если объект найден по реквизиту (колонка в FTS), в выдаче показывается только совпавший реквизит с типом значения — в том же формате дерева

This commit is contained in:
mihailkudravcev
2026-07-17 16:56:15 +03:00
parent 404659a545
commit 06e13395f8
6 changed files with 239 additions and 53 deletions
+205 -24
View File
@@ -3,12 +3,15 @@
from __future__ import annotations
import json
import re
import sqlite3
from pathlib import Path
from typing import Any
from .parse_meta import FieldRec, ObjectRec, parse_metadata_xml
_QUERY_TOKEN_RE = re.compile(r"[\wА-Яа-яЁё]+", flags=re.UNICODE)
FILL_LABEL = {
"ShowError": "обяз.",
@@ -43,6 +46,174 @@ def _field_matches(f: dict[str, Any], q: str, qn: str) -> bool:
return name_cf == qn or syn_cf == qn or qn in syn_parts
def _query_tokens(query: str) -> list[str]:
return _QUERY_TOKEN_RE.findall(query or "")
def _words_cf(text: str) -> list[str]:
return [w.casefold() for w in _QUERY_TOKEN_RE.findall(text or "")]
def _tokens_match_text(tokens: list[str], text: str) -> bool:
"""Как FTS search: каждый токен — префикс какого-либо слова в тексте."""
if not tokens:
return False
words = _words_cf(text)
if not words:
return False
for tok in tokens:
t = tok.casefold()
if not any(w.startswith(t) for w in words):
return False
return True
def object_header_matches_search(row: dict[str, Any], query: str) -> bool:
"""Совпадение по имени/синониму/комментарию объекта (не по реквизитам)."""
tokens = _query_tokens(query)
if not tokens:
return False
for part in (
row.get("name") or "",
row.get("full_name") or "",
row.get("synonym") or "",
row.get("comment") or "",
):
if _tokens_match_text(tokens, part):
return True
return False
def field_matches_search_query(f: dict[str, Any], query: str) -> bool:
"""Совпадение реквизита с поисковым запросом (префикс по словам, как FTS)."""
tokens = _query_tokens(query)
if not tokens:
return False
text = " ".join(
filter(
None,
[
f.get("kind") or "",
f.get("name") or "",
f.get("synonym") or "",
f.get("comment") or "",
f.get("tabular_section") or "",
" ".join(f.get("types") or []),
],
)
)
return _tokens_match_text(tokens, text)
def _field_dict_from_db_row(row: sqlite3.Row) -> dict[str, Any]:
return {
"kind": row["kind"],
"name": row["name"],
"synonym": row["synonym"] or "",
"comment": row["comment"] or "" if "comment" in row.keys() else "",
"types": [t.strip() for t in (row["types"] or "").split(",") if t.strip()],
"tabular_section": row["tabular_section"] or "",
"refs": json.loads(row["refs_json"] or "[]"),
"fill_checking": "",
}
def find_fields_for_search_query(
conn: sqlite3.Connection,
object_id: int,
query: str,
) -> list[dict[str, Any]]:
"""Реквизиты объекта, совпавшие с FTS-запросом search."""
rows = conn.execute(
"""
SELECT kind, name, synonym, comment, types, tabular_section, refs_json
FROM fields
WHERE object_id = ?
ORDER BY tabular_section, kind, name
""",
(object_id,),
)
matched: list[dict[str, Any]] = []
seen: set[tuple[str, str, str]] = set()
for row in rows:
fdict = _field_dict_from_db_row(row)
if not field_matches_search_query(fdict, query):
continue
sig = (fdict["tabular_section"], fdict["name"], fdict["kind"])
if sig in seen:
continue
seen.add(sig)
matched.append(fdict)
return matched
def enrich_matched_fields(obj: dict[str, Any], matched: list[dict[str, Any]]) -> None:
"""Подтянуть FillChecking и типы из актуальной структуры объекта."""
if not matched or not obj.get("fields"):
return
by_f = {
(f.get("tabular_section") or "", f.get("name") or "", f.get("kind") or ""): f
for f in obj["fields"]
}
for mf in matched:
key = (mf.get("tabular_section") or "", mf.get("name") or "", mf.get("kind") or "")
src = by_f.get(key)
if not src:
continue
mf["fill_checking"] = src.get("fill_checking") or ""
mf["types"] = src.get("types") or mf.get("types")
mf["synonym"] = src.get("synonym") or mf.get("synonym")
mf["refs"] = src.get("refs") or mf.get("refs")
def load_object_structure_for_row(
conn: sqlite3.Connection,
row: dict[str, Any],
) -> dict[str, Any]:
match = {
"config": row["config"],
"full_name": row["full_name"],
"path": row.get("path"),
"object_type": row.get("object_type"),
}
if row.get("json"):
match["json"] = row["json"]
elif row.get("object_id"):
jrow = conn.execute(
"SELECT json FROM objects WHERE id=?", (row["object_id"],)
).fetchone()
match["json"] = jrow["json"] if jrow else "{}"
else:
jrow = conn.execute(
"SELECT json FROM objects WHERE config=? AND full_name=?",
(row["config"], row["full_name"]),
).fetchone()
match["json"] = jrow["json"] if jrow else "{}"
return load_object_structure(match)
def annotate_search_hits(
conn: sqlite3.Connection,
rows: list[dict[str, Any]],
query: str,
) -> list[dict[str, Any]]:
"""Уточнить search: object — по шапке; field — по реквизитам в fields_text."""
for row in rows:
if object_header_matches_search(row, query):
row["match_kind"] = "object"
row["matched_fields"] = []
continue
oid = row.get("object_id")
matched = find_fields_for_search_query(conn, int(oid), query) if oid else []
if matched:
row["match_kind"] = "field"
row["matched_fields"] = matched
else:
row["match_kind"] = "object"
row["matched_fields"] = []
return rows
def find_structure_matches(
conn: sqlite3.Connection,
query: str,
@@ -239,12 +410,43 @@ def format_field_line(
return f"{indent}{branch}`{kind}.{name}`{syn_s} [{fill}] : {types_s}"
def _select_matched_field_nodes(
fields: list[dict[str, Any]],
matched_fields: list[dict[str, Any]],
) -> tuple[list[dict[str, Any]], set[tuple[str, str, str]], set[str]]:
want: set[tuple[str, str, str]] = set()
need_ts: set[str] = set()
for mf in matched_fields:
want.add(
(
mf.get("kind") or "",
mf.get("tabular_section") or "",
mf.get("name") or "",
)
)
if mf.get("tabular_section") and mf.get("kind") == "Column":
need_ts.add(mf["tabular_section"])
selected: list[dict[str, Any]] = []
for f in fields:
key = (
f.get("kind") or "",
f.get("tabular_section") or "",
f.get("name") or "",
)
if key in want:
selected.append(f)
elif f.get("kind") == "TabularSection" and f.get("name") in need_ts:
selected.append(f)
return selected, want, need_ts
def format_object_tree(
obj: dict[str, Any],
*,
clear: bool = False,
matched_fields: list[dict[str, Any]] | None = None,
match_kind: str = "object",
omit_header: bool = False,
) -> str:
"""Дерево реквизитов объекта. clear — только совпадения."""
lines: list[str] = []
@@ -254,35 +456,14 @@ def format_object_tree(
header = f"`{obj.get('full_name') or obj.get('object_type') + '.' + obj.get('name')}`{syn_s}"
if cfg:
header = f"[{cfg}] {header}"
lines.append(header)
if not omit_header:
lines.append(header)
fields: list[dict[str, Any]] = list(obj.get("fields") or [])
if clear:
if match_kind == "field" and matched_fields:
want: set[tuple[str, str, str]] = set()
need_ts: set[str] = set()
for mf in matched_fields:
want.add(
(
mf.get("kind") or "",
mf.get("tabular_section") or "",
mf.get("name") or "",
)
)
if mf.get("tabular_section") and mf.get("kind") == "Column":
need_ts.add(mf["tabular_section"])
selected: list[dict[str, Any]] = []
for f in fields:
key = (
f.get("kind") or "",
f.get("tabular_section") or "",
f.get("name") or "",
)
if key in want:
selected.append(f)
elif f.get("kind") == "TabularSection" and f.get("name") in need_ts:
selected.append(f)
selected, want, need_ts = _select_matched_field_nodes(fields, matched_fields)
if not selected:
lines.append(" (совпавшие реквизиты не найдены в структуре)")
return "\n".join(lines)