feat: Add block_simultaneous_read with stream_groups + fix default HTTP timeout#940

Closed

Alfredo Garcia (agarctfi) wants to merge 31 commits intomainfrom

devin/1772836274-block-simultaneous-read-with-timeout

Commits on Mar 6, 2026

Add block_simultaneous_read to DefaultStream

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
Change `block_simultaneous_read` to string

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
Fix StreamFacade

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
Fix NoneType error when all streams are blocked

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
Fix unit tests

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
Auto-fix lint and format issues

octavia-squidington-iii
authored and
devin-ai-integration[bot]
committed
Add retry deferred streams on stream completion

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
Fix unit tests

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
More fixes for unit tests

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
refactor: replace per-stream block_simultaneous_read with top-level stream_groups
devin-ai-integration[bot]
committed
refactor: move stream_name_to_group into ModelToComponentFactory
devin-ai-integration[bot]
committed
refactor: use stream_groups manifest in factory test instead of hardcoded dict
devin-ai-integration[bot]
committed
fix: only include parent stream in stream_groups to avoid deadlock
devin-ai-integration[bot]
committed
style: fix ruff format for long line
devin-ai-integration[bot]
committed
refactor: move _build_stream_name_to_group into ModelToComponentFactory
devin-ai-integration[bot]
committed
refactor: resolve stream_groups from actual stream instances instead of factory
devin-ai-integration[bot]
committed
Fix stream format in schema

Anatolii Yatsuk (tolik0)
authored and
devin-ai-integration[bot]
committed
refactor: add get_partition_router() helper to DefaultStream
devin-ai-integration[bot]
committed
feat: validate no parent-child streams share a group to prevent deadlock
devin-ai-integration[bot]
committed
feat: assert partition generation queue is empty when all streams are done
devin-ai-integration[bot]
committed
refactor: move inline imports to module level in default_stream.py and concurrent_declarative_source.py
devin-ai-integration[bot]
committed
fix: unwrap GroupingPartitionRouter in get_partition_router() to detect parent streams
devin-ai-integration[bot]
committed
fix: handle GroupingPartitionRouter at call sites instead of in get_partition_router()
devin-ai-integration[bot]
committed
feat: check active_groups is empty in is_done() safety check
devin-ai-integration[bot]
committed
test: add missing unit tests for GroupingPartitionRouter, active_groups check, and get_partition_router
devin-ai-integration[bot]
committed
fix: make deadlock validation check all ancestors, not just direct parents
devin-ai-integration[bot]
committed
style: alphabetize StreamGroup and BlockSimultaneousSyncsAction in schema definitions
devin-ai-integration[bot]
committed
style: move BlockSimultaneousSyncsAction next to StreamGroup for easier reading
devin-ai-integration[bot]
committed
fix: add default HTTP request timeout to prevent indefinite hangs

devin-ai-integration[bot]
and
Alfredo Garcia (agarctfi)
committed
fix: match test assertions against AirbyteTracedException message field

devin-ai-integration[bot]
and
Alfredo Garcia (agarctfi)
committed
style: fix ruff format in test_concurrent_read_processor.py

devin-ai-integration[bot]
and
Alfredo Garcia (agarctfi)
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: Add block_simultaneous_read with stream_groups + fix default HTTP timeout#940

feat: Add block_simultaneous_read with stream_groups + fix default HTTP timeout#940
Alfredo Garcia (agarctfi) wants to merge 31 commits intomainfrom
devin/1772836274-block-simultaneous-read-with-timeout

Commits on Mar 6, 2026

Add block_simultaneous_read to DefaultStream

Change `block_simultaneous_read` to string

Fix StreamFacade

Fix NoneType error when all streams are blocked

Fix unit tests

Auto-fix lint and format issues

Add retry deferred streams on stream completion

Fix unit tests

More fixes for unit tests

refactor: replace per-stream block_simultaneous_read with top-level stream_groups

refactor: move stream_name_to_group into ModelToComponentFactory

refactor: use stream_groups manifest in factory test instead of hardcoded dict

fix: only include parent stream in stream_groups to avoid deadlock

style: fix ruff format for long line

refactor: move _build_stream_name_to_group into ModelToComponentFactory

refactor: resolve stream_groups from actual stream instances instead of factory

Fix stream format in schema

refactor: add get_partition_router() helper to DefaultStream

feat: validate no parent-child streams share a group to prevent deadlock

feat: assert partition generation queue is empty when all streams are done

refactor: move inline imports to module level in default_stream.py and concurrent_declarative_source.py

fix: unwrap GroupingPartitionRouter in get_partition_router() to detect parent streams

fix: handle GroupingPartitionRouter at call sites instead of in get_partition_router()

feat: check active_groups is empty in is_done() safety check

test: add missing unit tests for GroupingPartitionRouter, active_groups check, and get_partition_router

fix: make deadlock validation check all ancestors, not just direct parents

style: alphabetize StreamGroup and BlockSimultaneousSyncsAction in schema definitions

style: move BlockSimultaneousSyncsAction next to StreamGroup for easier reading

fix: add default HTTP request timeout to prevent indefinite hangs

fix: match test assertions against AirbyteTracedException message field

style: fix ruff format in test_concurrent_read_processor.py