fix(dwh): validate adapter limits strictly

This commit is contained in:
2026-07-11 20:19:05 +02:00
parent 216984aac8
commit f6302b31dd
7 changed files with 84 additions and 19 deletions
+5 -4
View File
@@ -4,6 +4,7 @@ from dataclasses import dataclass
from sqlalchemy import Engine, text
from tht.config import ExamplesConfig, LshConfig
from tht.execute import require_positive_int
from tht.mschema.models import Annotations, PhysicalSchema
from tht.ports.dwh import DistinctValues
@@ -26,8 +27,7 @@ def _quoted_top_values_query(engine: Engine, schema: str, table: str, column: st
def sample_column(
engine: Engine, schema: str, table: str, column: str, *, limit: int
) -> list[object]:
if limit <= 0:
raise ValueError("limit must be a positive integer")
limit = require_positive_int(limit, name="limit")
query = _quoted_top_values_query(engine, schema, table, column)
with engine.connect() as conn:
rows = conn.execute(query, {"lim": limit}).fetchall()
@@ -37,8 +37,7 @@ def sample_column(
def sample_column_rest(
client, schema: str, table: str, column: str, *, limit: int
) -> list[object]:
if limit <= 0:
raise ValueError("limit must be a positive integer")
limit = require_positive_int(limit, name="limit")
rows = client.top_values(schema, table, column, limit)
return [row["value"] for row in rows if row.get("value") is not None]
@@ -51,6 +50,7 @@ def distinct_values(
*,
max_values: int = DEFAULT_DISTINCT_VALUES_LIMIT,
) -> DistinctValues:
max_values = require_positive_int(max_values, name="max_values")
values = sample_column(engine, schema, table, column, limit=max_values + 1)
return DistinctValues(values=values[:max_values], truncated=len(values) > max_values)
@@ -63,6 +63,7 @@ def distinct_values_rest(
*,
max_values: int = DEFAULT_DISTINCT_VALUES_LIMIT,
) -> DistinctValues:
max_values = require_positive_int(max_values, name="max_values")
values = sample_column_rest(client, schema, table, column, limit=max_values + 1)
return DistinctValues(values=values[:max_values], truncated=len(values) > max_values)