fix(ph-ai): async memory nodes #42888
Conversation
This stack of pull requests is managed by Graphite. Learn more about stacking. |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Additional Comments (1)
-
ee/hogai/chat_agent/slash_commands/commands/remember/command.py, line 51 (link)syntax: calling sync method
append_core_memory()which doesn't exist anymore - should useasync_append_core_memory()
7 files reviewed, 1 comment
There was a problem hiding this comment.
Pull request overview
This PR converts memory-related nodes from synchronous to asynchronous operations to prevent deadlocks when running subagent workflows. The changes include converting database operations to their async equivalents, updating method signatures to use async/await, and converting all tests to be async-compatible.
Key Changes
- Converted memory-related model methods and node operations to async/await pattern
- Added
asyncio_mode = autotoee/pytest.inifor async test support - Updated all test classes to use
NonAtomicBaseTestand converted test methods to async
Reviewed changes
Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| ee/pytest.ini | Added asyncio_mode configuration for pytest async support |
| ee/models/assistant.py | Converted CoreMemory methods to async (achange_status_to_pending, async_append_question_to_initial_text, etc.) |
| ee/hogai/chat_agent/slash_commands/test/test_slash_command_handler.py | Converted router tests to async |
| ee/hogai/chat_agent/slash_commands/nodes.py | Converted router method to arouter and made it async |
| ee/hogai/chat_agent/memory/test/test_nodes.py | Converted all test classes to NonAtomicBaseTest and made test methods async with proper async database operations |
| ee/hogai/chat_agent/memory/nodes.py | Converted memory node methods to async (arun, arouter), replaced sync database calls with async equivalents |
| ee/hogai/chat_agent/graph.py | Updated router method calls to arouter |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
kappa90
force-pushed
the
12-06-fix_ph-ai_async_memory_nodes
branch
2 times, most recently
from
3edc6c1 to
e456c9e
Compare
kappa90
force-pushed
the
12-06-fix_ph-ai_async_memory_nodes
branch
5 times, most recently
from
0223120 to
009a1ec
Compare
Co-authored-by: Copilot <[email protected]>
kappa90
force-pushed
the
12-06-fix_ph-ai_async_memory_nodes
branch
from
009a1ec to
f61f0eb
Compare
skoob13
left a comment
skoob13
left a comment
There was a problem hiding this comment.
Approved. Pushed fixes for a few sync issues.
🧠 AI eval resultsEvaluated 35 experiments, comprising 64 metrics. Showing experiments with largest regressions first. funnel🔴 plan_correctness: 5.00%, -2.50% (improvements: 2, regressions: 2) Baseline: master-1764965827 • Avg. case performance: ⏱️ 157.35 s, 🔢 1955 tokens, 💵 $0.0011 in tokens insight_evaluation_accuracy🔴 InsightEvaluationAccuracy: 25.00%, -25.00% (improvements: 0, regressions: 1) Baseline: master-1764966032 • Avg. case performance: ⏱️ 15.08 s, 🔢 1547 tokens, 💵 $0.0030 in tokens memory_onboarding🔴 has_correct_style: 50.00%, -16.67% (improvements: 0, regressions: 1) Baseline: master-1764966063 • Avg. case performance: ⏱️ 67.30 s, 🔢 406 tokens, 💵 $0.0029 in tokens tool_search_session_recordings🔴 date_time_filtering_correctness: 86.84%, -10.71% (improvements: 0, regressions: 2) Baseline: master-1764966802 • Avg. case performance: ⏱️ 10.09 s, 🔢 16787 tokens, 💵 $0.0344 in tokens retention🔵 QueryKindSelection: 100.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966182 • Avg. case performance: ⏱️ 49.36 s, 🔢 0 tokens tool_search_session_recordings🔴 date_time_filtering_correctness: 91.67%, -5.56% (improvements: 0, regressions: 1) Baseline: master-1764966802 • Avg. case performance: ⏱️ 18.86 s, 🔢 16497 tokens, 💵 $0.0338 in tokens tool_generate_hogql_query🔵 no_mustache: 100.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966728 • Avg. case performance: ⏱️ 6.31 s, 🔢 21984 tokens, 💵 $0.0446 in tokens memory🟢 ToolRelevance: 88.80%, +18.80% (improvements: 2, regressions: 0) Baseline: master-1764966049 • Avg. case performance: ⏱️ 1.89 s, 🔢 1142 tokens, 💵 $0.0032 in tokens yaml_fixing🟢 ExactMatch: 100.00%, +16.67% (improvements: 1, regressions: 0) Baseline: master-1764966888 • Avg. case performance: ⏱️ 0.64 s, 🔢 169 tokens, 💵 $0.0001 in tokens surveys🔵 feature_flag_integration: 58.33%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966574 • Avg. case performance: ⏱️ 3.75 s, 🔢 5470 tokens, 💵 $0.0117 in tokens 25 experiments with no significant changestrends🔵 plan_correctness: 77.50%, ±0.00% (improvements: 1, regressions: 1) Baseline: master-1764966595 • Avg. case performance: ⏱️ 30.34 s, 🔢 0 tokens tool_routing_session_replay🔵 ToolRelevance: 28.20%, -1.05% (improvements: 1, regressions: 2) Baseline: master-1764966353 • Avg. case performance: ⏱️ 8.63 s, 🔢 0 tokens root_documentation🔵 ToolRelevance: 95.18%, +0.24% (improvements: 15, regressions: 10) Baseline: master-1764966282 • Avg. case performance: ⏱️ 16.51 s, 🔢 0 tokens tool_routing_dashboard_creation🔵 ToolRelevance: 58.68%, -0.14% (improvements: 2, regressions: 2) Baseline: master-1764965807 • Avg. case performance: ⏱️ 9.30 s, 🔢 0 tokens root🔵 ToolRelevance: 10.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966266 • Avg. case performance: ⏱️ 8.07 s, 🔢 0 tokens root_entity_search🔵 ToolRelevance: 98.22%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966314 • Avg. case performance: ⏱️ 4.92 s, 🔢 0 tokens root_style🔵 style_checker: 100.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966325 • Avg. case performance: ⏱️ 8.88 s, 🔢 0 tokens session_summarization_no_context🔵 ToolRelevance: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966370 • Avg. case performance: ⏱️ 5.15 s, 🔢 0 tokens session_summarization_limit_with_context🔵 ToolRelevance: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966381 • Avg. case performance: ⏱️ 7.33 s, 🔢 0 tokens session_summarization_limit_without_context🔵 ToolRelevance: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966396 • Avg. case performance: ⏱️ 6.05 s, 🔢 0 tokens session_summarization_current_session🔵 ToolRelevance: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966408 • Avg. case performance: ⏱️ 6.72 s, 🔢 0 tokens sql🔵 plan_correctness: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966423 • Avg. case performance: ⏱️ 45.47 s, 🔢 0 tokens survey_analysis🔵 recommendation_quality: 100.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966550 • Avg. case performance: ⏱️ 9.94 s, 🔢 2697 tokens, 💵 $0.0088 in tokens ui_context_actions🔵 ToolRelevance: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966651 • Avg. case performance: ⏱️ 7.08 s, 🔢 0 tokens ui_context_events🔵 ToolRelevance: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966667 • Avg. case performance: ⏱️ 7.59 s, 🔢 0 tokens create_experiment🔵 ExperimentOutputScorer: 87.13%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966681 • Avg. case performance: ⏱️ 1.50 s, 🔢 0 tokens create_feature_flag🔵 FeatureFlagOutputScorer: 89.62%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966688 • Avg. case performance: ⏱️ 15.49 s, 🔢 13518 tokens, 💵 $0.0277 in tokens combined_rename_and_add🔵 SemanticSimilarity: 88.93%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966724 • Avg. case performance: ⏱️ 0.05 s, 🔢 0 tokens tool_filter_revenue_analytics🔵 date_time_filtering_correctness: 100.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966751 • Avg. case performance: ⏱️ 1.91 s, 🔢 3826 tokens, 💵 $0.0081 in tokens tool_filter_revenue_analytics_ask_user_for_help🔵 ask_user_for_help_scorer: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966757 • Avg. case performance: ⏱️ 0.90 s, 🔢 3801 tokens, 💵 $0.0080 in tokens tool_search_session_recordings_ask_user_for_help🔵 ask_user_for_help_scorer: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966824 • Avg. case performance: ⏱️ 12.48 s, 🔢 45872 tokens, 💵 $0.0927 in tokens tool_search_session_recordings_ask_user_for_help🔵 ask_user_for_help_scorer: 0.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966824 • Avg. case performance: ⏱️ 11.47 s, 🔢 73793 tokens, 💵 $0.1486 in tokens tool_call_dashboard_creation🔵 dashboard_creation_accuracy: 40.00%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966840 • Avg. case performance: ⏱️ 14.63 s, 🔢 7344 tokens, 💵 $0.0034 in tokens filter_query_generation🔵 SemanticSimilarity: 99.21%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966878 • Avg. case performance: ⏱️ 0.47 s, 🔢 579 tokens, 💵 $0.0012 in tokens insights_addition🔵 SemanticSimilarity: 55.14%, ±0.00% (improvements: 0, regressions: 0) Baseline: master-1764966893 • Avg. case performance: ⏱️ 0.31 s, 🔢 0 tokens Triggered by this commit. |
Problem
Memory nodes are sync and lead to deadlocks when running subagent workflows. Let's convert memory-related nodes to async to remove blocking operations.
Changes
How did you test this code?
Updated all tests