Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
import time
from datetime import UTC, datetime, timedelta, timezone
import pytest
from pydantic import TypeAdapter, ValidationError
from fields.timestamp import Timestamp
def test_timestamp_preserves_datetime_until_json_serialization() -> None:
source = datetime(2026, 1, 12, 8, tzinfo=UTC)
adapter = TypeAdapter(Timestamp)
timestamp = adapter.validate_python(source)
assert timestamp is source
assert adapter.dump_python(timestamp, mode="json") == 1768204800
assert adapter.json_schema(mode="serialization")["type"] == "integer"
@pytest.mark.parametrize("host_timezone", ["UTC", "Asia/Shanghai"])
@pytest.mark.parametrize(
"source",
[
datetime(2026, 8, 20, 8),
datetime(2026, 8, 20, 8, tzinfo=UTC),
datetime(2026, 8, 20, 17, tzinfo=timezone(timedelta(hours=9))),
],
)
def test_timestamp_interprets_naive_values_as_utc_in_every_host_timezone(
source: datetime, host_timezone: str, monkeypatch: pytest.MonkeyPatch
) -> None:
adapter = TypeAdapter(Timestamp)
try:
with monkeypatch.context() as timezone_environment:
timezone_environment.setenv("TZ", host_timezone)
time.tzset()
assert adapter.dump_python(adapter.validate_python(source), mode="json") == 1787212800
finally:
time.tzset()
@pytest.mark.parametrize("value", [1768204800, 1768204800.5])
def test_timestamp_rejects_preconverted_numeric_values(value: int | float) -> None:
with pytest.raises(ValidationError, match="Timestamp source must be a datetime"):
TypeAdapter(Timestamp).validate_python(value)