy, order=order, enabled=enabled, line_no=line_no)
def _parse_builtin(value: str, *, path: Path, line_no: int) -> ConfBuiltin:
fields = _parse_csv_fields(value, path=path, line_no=line_no)
if len(fields) != 6:
raise ConfConfigError(path, line_no, "builtin requires exactly 6 fields: key,type,value,policy,order,enabled")
key, builtin_type, builtin_value, policy, order_text, enabled_text = fields
if builtin_type not in {"GEOIP", "FINAL"}:
raise ConfConfigError(path, line_no, f"unsupported builtin type: {builtin_type}")
if not key or not policy:
raise ConfConfigError(path, line_no, "builtin key and policy cannot be empty")
if builtin_type == "GEOIP" and not builtin_value:
raise ConfConfigError(path, line_no, "builtin GEOIP requires a value")
try:
order = int(order_text)
except ValueError as exc:
raise ConfConfigError(path, line_no, f"invalid integer for builtin.order: {order_text}") from exc
enabled = _parse_bool(enabled_text, path=path, line_no=line_no, key="builtin.enabled")
return ConfBuiltin(
key=key,
builtin_type=builtin_type,
value=builtin_value,
policy=policy,
order=order,
enabled=enabled,
line_no=line_no,
)
def _parse_string_list(value: str, *, path: Path, line_no: int) -> list[str]:
fields = _parse_csv_fields(_expand_env(value), path=path, line_no=line_no)
return [field for field in fields if field]
def _parse_override_policy(value: str, *, path: Path, line_no: int) -> ConfPolicyOverride:
fields = _parse_csv_fields(value, path=path, line_no=line_no)
if len(fields) != 2:
raise ConfConfigError(path, line_no, "override_policy requires exactly 2 fields: module_key,new_policy")
module_key, policy = fields
if not module_key or not policy:
raise ConfConfigError(path, line_no, "override_policy module key and policy cannot be empty")
return ConfPolicyOverride(module_key=module_key, policy=policy, line_no=line_no)
def _parse_inline_rule(value: str, *, path: Path, line_no: int, key: str) -> ConfInlineRule:
expanded = _expand_env(value).strip()
if not expanded:
raise ConfConfigError(path, line_no, f"{key} cannot be empty")
return ConfInlineRule(value=expanded, line_no=line_no)
def _validate_profile_references(
profile: ConfProfileConfig,
base: ConfBaseConfig,
*,
path: Path,
field_lines: dict[str, int],
) -> None:
source_keys = {item.key for item in base.sources}
module_keys = {item.key for item in base.modules}
builtin_keys = {item.key for item in base.builtins}
for source in profile.sources:
if source not in source_keys:
raise ConfConfigError(path, field_lines.get("sources", 1), f"profile references unknown source: {source}")
for module in profile.include_modules:
if module not in module_keys:
raise ConfConfigError(
path,
field_lines.get("include_modules", 1),
f"profile references unknown module in include_modules: {module}",
)
for module in profile.exclude_modules:
if module not in module_keys:
raise ConfConfigError(
path,
field_lines.get("exclude_modules", 1),
f"profile references unknown module in exclude_modules: {module}",
)
for builtin in profile.include_builtins:
if builtin not in builtin_keys:
raise ConfConfigError(
path,
field_lines.get("include_builtins", 1),
f"profile references unknown builtin in include_builtins: {builtin}",
)
for override in profile.override_policies:
if override.module_key not in module_keys:
raise ConfConfigError(path, override.line_no, f"override_policy references unknown module: {override.module_key}")
def load_conf_base(path: Path) -> ConfBaseConfig:
config = ConfBaseConfig()
scalar_values: dict[str, Any] = {}
config_dir = path.resolve().parent
for line_no, raw_line in _iter_conf_lines(path):
key, value = _split_key_value(raw_line, path=path, line_no=line_no)
if key in _SCALAR_KEYS:
if key in scalar_values:
raise ConfConfigError(path, line_no, f"duplicate scalar key: {key}")
expanded_value = _expand_env(value)
if key in _BOOL_KEYS:
scalar_values[key] = _parse_bool(expanded_value, path=path, line_no=line_no, key=key)
else:
scalar_values[key] = expanded_value
continue
if key not in _MULTI_KEYS:
raise ConfConfigError(path, line_no, f"unknown directive: {key}")
if key == "source":
config.sources.append(_parse_source(value, path=path, line_no=line_no))
elif key == "selector":
config.selectors.append(_parse_selector(value, path=path, line_no=line_no))
elif key == "group":
config.groups.append(_parse_group(value, path=path, line_no=line_no))
elif key == "module":
config.modules.append(_parse_module(value, path=path, line_no=line_no, config_dir=config_dir))
elif key == "builtin":
config.builtins.append(_parse_builtin(value, path=path, line_no=line_no))
for key, parsed_value in scalar_values.items():
setattr(config, key, parsed_value)
_ensure_named_records(config.sources, key="key", path=path)
_ensure_named_records(config.selectors, key="key", path=path)
_ensure_named_records(config.groups, key="name", path=path)
_ensure_named_records(config.modules, key="key", path=path)
_ensure_named_records(config.builtins, key="key", path=path)
return config
def load_conf_profile(path: Path, *, base_config: ConfBaseConfig) -> ConfProfileConfig:
profile = ConfProfileConfig()
scalar_values: dict[str, Any] = {}
field_lines: dict[str, int] = {}
for line_no, raw_line in _iter_conf_lines(path):
key, value = _split_key_value(raw_line, path=path, line_no=line_no)
if key in _PROFILE_SCALAR_KEYS:
if key in scalar_values:
raise ConfConfigError(path, line_no, f"duplicate scalar key: {key}")
field_lines[key] = line_no
if key == "enabled":
scalar_values[key] = _parse_bool(_expand_env(value), path=path, line_no=line_no, key=key)
elif key in {"sources", "include_modules", "exclude_modules", "include_builtins"}:
scalar_values[key] = _parse_string_list(value, path=path, line_no=line_no)
else:
scalar_values[key] = _expand_env(value)
continue
if key not in _PROFILE_MULTI_KEYS:
raise ConfConfigError(path, line_no, f"unknown directive: {key}")
if key == "override_policy":
profile.override_policies.append(_parse_override_policy(value, path=path, line_no=line_no))
elif key == "prepend_rule":
profile.prepend_rules.append(_parse_inline_rule(value, path=path, line_no=line_no, key=key))
elif key == "append_rule":
profile.append_rules.append(_parse_inline_rule(value, path=path, line_no=line_no, key=key))
for key, parsed_value in scalar_values.items():
setattr(profile, key, parsed_value)
_validate_profile_references(profile, base_config, path=path, field_lines=field_lines)
return profile
def load_conf_profiles(directory: Path, *, base_config: ConfBaseConfig) -> dict[str, ConfProfileConfig]:
profiles: dict[str, ConfProfileConfig] = {}
for path in sorted(directory.glob("*.conf")):
profile = load_conf_profile(path, base_config=base_config)
profile_key = path.stem
if profile_key in profiles:
raise ConfConfigError(path, 1, f"duplicate profile file key: {profile_key}")
profiles[profile_key] = profile
return profiles