Initial SQL-only 1C adapter baseline

This commit is contained in:
2026-07-22 03:03:47 +03:00
commit e2503b77e7
545 changed files with 184711 additions and 0 deletions
+534
View File
@@ -0,0 +1,534 @@
#!/usr/bin/env python3
"""Resolve a 1C fact against an explicit current-configuration source.
This command is intentionally narrower than RAG: it verifies whether a named
1C object member exists in the provided configuration evidence and returns the
source/provenance. It does not search examples unless a snapshot path is passed
explicitly by the caller.
"""
from __future__ import annotations
import argparse
import base64
import json
import re
import sys
from pathlib import Path
from typing import Any
sys.path.insert(0, str(Path(__file__).resolve().parent))
from get_1c_object_brief_context import view_forms, view_modules # noqa: E402
from get_1c_object_metadata import build_object_metadata # noqa: E402
from resolve_1c_object import canonical_kind, load_json, normalize, parse_query, resolve_object # noqa: E402
KIND_TO_RU = {
"Catalog": "Справочник",
"Document": "Документ",
"Enum": "Перечисление",
"InformationRegister": "РегистрСведений",
"AccumulationRegister": "РегистрНакопления",
"AccountingRegister": "РегистрБухгалтерии",
"Report": "Отчет",
"DataProcessor": "Обработка",
"CommonModule": "ОбщийМодуль",
"Form": "Форма",
"Template": "Макет",
"Command": "Команда",
}
SNAPSHOT_KIND_ALIASES = {
"Catalog": {"catalog", "справочник"},
"Document": {"document", "документ"},
"Enum": {"enum", "перечисление"},
"InformationRegister": {"register", "information_register", "регистрсведений", "регистр сведений"},
"AccumulationRegister": {"register", "accumulation_register", "регистрнакопления", "регистр накопления"},
"Report": {"report", "отчет"},
"DataProcessor": {"processing", "processor", "обработка"},
"CommonModule": {"common_module", "общиймодуль", "общий модуль"},
"Form": {"form", "форма"},
}
MEMBER_AREAS = {
"attribute": {"attribute", "attr", "реквизит", "реквизиты"},
"tabular_section": {"tabular_section", "table", "табличнаячасть", "табличная часть", "тч"},
"tabular_section_attribute": {"tabular_section_attribute", "table_attribute", "реквизиттч", "реквизит табличной части"},
"form": {"form", "форма", "формы"},
"command": {"command", "команда", "команды"},
"module": {"module", "модуль", "модули"},
"any": {"any", "любой", "все"},
}
def decode_arg(value: str | None, encoded: str | None) -> str | None:
if encoded:
return base64.b64decode(encoded).decode("utf-8")
return value
def compact_text(value: Any) -> str:
return re.sub(r"[\s._-]+", "", str(value or "")).casefold()
def member_area(value: str | None) -> str:
if not value:
return "any"
wanted = compact_text(value)
for area, aliases in MEMBER_AREAS.items():
if wanted in {compact_text(alias) for alias in aliases}:
return area
return value
def split_fact_path(raw: str) -> tuple[str | None, str, str | None, str | None]:
"""Parse common 1C fact paths.
Supported examples:
- Справочник.Номенклатура
- Справочник.Номенклатура.Артикул
- Справочник.Номенклатура.Цены.Цена
"""
parts = [part.strip() for part in str(raw or "").split(".") if part.strip()]
if len(parts) >= 2:
kind = parts[0]
object_name = parts[1]
if len(parts) == 2:
return kind, object_name, None, None
if len(parts) == 3:
return kind, object_name, None, parts[2]
return kind, object_name, parts[2], parts[3]
if len(parts) == 1:
return None, parts[0], None, None
raise ValueError("fact path is empty")
def path_join(*parts: Any) -> str:
return ".".join(str(part).strip() for part in parts if str(part or "").strip())
def object_canonical_path(kind: Any, name: Any) -> str:
normalized_kind = canonical_kind(str(kind or "")) or str(kind or "")
kind_text = KIND_TO_RU.get(normalized_kind, normalized_kind)
return path_join(kind_text, name)
def member_canonical_path(object_path: str, *, table_section: Any = None, member: Any = None) -> str:
return path_join(object_path, table_section, member)
def same_name(left: Any, right: Any) -> bool:
return normalize(str(left or "")) == normalize(str(right or ""))
def matches_name(item: dict[str, Any], wanted: str) -> bool:
return same_name(item.get("name"), wanted) or same_name(item.get("synonym"), wanted)
def nearest(items: list[dict[str, Any]], wanted: str, limit: int = 8) -> list[dict[str, Any]]:
wanted_norm = normalize(wanted)
result = []
for item in items:
name = str(item.get("name") or "")
synonym = str(item.get("synonym") or "")
if wanted_norm and (wanted_norm in normalize(name) or wanted_norm in normalize(synonym) or normalize(name) in wanted_norm):
result.append(compact_member(item))
return result[:limit]
def compact_member(item: dict[str, Any], *, area: str | None = None) -> dict[str, Any]:
payload = {
"area": area,
"name": item.get("name"),
"synonym": item.get("synonym"),
"type": item.get("type"),
"types": item.get("types"),
"canonical_types": item.get("canonical_types"),
"origin": item.get("origin"),
"effective_action": item.get("effective_action"),
"uuid": item.get("uuid"),
}
return {key: value for key, value in payload.items() if value not in (None, [], "")}
def with_path_fields(
member_payload: dict[str, Any],
*,
object_path: str,
path_kind: str,
table_section: str | None = None,
) -> dict[str, Any]:
payload = dict(member_payload)
payload["canonical_path"] = member_canonical_path(
object_path,
table_section=table_section,
member=payload.get("name"),
)
payload["path_kind"] = path_kind
if table_section:
payload["context_path"] = path_join(table_section, payload.get("name"))
return payload
def result_base(*, query: dict[str, Any], source: dict[str, Any], view: str, extension: str | None) -> dict[str, Any]:
return {
"schema": "onec_fact_resolution.v1",
"query": query,
"source": source,
"view": view,
"extension": extension if view == "extension" else None,
}
def resolve_from_route_index(
index: dict[str, Any],
*,
index_path: Path,
kind: str | None,
object_name: str,
member: str | None,
area: str,
table_section: str | None,
view: str,
extension: str | None,
) -> dict[str, Any]:
canonical_kind_value, parsed_name = parse_query(kind, object_name)
query = {
"kind": canonical_kind_value,
"name": parsed_name,
"member": member,
"table_section": table_section,
"area": area,
}
result = result_base(
query=query,
source={"kind": "route_index", "path": str(index_path)},
view=view,
extension=extension,
)
resolution = resolve_object(index, kind=canonical_kind_value, name=parsed_name, limit=50)
canonical = resolution.get("canonical")
if not canonical:
result.update(
{
"exists": False,
"confidence": "none",
"reason": "object_not_found",
"object": None,
"nearest": resolution.get("matches") or [],
}
)
return result
result["object"] = {
"kind": canonical.get("kind"),
"kind_ru": KIND_TO_RU.get(str(canonical.get("kind") or ""), canonical.get("kind")),
"name": canonical.get("name"),
"synonym": canonical.get("synonym"),
"uuid": canonical.get("guid"),
}
object_path = object_canonical_path(result["object"].get("kind"), result["object"].get("name"))
result["object"]["canonical_path"] = object_path
result["object"]["path_kind"] = "metadata_object"
result["active_extensions"] = sorted(
{
(overlay.get("origin") or {}).get("extension") or overlay.get("extension_name")
for overlay in (resolution.get("extension_overlays") or [])
if (overlay.get("origin") or {}).get("extension") or overlay.get("extension_name")
}
)
if not member and not table_section:
result.update({"exists": True, "confidence": "verified", "area": "object", "match": result["object"]})
return result
metadata = build_object_metadata(index, kind=canonical_kind_value or "", name=parsed_name, view=view, extension=extension, include_storage=False)
metadata_extensions = set(result.get("active_extensions") or [])
for item in metadata.get("attributes") or []:
origin = item.get("origin") or {}
if origin.get("extension"):
metadata_extensions.add(origin["extension"])
for extension_name in origin.get("extensions") or []:
if extension_name:
metadata_extensions.add(extension_name)
result["active_extensions"] = sorted(metadata_extensions)
checks: list[tuple[str, list[dict[str, Any]]]] = []
if area in {"any", "attribute"} and member:
checks.append(("attribute", metadata.get("attributes") or []))
if area in {"any", "tabular_section"} and not (area == "any" and table_section and member):
wanted = table_section or member
sections = metadata.get("tabular_sections") or []
if wanted:
match = next((item for item in sections if matches_name(item, wanted)), None)
if match:
result.update(
{
"exists": True,
"confidence": "verified",
"area": "tabular_section",
"match": with_path_fields(
compact_member(match, area="tabular_section"),
object_path=object_path,
path_kind="metadata_member",
),
}
)
return result
checks.append(("tabular_section", sections))
if area in {"any", "tabular_section_attribute"} and table_section and member:
sections = metadata.get("tabular_sections") or []
section = next((item for item in sections if matches_name(item, table_section)), None)
attrs = section.get("attributes") if section else []
if attrs:
match = next((item for item in attrs if matches_name(item, member)), None)
if match:
result.update(
{
"exists": True,
"confidence": "verified",
"area": "tabular_section_attribute",
"table_section": with_path_fields(
compact_member(section, area="tabular_section"),
object_path=object_path,
path_kind="metadata_member",
),
"match": with_path_fields(
compact_member(match, area="tabular_section_attribute"),
object_path=object_path,
path_kind="metadata_member",
table_section=str(section.get("name") or table_section),
),
}
)
return result
checks.append(("tabular_section_attribute", attrs))
if area in {"any", "form"} and member:
forms, _ = view_forms(canonical, resolution, view=view, extension=extension)
checks.append(("form", forms))
if area in {"any", "module"} and member:
modules, _ = view_modules(canonical, resolution, view=view, extension=extension)
checks.append(("module", modules))
for check_area, items in checks:
if not member:
continue
match = next((item for item in items if matches_name(item, member)), None)
if match:
result.update(
{
"exists": True,
"confidence": "verified",
"area": check_area,
"match": with_path_fields(
compact_member(match, area=check_area),
object_path=object_path,
path_kind="metadata_member" if check_area in {"attribute", "module", "form", "command"} else check_area,
),
}
)
return result
near = []
wanted = member or table_section or ""
for check_area, items in checks:
for item in nearest(items, wanted):
item["area"] = check_area
near.append(item)
result.update(
{
"exists": False,
"confidence": "verified_absent",
"reason": "member_not_found_in_current_source",
"nearest": near[:10],
"counts": {
"attributes": len(metadata.get("attributes") or []),
"tabular_sections": len(metadata.get("tabular_sections") or []),
},
}
)
return result
def snapshot_kind_matches(actual: str, wanted: str | None) -> bool:
if not wanted:
return True
canonical = canonical_kind(wanted)
aliases = SNAPSHOT_KIND_ALIASES.get(canonical or wanted, {wanted})
return compact_text(actual) in {compact_text(alias) for alias in aliases}
def resolve_from_snapshot(
snapshot: dict[str, Any],
*,
snapshot_path: Path,
kind: str | None,
object_name: str,
member: str | None,
area: str,
table_section: str | None,
view: str,
extension: str | None,
) -> dict[str, Any]:
query = {"kind": canonical_kind(kind), "name": object_name, "member": member, "table_section": table_section, "area": area}
result = result_base(
query=query,
source={"kind": "metadata_snapshot", "path": str(snapshot_path), "name": (snapshot.get("source") or {}).get("name")},
view=view,
extension=extension,
)
objects = snapshot.get("objects") or []
obj = next(
(
item
for item in objects
if snapshot_kind_matches(str(item.get("kind") or ""), kind)
and (same_name(item.get("name"), object_name) or same_name(item.get("full_name"), object_name) or same_name(item.get("synonym"), object_name))
),
None,
)
if not obj:
result.update({"exists": False, "confidence": "none", "reason": "object_not_found", "nearest": []})
return result
result["object"] = {key: obj.get(key) for key in ("kind", "name", "full_name", "synonym") if obj.get(key)}
object_path = str(obj.get("full_name") or object_canonical_path(obj.get("kind"), obj.get("name")))
result["object"]["canonical_path"] = object_path
result["object"]["path_kind"] = "metadata_object"
if not member and not table_section:
result.update({"exists": True, "confidence": "verified", "area": "object", "match": result["object"]})
return result
checks: list[tuple[str, list[dict[str, Any]]]] = []
if area in {"any", "attribute"} and member:
checks.append(("attribute", obj.get("attributes") or []))
if area in {"any", "tabular_section"} and not (area == "any" and table_section and member):
wanted = table_section or member
sections = obj.get("tabular_sections") or []
if wanted:
section = next((item for item in sections if matches_name(item, wanted)), None)
if section:
result.update(
{
"exists": True,
"confidence": "verified",
"area": "tabular_section",
"match": with_path_fields(
compact_member(section, area="tabular_section"),
object_path=object_path,
path_kind="metadata_member",
),
}
)
return result
checks.append(("tabular_section", sections))
if area in {"any", "tabular_section_attribute"} and table_section and member:
section = next((item for item in obj.get("tabular_sections") or [] if matches_name(item, table_section)), None)
attrs = section.get("attributes") if section else []
checks.append(("tabular_section_attribute", attrs or []))
if area in {"any", "form"} and member:
checks.append(("form", obj.get("forms") or []))
if area in {"any", "command"} and member:
checks.append(("command", obj.get("commands") or []))
if area in {"any", "module"} and member:
checks.append(("module", obj.get("modules") or []))
for check_area, items in checks:
if member:
match = next((item for item in items if matches_name(item, member)), None)
if match:
context_section = None
if check_area == "tabular_section_attribute" and table_section:
context_section = table_section
result.update(
{
"exists": True,
"confidence": "verified",
"area": check_area,
"match": with_path_fields(
compact_member(match, area=check_area),
object_path=object_path,
path_kind="metadata_member",
table_section=context_section,
),
}
)
return result
near = []
wanted = member or table_section or ""
for check_area, items in checks:
for item in nearest(items, wanted):
item["area"] = check_area
near.append(item)
result.update({"exists": False, "confidence": "verified_absent", "reason": "member_not_found_in_explicit_snapshot", "nearest": near[:10]})
return result
def main() -> int:
parser = argparse.ArgumentParser(description="Resolve a concrete 1C fact against an explicit configuration source.")
source = parser.add_mutually_exclusive_group(required=True)
source.add_argument("--index", type=Path, help="Unified object route index for current configuration.")
source.add_argument("--snapshot", type=Path, help="Explicit metadata snapshot. Use only when this is the intended source.")
parser.add_argument("--path", help="Fact path, for example Справочник.Номенклатура.Артикул")
parser.add_argument("--path-b64")
parser.add_argument("--kind")
parser.add_argument("--name")
parser.add_argument("--member")
parser.add_argument("--table-section")
parser.add_argument("--area", default="any")
parser.add_argument("--view", choices=["effective", "base", "extension"], default="effective")
parser.add_argument("--extension")
parser.add_argument("--output", type=Path)
args = parser.parse_args()
fact_path = decode_arg(args.path, args.path_b64)
if fact_path:
path_kind, path_name, path_section, path_member = split_fact_path(fact_path)
kind = args.kind or path_kind
name = args.name or path_name
table_section = args.table_section or path_section
member = args.member or path_member
else:
kind = args.kind
name = args.name
table_section = args.table_section
member = args.member
if not name:
raise SystemExit("Use --path or --kind/--name.")
area = member_area(args.area)
if args.index:
result = resolve_from_route_index(
load_json(args.index),
index_path=args.index,
kind=kind,
object_name=name,
member=member,
area=area,
table_section=table_section,
view=args.view,
extension=args.extension,
)
else:
result = resolve_from_snapshot(
load_json(args.snapshot),
snapshot_path=args.snapshot,
kind=kind,
object_name=name,
member=member,
area=area,
table_section=table_section,
view=args.view,
extension=args.extension,
)
text = json.dumps(result, ensure_ascii=False, indent=2) + "\n"
if args.output:
args.output.parent.mkdir(parents=True, exist_ok=True)
args.output.write_text(text, encoding="utf-8")
print(json.dumps({"output": str(args.output), "exists": result.get("exists"), "confidence": result.get("confidence")}, ensure_ascii=False))
else:
print(text)
return 0
if __name__ == "__main__":
raise SystemExit(main())