fix/json-serializer-pydantic-generics - Branch - PrefectHQ/prefect

fix: handle pydantic generic models in JSON serializer

#19868Merged

Comparing

fix/json-serializer-pydantic-generics

(

139b051

) with

main

(

080230e

)

Untouched: 2

Ignored: 6

Benchmarks

Passed

bench_import_prefect_flow

benches/bench_import.py

1.4 s1.4 s

bench_task_decorator

benches/bench_tasks.py

462 µs462.4 µs

Ignored

bench_flow_call[options0]

benches/bench_flows.py

Ignored

-3%

162.9 ms168.5 ms

bench_task_call

benches/bench_tasks.py

Ignored

+1%

34.1 ms33.7 ms

bench_task_submit

benches/bench_tasks.py

Ignored

-18%

8.4 ms10.2 ms

bench_flow_call[options1]

benches/bench_flows.py

Ignored

107.2 ms107.5 ms

bench_flow_decorator

benches/bench_flows.py

Ignored

5.5 ms5.5 ms

bench_import_prefect

benches/bench_import.py

Ignored

1.1 s1.1 s

Commits

Click on a commit to change the comparison range

Base

main

080230e

+0.01%

fix: handle pydantic generic models in JSON serializer When using parameterized Pydantic generic models like `APIResult[str]`, the JSON serializer was storing class names with brackets (e.g., `module.APIResult[str]`) which cannot be imported by Python's import system, causing deserialization to fail. This fix extracts the origin class from Pydantic's generic metadata, ensuring the serialized class name is importable (e.g., `module.APIResult`). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

7a91c0c

2 days ago

by zzstoatzz

test: add coverage for non-Pydantic generic dataclass serialization Ensures the Pydantic generic fix doesn't break standard Generic dataclasses. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

139b051

2 days ago

by zzstoatzz

Home Terms Privacy Docs