refactor(agents-api): Switch to psycopg from asyncpg

[creatorrr]

User description

Signed-off-by: Diwank Singh Tomer diwank.singh@gmail.com

---

PR Type

Enhancement, Other

---

Description

• Major refactoring: Added new SQLBuilder class for safe dynamic SQL query construction with parameter placeholders and SQL injection prevention
• Database query improvements: Refactored agent and user patching queries to use new SQL builder utilities, replacing complex CASE/WHEN patterns
• Query optimization: Enhanced agent listing and execution transitions queries with dynamic filtering capabilities
• Code formatting: Applied comprehensive formatting improvements across the entire codebase including:

• Multi-line formatting for long function signatures and parameter lists
• Consistent formatting for decorators, imports, and method calls
• Improved readability of test assertions and data structures
• Better line breaking for complex expressions and constructor calls
  • Performance improvements: Test runtime measurements show significant performance gains across the test suite
  • Maintainability: Improved separation of concerns and type safety in database operations

---

Changes walkthrough 📝

	Relevant files
Enhancement	5 files sql_builder.py Add SQL builder utilities for dynamic query construction  agents-api/agents_api/queries/sql_builder.py • Added a new SQLBuilder class for safe dynamic SQL query construction • Implemented methods for building UPDATE, WHERE, ORDER BY, and LIMIT clauses • Added utility functions for building patch queries and JSONB updates • Includes parameter placeholders and SQL injection prevention +342/-0  patch_agent.py Refactor agent patching to use SQL builder utilities          agents-api/agents_api/queries/agents/patch_agent.py • Refactored to use the new build_patch_query utility from sql_builder • Replaced complex CASE/WHEN SQL patterns with dynamic query building • Simplified parameter handling and improved type safety • Added special handling for metadata JSONB merge operations +70/-61  patch_user.py Refactor user patching to use SQL builder utilities            agents-api/agents_api/queries/users/patch_user.py • Refactored to use the new build_patch_query utility from sql_builder • Replaced complex CASE/WHEN SQL patterns with dynamic query building • Simplified parameter handling and improved maintainability • Added special handling for metadata JSONB merge operations +65/-43  list_agents.py Refactor agent listing to use metadata filter utilities    agents-api/agents_api/queries/agents/list_agents.py • Refactored to use build_metadata_filter_conditions utility function • Simplified SQL query construction and parameter handling • Improved separation of concerns between base query and dynamic filters +31/-23  list_execution_transitions.py Refactor execution transitions query with dynamic filtering agents-api/agents_api/queries/executions/list_execution_transitions.py • Refactored SQL query construction to use dynamic string formatting • Improved separation of base query and conditional filters • Applied consistent formatting to decorator and function definitions +26/-19
Formatting	77 files task_validation.py Code formatting improvements for task validation utilities agents-api/agents_api/common/utils/task_validation.py • Reformatted import statements to use multi-line format • Applied consistent code formatting to list comprehensions and dictionary structures • Improved readability of nested conditional statements and function calls +72/-46  test_docs_routes.py Code formatting improvements for document route tests        agents-api/tests/test_docs_routes.py • Applied consistent formatting to function calls and method invocations • Reformatted long parameter lists to use multi-line format • Improved readability of test assertions and data structures +53/-17  test_task_execution_workflow.py Code formatting improvements for task execution workflow tests agents-api/tests/test_task_execution_workflow.py • Reformatted long function calls and parameter lists for better readability • Applied consistent multi-line formatting to patch decorators and assertions • Improved code structure in test method definitions +65/-31  test_mmr.py Code formatting improvements for MMR algorithm tests          agents-api/tests/test_mmr.py • Reformatted test data creation and function calls for better readability • Applied consistent multi-line formatting to list comprehensions and assertions • Improved structure of test method parameters and return values +42/-14  test_chat_streaming.py Code formatting improvements for chat streaming tests        agents-api/tests/test_chat_streaming.py • Applied consistent formatting to patch decorators and function calls • Reformatted long parameter lists and assertions for better readability • Improved code structure in streaming chat test methods +44/-14  env.py Code formatting improvements for environment configuration agents-api/agents_api/env.py • Applied consistent multi-line formatting to environment variable definitions • Reformatted long parameter lists and function calls for better readability +21/-7    db_exceptions.py Code formatting improvements for database exception handling agents-api/agents_api/common/utils/db_exceptions.py • Applied consistent multi-line formatting to exception mapping dictionaries • Reformatted long parameter lists and function calls for better readability • Improved code structure in error handling utilities +36/-22  __init__.py Code formatting improvements for task execution workflows agents-api/agents_api/workflows/task_execution/init.py • Applied consistent formatting to conditional statements and function calls • Reformatted long parameter lists and timeout configurations • Improved readability of workflow step handling methods +33/-11  test_task_validation.py Code formatting improvements for task validation tests      agents-api/tests/test_task_validation.py • Applied consistent multi-line formatting to import statements and test data • Reformatted test method definitions and assertions for better readability • Improved structure of test case parameters and expected results +38/-11  test_task_routes.py Code formatting improvements for task route tests                agents-api/tests/test_task_routes.py • Applied consistent formatting to function calls and method invocations • Reformatted long parameter lists and test assertions • Improved readability of route testing methods +25/-8    test_secrets_usage.py Code formatting improvements for secrets usage tests          agents-api/tests/test_secrets_usage.py • Applied consistent multi-line formatting to patch decorators and function calls • Reformatted long parameter lists and test data structures • Improved readability of secrets usage test methods +31/-9    test_usage_cost.py Code formatting improvements for usage cost tests                agents-api/tests/test_usage_cost.py • Applied consistent formatting to function calls and method invocations • Reformatted long parameter lists and test assertions • Improved readability of usage cost testing methods +24/-8    test_middleware.py Code formatting improvements for middleware tests                agents-api/tests/test_middleware.py • Applied consistent multi-line formatting to patch decorators and function calls • Reformatted long parameter lists and test assertions • Improved readability of middleware testing methods +29/-10  test_usage_tracking.py Code formatting improvements for usage tracking tests        agents-api/tests/test_usage_tracking.py • Applied consistent formatting to function calls and method invocations • Reformatted long parameter lists and test data structures • Improved readability of usage tracking test methods +28/-12  evaluator.py Code formatting improvements for expression evaluator utilities agents-api/agents_api/common/utils/evaluator.py • Applied consistent multi-line formatting to import statements and function calls • Reformatted long parameter lists and conditional statements • Improved readability of evaluator utility functions +17/-5    openapi_model.py Code formatting improvements for OpenAPI model definitions agents-api/agents_api/autogen/openapi_model.py • Applied consistent multi-line formatting to model configuration dictionaries • Reformatted long parameter lists and validation methods • Improved readability of OpenAPI model definitions +30/-18  utils.py Code formatting improvements for database query utilities agents-api/agents_api/queries/utils.py • Applied consistent multi-line formatting to function calls and data structures • Reformatted long parameter lists and conditional statements • Improved readability of database query utility functions +36/-16  tasks.py Code formatting improvements for task protocol utilities  agents-api/agents_api/common/protocol/tasks.py • Applied consistent formatting to conditional statements and function calls • Reformatted long parameter lists and method invocations • Improved readability of task protocol implementation +13/-4    helpers.py Code formatting improvements for task execution helpers    agents-api/agents_api/workflows/task_execution/helpers.py • Applied consistent multi-line formatting to import statements and function calls • Reformatted long parameter lists and workflow configurations • Improved readability of task execution helper functions +19/-7    web.py Code formatting improvements for web application utilities agents-api/agents_api/web.py • Applied consistent formatting to conditional statements and string operations • Reformatted long parameter lists and error handling logic • Improved readability of web application utilities +15/-5    gather_messages.py Code formatting improvements for chat message gathering    agents-api/agents_api/queries/chat/gather_messages.py • Applied consistent multi-line formatting to decorator and function definitions • Reformatted long parameter lists and list comprehensions • Improved readability of message gathering utilities +13/-7    test_docs_queries.py Code formatting improvements for document query tests        agents-api/tests/test_docs_queries.py • Applied consistent formatting to function calls and method invocations • Reformatted long parameter lists and test assertions • Improved readability of document query testing methods +15/-5    create_response.py Code formatting improvements for response creation router agents-api/agents_api/routers/responses/create_response.py • Applied consistent formatting to conditional statements and function calls • Reformatted long parameter lists and list comprehensions • Improved readability of response creation logic +22/-12  expressions.py Code formatting improvements for expression evaluation utilities agents-api/agents_api/common/utils/expressions.py • Reorganized import statements and moved constants after imports • Applied consistent formatting to conditional statements and function calls • Improved readability of expression evaluation utilities +11/-8    execute_api_call.py Code formatting improvements for API call execution            agents-api/agents_api/activities/execute_api_call.py • Split long lines for follow_redirects and include_response_content parameter extraction • Split long line for error body truncation in HTTP error handling • Split long line for response content base64 encoding +12/-4    test_agent_queries.py Function signature formatting improvements in agent tests agents-api/tests/test_agent_queries.py • Split long function parameter lists across multiple lines for better readability • Applied consistent line breaking to function signatures with multiple parameters +12/-4    test_chat_routes.py Line length formatting improvements in chat route tests    agents-api/tests/test_chat_routes.py • Split long lines in test assertions and variable assignments • Improved readability of template strings and assertion statements +13/-4    test_secrets_queries.py Function signature and assertion formatting in secrets tests agents-api/tests/test_secrets_queries.py • Split long function signatures across multiple lines • Improved formatting of assertion statements and metadata comparisons +12/-4    test_prepare_for_step.py Transition target constructor formatting improvements        agents-api/tests/test_prepare_for_step.py • Split long TransitionTarget constructor calls across multiple lines • Improved readability of workflow transition test setup +15/-5    tool_executor.py Tool executor constructor and expression formatting            agents-api/agents_api/activities/tool_executor.py • Split long WebPreviewToolCall constructor calls across multiple lines • Improved formatting of conditional expressions for tool output handling +9/-3      app.py Application initialization formatting improvements              agents-api/agents_api/app.py • Split long function calls for database pool creation and state checks • Improved formatting of Instrumentator method chaining +12/-4    model_converters.py Model converter formatting and exception handling improvements agents-api/agents_api/routers/utils/model_converters.py • Split long lines in HTTP exception handling and message construction • Improved formatting of dictionary and list comprehensions +14/-5    agents.py Agent exception constructor formatting improvements            agents-api/agents_api/common/exceptions/agents.py • Split long exception initialization calls across multiple lines • Improved readability of error message formatting +6/-2      interceptors.py Interceptor import and method signature formatting              agents-api/agents_api/common/interceptors.py • Split long import statements and function signatures • Improved formatting of workflow and activity interceptor methods +12/-3    test_user_queries.py User query test formatting improvements                                    agents-api/tests/test_user_queries.py • Split long function signatures and assertion statements • Improved formatting consistency across user query tests +12/-4    conftest.py Test configuration and fixture formatting improvements      agents-api/tests/conftest.py • Split long function calls and container initialization • Improved formatting of test fixture setup and teardown +12/-6    test_execution_workflow.py Execution workflow test formatting improvements                    agents-api/tests/test_execution_workflow.py • Split long list comprehensions and event filtering logic • Improved formatting of workflow execution test assertions +12/-4    test_tool_call_step.py Tool call step test parameter formatting                                  agents-api/tests/test_tool_call_step.py • Split long dictionary definitions for tool parameters • Improved formatting of UUID and tool call assertions +11/-3    litellm.py LiteLLM client formatting improvements                                      agents-api/agents_api/clients/litellm.py • Split long list comprehensions and dictionary construction • Improved formatting of HTTP headers and API client setup +6/-2      stream_execution_status.py Execution status streaming formatting improvements              agents-api/agents_api/routers/tasks/stream_execution_status.py • Split long function calls and error message formatting • Improved readability of execution status streaming logic +14/-6    prompt_step.py Prompt step conditional expression formatting                        agents-api/agents_api/activities/task_steps/prompt_step.py • Split long conditional expressions and prompt handling logic • Improved formatting of agent model selection and prompt wrapping +6/-2      test_transitions_queries.py Transition query test parameter formatting                              agents-api/tests/test_transitions_queries.py • Split long dictionary definitions and function calls • Improved formatting of transition query test parameters +13/-3    stream_transitions_events.py Transition event streaming formatting improvements              agents-api/agents_api/routers/tasks/stream_transitions_events.py • Split long attribute access chains and dictionary construction • Improved formatting of event streaming and payload handling +14/-8    humanization_utils.py Translation utility formatting improvements                            agents-api/agents_api/common/utils/humanization_utils.py • Split long GoogleTranslator constructor and method calls • Improved formatting of translation utility functions +9/-3      test_entry_queries.py Entry query test function signature formatting                      agents-api/tests/test_entry_queries.py • Split long function signatures across multiple lines • Improved formatting consistency in entry query tests +9/-3      test_session_routes.py Session route test formatting improvements                              agents-api/tests/test_session_routes.py • Split long function calls and HTTP request formatting • Improved readability of session route test methods +9/-3      test_session_queries.py Session query test function signature formatting                  agents-api/tests/test_session_queries.py • Split long function signatures across multiple lines • Improved formatting consistency in session query tests +9/-3      mmr.py MMR algorithm expression and signature formatting                agents-api/agents_api/common/utils/mmr.py • Split long mathematical expressions and function signatures • Improved formatting of maximal marginal relevance calculations +7/-2      test_model_validation.py Model validation test patch formatting                                      agents-api/tests/test_model_validation.py • Split long patch decorator paths across multiple lines • Improved formatting of model validation test setup +9/-3      create_doc.py Document creation query parameter formatting                          agents-api/agents_api/queries/docs/create_doc.py • Split long tuple definitions for query parameters • Improved formatting of document creation query construction +14/-10  render.py Session render function signature and return formatting    agents-api/agents_api/routers/sessions/render.py • Split long function signatures and return type annotations • Improved formatting of chat input rendering logic +14/-3    test_nlp_utilities.py NLP utility test parameter formatting improvements              agents-api/tests/test_nlp_utilities.py • Split long test parameter tuples and expected result sets • Improved formatting of NLP utility test cases +13/-3    workflows.py Workflow utility assertion formatting improvements              agents-api/agents_api/common/utils/workflows.py • Split long assertion statements with complex conditions • Improved readability of workflow name extraction logic +6/-6      list_execution_inputs_data.py Execution input data query exception decorator formatting agents-api/agents_api/queries/executions/list_execution_inputs_data.py • Split long @rewrap_exceptions decorator dictionary across multiple lines • Improved formatting of exception handling configuration +15/-13  list_execution_state_data.py Execution state data query exception decorator formatting agents-api/agents_api/queries/executions/list_execution_state_data.py • Split long @rewrap_exceptions decorator dictionary across multiple lines • Improved formatting of exception handling configuration +15/-13  search_docs.py Document search parameter type annotation formatting          agents-api/agents_api/routers/docs/search_docs.py • Split long union type annotations across multiple lines • Improved formatting of document search parameter types +6/-2      nlp.py NLP keyword extraction formatting improvements                      agents-api/agents_api/common/nlp.py • Split long list comprehensions and normalization logic • Improved formatting of keyword extraction and processing +6/-2      tool_call_step.py Tool call step function signature formatting                          agents-api/agents_api/activities/task_steps/tool_call_step.py • Split long function signature across multiple lines • Improved formatting of tool call construction function +3/-1      list_executions.py Execution list query exception decorator formatting            agents-api/agents_api/queries/executions/list_executions.py • Split long @rewrap_exceptions decorator dictionary across multiple lines • Improved formatting of execution list query exception handling +15/-13  test_files_queries.py File query test function signature formatting                        agents-api/tests/test_files_queries.py • Split long function signatures across multiple lines • Improved formatting consistency in file query tests +6/-2      secrets.py Secret utility function signature formatting                          agents-api/agents_api/common/utils/secrets.py • Split long function signature across multiple lines • Improved formatting of secret retrieval function +3/-1      create_task.py Task creation workflow parameter formatting                            agents-api/agents_api/queries/tasks/create_task.py • Split long list append operations across multiple lines • Improved formatting of workflow parameter construction +11/-9    codec.py Worker codec formatting improvements                                          agents-api/agents_api/worker/codec.py • Split long string formatting and version checking logic • Improved formatting of payload encoding/decoding operations +6/-2      test_expression_validation.py Expression validation test assertion formatting                    agents-api/tests/test_expression_validation.py • Split long assertion statements with complex conditions • Improved formatting of expression validation test assertions +6/-2      create_usage_record.py Usage record creation formatting improvements                        agents-api/agents_api/queries/usage/create_usage_record.py • Split long conditional expressions and print statements • Improved formatting of usage cost calculation logic +6/-2      update_execution.py Execution update formatting improvements                                  agents-api/agents_api/routers/tasks/update_execution.py • Split long function calls and exception handling • Improved formatting of execution update logic +6/-2      developer_id.py Developer ID dependency formatting improvements                    agents-api/agents_api/dependencies/developer_id.py • Split long assertion statements and function calls • Improved formatting of developer ID dependency logic +6/-2      base_evaluate.py Base evaluate function call formatting                                      agents-api/agents_api/activities/task_steps/base_evaluate.py • Split long function call across multiple lines • Improved formatting of expression evaluation function call +3/-1      create_or_update_task.py Task create/update workflow parameter formatting                  agents-api/agents_api/queries/tasks/create_or_update_task.py • Split long list append operations across multiple lines • Improved formatting of workflow parameter construction +10/-8    patch_task.py Task patch workflow parameter formatting                                  agents-api/agents_api/queries/tasks/patch_task.py • Split long list append operations across multiple lines • Improved formatting of workflow parameter construction +10/-8    update_task.py Task update workflow parameter formatting                                agents-api/agents_api/queries/tasks/update_task.py • Split long list append operations across multiple lines • Improved formatting of workflow parameter construction +10/-8    delete_doc.py Document deletion route decorator formatting                          agents-api/agents_api/routers/docs/delete_doc.py • Split long decorator lines across multiple lines • Improved formatting of route decorator definitions +6/-2      create_or_update_task.py Task create/update route formatting improvements                  agents-api/agents_api/routers/tasks/create_or_update_task.py • Split long decorator and exception handling lines • Improved formatting of task creation route logic +6/-2      sync_s3.py S3 client configuration formatting improvements                    agents-api/agents_api/clients/sync_s3.py • Split long configuration dictionary across multiple lines • Improved formatting of S3 client configuration +3/-1      users.py User exception constructor formatting                                        agents-api/agents_api/common/exceptions/users.py • Split long exception initialization call across multiple lines • Improved formatting of user exception constructor +3/-1      chat.py Chat session response formatting improvements                        agents-api/agents_api/routers/sessions/chat.py • Split long dictionary access and conditional expressions • Improved formatting of chat response streaming logic +6/-2      get_response.py Response retrieval formatting improvements                              agents-api/agents_api/routers/responses/get_response.py • Split long exception and object construction calls • Improved formatting of response retrieval logic +6/-2
Tests	1 files .pytest-runtimes Updated test runtime measurements showing performance improvements agents-api/.pytest-runtimes • Updated test runtime measurements for all test cases • Most test execution times have been reduced significantly • Some tests show dramatic improvements (e.g., test_chat_check_that_render_route_works_and_does_not_call_completion_mock from 80.86s to 0.29s) • Overall pattern suggests performance improvements across the test suite +286/-286
Additional files	43 files execute_integration.py +3/-1      temporal.py +5/-3      sessions.py +3/-1      template.py +2/-1      types.py +3/-1      usage.py +3/-1      auth.py +3/-1      query_filter.py +3/-1      counters.py +2/-10    create_agent.py +3/-1      create_or_update_agent.py +3/-1      list_docs.py +1/-3      utils.py +3/-1      get_history.py +3/-1      list_entries.py +3/-1      create_execution.py +3/-1      create_execution_transition.py +3/-1      prepare_execution_input.py +3/-1      create_file.py +7/-1      list_files.py +3/-1      list.py +3/-1      create_or_update_session.py +3/-1      create_tools.py +3/-1      entities.py +4/-1      summarize.py +4/-1      trim.py +4/-1      utils.py +3/-1      create_file.py +3/-1      delete_file.py +3/-1      check_health.py +6/-1      delete_session.py +6/-2      create_task.py +3/-1      create_task_execution.py +3/-1      pyproject.toml +3/-0      test_find_selector.py +3/-1      test_activities_utils.py +3/-1      test_execution_queries.py +3/-1      test_file_routes.py +3/-1      test_get_doc_search.py +4/-1      test_metadata_filter_utils.py +3/-1      test_secrets_routes.py +3/-1      test_task_queries.py +3/-1      uuid_extensions.pyi +0/-7

---

Need help?

Type /help how to ... in the comments thread for any questions about Qodo Merge usage.

Check out the documentation for more information.

[claude]

Claude finished @creatorrr's task —— View job

---

PR Review: Switch to psycopg from asyncpg

Reviewing this major database driver migration across 100 files. This is a significant infrastructure change that requires careful analysis.

Review Checklist:

• Read repository guidelines and setup instructions
• Examine database connection and configuration changes
• Review query execution and transaction handling
• Analyze error handling and exception management
• Check connection pooling and resource management
• Evaluate performance implications
• Review security considerations
• Assess test coverage and changes
• Provide comprehensive feedback and recommendations

---

🚨 Critical Finding: Migration is Incomplete

Despite the PR title indicating "Switch to psycopg from asyncpg", the actual migration has not been completed. The codebase still predominantly uses asyncpg:

Evidence of Incomplete Migration:

1. Main Dependencies Still asyncpg:

   • pyproject.toml:50 - asyncpg>=0.30.0 remains as main dependency
   • pyproject.toml:81 - psycopg[binary]>=3.2.5 only in dev dependencies with comment "only for use inside tests for now"

2. Core Database Client Still Uses asyncpg:

   • agents_api/clients/pg.py:3 - import asyncpg
   • agents_api/clients/pg.py:19 - await asyncpg.create_pool()
   • agents_api/app.py:10 - from asyncpg.pool import Pool

3. Query Infrastructure Built on asyncpg:

   • agents_api/queries/utils.py:17 - import asyncpg
   • agents_api/queries/utils.py:138 - cast(asyncpg.Pool, getattr(app.state, "postgres_pool", None))
   • agents_api/queries/utils.py:165 - raise asyncpg.NoDataFoundError(msg)

4. Exception Handling for asyncpg:

   • agents_api/common/utils/db_exceptions.py - Entire module designed around asyncpg exceptions (71+ references)

---

📋 Code Quality Assessment

✅ Positive Aspects:

1. SQL Builder Addition: The new sql_builder.py module provides good utilities for dynamic query construction with proper parameterization to prevent SQL injection.

2. Type Safety Improvements: Many files show improved type annotations and imports.

3. Consistent Error Handling: Exception handling patterns are well-structured with hierarchical, typed exceptions.

4. Test Infrastructure: Comprehensive test coverage maintained with proper fixtures.

⚠️ Areas of Concern:

1. Misleading PR Title: The migration is not actually happening, which creates confusion about the PR's purpose.

2. Mixed Driver References: Some files mention psycopg in comments but use asyncpg in implementation.

3. Incomplete Documentation: No clear migration strategy or timeline documented.

---

🔍 Technical Analysis

Performance Considerations:

• Cannot assess performance impact since the actual driver switch hasn't occurred
• Current asyncpg implementation appears well-optimized with connection pooling
• psycopg 3.x should provide similar performance characteristics to asyncpg

Security Review:

• ✅ SQL injection protection via parameterized queries maintained
• ✅ Connection string handling secure
• ✅ No hardcoded credentials found
• ✅ Proper error handling doesn't leak sensitive information

Compatibility Issues:

• API Differences: asyncpg and psycopg have different APIs that will require significant changes:
  • Connection creation (asyncpg.create_pool() vs psycopg_pool.AsyncConnectionPool())
  • Query execution methods differ
  • Exception types are different
  • Type codec registration varies

---

🧪 Test Coverage

• Test fixtures still use asyncpg: conftest.py creates pools using create_db_pool which uses asyncpg
• Test runtime changes: The .pytest-runtimes file shows significant changes (286 lines modified), suggesting test behavior differences
• psycopg only in dev dependencies: Suggests testing-only usage rather than production migration

---

💡 Recommendations

Immediate Actions:

1. Clarify PR Intent:

   • Update PR title and description to reflect actual changes being made
   • If this is preparation work, make that explicit
   • If this is an incomplete migration, provide a migration plan

2. Resolve Inconsistencies:

   • Either complete the migration or remove psycopg references
   • Update comments that incorrectly reference the new driver

If Continuing Migration:

1. Create Migration Plan:

   • Document API differences between asyncpg and psycopg
   • Plan staged migration approach
   • Create compatibility layer if needed

2. Update Core Components:

   • Replace clients/pg.py to use psycopg
   • Update connection pool management
   • Migrate exception handling in db_exceptions.py
   • Update query utilities in utils.py

3. Testing Strategy:

   • Run both drivers in parallel during transition
   • Comprehensive integration testing
   • Performance benchmarking

If This is Preparation Work:

1. Document Intent: Add clear comments explaining this is prep work
2. Version Dependencies: Consider feature flags for driver selection
3. Plan Timeline: Provide roadmap for actual migration

---

🎯 Verdict

❌ Do Not Merge As-Is: This PR creates confusion by claiming to switch database drivers while not actually doing so. The changes appear to be either:

• Incomplete migration work that needs completion
• Preparation work that needs better documentation
• Accidental inclusion of psycopg without actual migration

Next Steps: Clarify the intent, either complete the migration or update the PR scope to reflect the actual changes being made.

[qodo-merge-for-open-source]

CI Feedback 🧐

(Feedback updated until commit d3888cf)

A test triggered by this PR failed. Here is an AI-generated analysis of the failure:

Action: Typecheck

Failed stage: Typecheck [❌]

Failure summary: The action failed due to pytype static type checking errors: • Invalid type annotation in agents_api/queries/sql_builder.py:296:23 - callable should be Callable (line 1674) • Multiple "not-callable" errors in tests/test_transitions_queries.py where functions list_execution_inputs_data and list_execution_state_data from agents_api.queries.executions are being called but pytype thinks they are not callable (lines 1877-1957) • The build process stopped due to these type checking failures preventing further progress

Relevant error logs: 1: ##[group]Operating System 2: Ubuntu ... 978: prune-cache: true 979: ignore-nothing-to-cache: false 980: ##[endgroup] 981: Downloading uv from "https://github.com/astral-sh/uv/releases/download/0.7.14/uv-x86_64-unknown-linux-gnu.tar.gz" ... 982: [command]/usr/bin/tar xz --warning=no-unknown-keyword --overwrite -C /home/runner/work/_temp/33950ac2-4da7-4bff-b065-59971f16119b -f /home/runner/work/_temp/467a726a-e055-42bc-bdb5-4aff4021638f 983: Added /opt/hostedtoolcache/uv/0.7.14/x86_64 to the path 984: Added /home/runner/.local/bin to the path 985: Set UV_CACHE_DIR to /home/runner/work/_temp/setup-uv-cache 986: Successfully installed uv version 0.7.14 987: Searching files using cache dependency glob: **/uv.lock 988: /home/runner/work/julep/julep/agents-api/uv.lock 989: /home/runner/work/julep/julep/cli/uv.lock 990: /home/runner/work/julep/julep/integrations-service/uv.lock 991: Found 3 files to hash. 992: Trying to restore uv cache from GitHub Actions cache with key: setup-uv-1-x86_64-unknown-linux-gnu-0.7.14-19a373a92bf560cb56116522380dcdd14cf478537e731f07ba47c25f51b31e85 993: ##[warning]Failed to restore: getCacheEntry failed: Cache service responded with 503 994: No GitHub Actions cache found for key: setup-uv-1-x86_64-unknown-linux-gnu-0.7.14-19a373a92bf560cb56116522380dcdd14cf478537e731f07ba47c25f51b31e85 ... 1657: [16/370] check agents_api.autogen.Jobs 1658: [17/370] check agents_api.autogen.Files 1659: [18/370] check agents_api.autogen.Executions 1660: [19/370] check agents_api.autogen.Entries 1661: [20/370] check agents_api.autogen.Agents 1662: [21/370] check agents_api.autogen.Chat 1663: [22/370] check agents_api.common.utils.datetime 1664: [23/370] check agents_api.common.exceptions.validation 1665: [24/370] check agents_api.app 1666: [25/370] check agents_api.worker.codec 1667: [26/370] check agents_api.common.utils.db_exceptions 1668: [27/370] check agents_api.autogen.openapi_model 1669: [28/370] check agents_api.queries.docs.utils 1670: [29/370] check agents_api.queries.utils 1671: [30/370] check agents_api.queries.sql_builder 1672: FAILED: /home/runner/work/julep/julep/agents-api/.pytype/pyi/agents_api/queries/sql_builder.pyi 1673: /home/runner/work/julep/julep/agents-api/.venv/bin/python -m pytype.main --disable pyi-error --imports_info /home/runner/work/julep/julep/agents-api/.pytype/imports/agents_api.queries.sql_builder.imports --module-name agents_api.queries.sql_builder --platform linux -V 3.12 -o /home/runner/work/julep/julep/agents-api/.pytype/pyi/agents_api/queries/sql_builder.pyi --analyze-annotated --nofail --none-is-not-bool --quick --strict-none-binding /home/runner/work/julep/julep/agents-api/agents_api/queries/sql_builder.py 1674: /home/runner/work/julep/julep/agents-api/agents_api/queries/sql_builder.py:296:23: error: in <module>: Invalid type annotation '<instance of Callable>' [invalid-annotation] 1675: Not a type 1676: special_handlers: dict[str, callable] | None = None, 1677: ~~~~~~~~~~~~~~~~~~~ 1678: For more details, see https://google.github.io/pytype/errors.html#invalid-annotation 1679: [31/370] check agents_api.metrics.counters ... 1860: [212/370] check agents_api.routers.internal.__init__ 1861: [213/370] check agents_api.routers.files.__init__ 1862: [214/370] check agents_api.rec_sum.utils 1863: [215/370] check agents_api.workflows.demo 1864: [216/370] check agents_api.rec_sum.data 1865: [217/370] check tests.test_user_routes 1866: [218/370] check agents_api.common.protocol.state_machine 1867: [219/370] check tests.test_secrets_routes 1868: [220/370] check agents_api.metrics.__init__ 1869: [221/370] check tests.test_files_queries 1870: [222/370] check agents_api.dependencies.content_length 1871: [223/370] check tests.test_query_utils 1872: [224/370] check tests.test_usage_tracking 1873: [225/370] check tests.test_task_queries 1874: [226/370] check tests.test_transitions_queries 1875: FAILED: /home/runner/work/julep/julep/agents-api/.pytype/pyi/tests/test_transitions_queries.pyi 1876: /home/runner/work/julep/julep/agents-api/.venv/bin/python -m pytype.main --disable pyi-error --imports_info /home/runner/work/julep/julep/agents-api/.pytype/imports/tests.test_transitions_queries.imports --module-name tests.test_transitions_queries --platform linux -V 3.12 -o /home/runner/work/julep/julep/agents-api/.pytype/pyi/tests/test_transitions_queries.pyi --analyze-annotated --nofail --none-is-not-bool --quick --strict-none-binding /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py 1877: /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py:98:25: error: in test_query_list_execution_inputs_data: 'agents_api.queries.executions.list_execution_inputs_data' object is not callable [not-callable] 1878: transitions = await list_execution_inputs_data( 1879: ~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1880: execution_id=execution.id, 1881: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1882: scope_id=scope_id, 1883: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1884: direction="asc", 1885: ~~~~~~~~~~~~~~~~~~~~~~~~ 1886: connection_pool=pool, 1887: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1888: ) 1889: ~~~~~ 1890: /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py:150:25: error: in test_query_list_execution_state_data: 'agents_api.queries.executions.list_execution_state_data' object is not callable [not-callable] 1891: transitions = await list_execution_state_data( 1892: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1893: execution_id=execution.id, 1894: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1895: scope_id=scope_id, 1896: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1897: direction="asc", 1898: ~~~~~~~~~~~~~~~~~~~~~~~~ 1899: connection_pool=pool, 1900: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1901: ) 1902: ~~~~~ 1903: /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py:296:44: error: in test_query_list_execution_inputs_data_search_window: 'agents_api.queries.executions.list_execution_inputs_data' object is not callable [not-callable] 1904: transitions_with_search_window = await list_execution_inputs_data( 1905: ~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1906: execution_id=execution_id, 1907: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1908: scope_id=scope_id, 1909: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1910: direction="asc", 1911: ~~~~~~~~~~~~~~~~~~~~~~~~ 1912: connection_pool=pool, 1913: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1914: search_window=timedelta(days=5), 1915: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1916: ) 1917: ~~~~~ 1918: /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py:307:47: error: in test_query_list_execution_inputs_data_search_window: 'agents_api.queries.executions.list_execution_inputs_data' object is not callable [not-callable] 1919: transitions_without_search_window = await list_execution_inputs_data( 1920: ~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1921: execution_id=execution_id, 1922: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1923: scope_id=scope_id, 1924: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1925: direction="asc", 1926: ~~~~~~~~~~~~~~~~~~~~~~~~ 1927: connection_pool=pool, 1928: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1929: ) 1930: ~~~~~ 1931: /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py:365:44: error: in test_query_list_execution_state_data_search_window: 'agents_api.queries.executions.list_execution_state_data' object is not callable [not-callable] 1932: transitions_with_search_window = await list_execution_state_data( 1933: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1934: execution_id=execution_id, 1935: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1936: scope_id=scope_id, 1937: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1938: direction="asc", 1939: ~~~~~~~~~~~~~~~~~~~~~~~~ 1940: connection_pool=pool, 1941: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1942: search_window=timedelta(days=5), 1943: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1944: ) 1945: ~~~~~ 1946: /home/runner/work/julep/julep/agents-api/tests/test_transitions_queries.py:376:47: error: in test_query_list_execution_state_data_search_window: 'agents_api.queries.executions.list_execution_state_data' object is not callable [not-callable] 1947: transitions_without_search_window = await list_execution_state_data( 1948: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1949: execution_id=execution_id, 1950: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1951: scope_id=scope_id, 1952: ~~~~~~~~~~~~~~~~~~~~~~~~~~ 1953: direction="asc", 1954: ~~~~~~~~~~~~~~~~~~~~~~~~ 1955: connection_pool=pool, 1956: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1957: ) 1958: ~~~~~ 1959: For more details, see https://google.github.io/pytype/errors.html#not-callable 1960: [227/370] check tests.test_session_queries ... 1987: [254/370] check tests.test_agent_metadata_filtering 1988: [255/370] check agents_api.routers.sessions.exceptions 1989: [256/370] check agents_api.common.exceptions.users 1990: [257/370] check tests.test_session_routes 1991: [258/370] check tests.test_middleware 1992: [259/370] check agents_api.activities.__init__ 1993: [260/370] check tests.test_metadata_filter_utils 1994: [261/370] check agents_api.common.utils.json 1995: [262/370] check tests.test_nlp_utilities 1996: [263/370] check agents_api.clients.__init__ 1997: [264/370] check tests.test_developer_queries 1998: [265/370] check tests.test_usage_cost 1999: [266/370] check agents_api.common.__init__ 2000: [267/370] check tests.test_get_doc_search 2001: [268/370] check tests.test_memory_utils 2002: ninja: build stopped: cannot make progress due to previous errors. 2003: Generated API key since not set in the environment: 65126120249813900152582112654768 2004: Generated API key since not set in the environment: 55716488390152661164747838795281 2005: Generated API key since not set in the environment: 67513108634390592355877737711932 2006: Analyzing 357 sources with 0 local dependencies 2007: Leaving directory '.pytype' 2008: ##[error]Process completed with exit code 1. 2009: Post job cleanup.

[qodo-merge-for-open-source]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 4 🔵🔵🔵🔵⚪

🧪 PR contains tests

🔒 Security concerns SQL injection: The new SQLBuilder class introduces a safe_identifier method that uses basic validation (only alphanumeric, underscore, dot characters) but may not be robust enough for production use. The method could potentially allow sophisticated SQL injection attacks to bypass the simple character filtering. Additionally, the manual parameter placeholder management across different methods could lead to parameter numbering conflicts that might be exploitable.

⚡ Recommended focus areas for review SQL Injection Risk The safe_identifier method uses basic validation that may not be sufficient for production use. The comment acknowledges this limitation but the implementation could still allow malicious identifiers to pass through, especially with complex SQL injection techniques. def safe_identifier(self, name: str) -> str: """ Safely quote an SQL identifier (table/column name). Note: This is a simple implementation. In production, consider using a more robust identifier escaping mechanism. """ # Special case for SELECT * if name == "*": return name # Basic validation - only allow alphanumeric, underscore, dot if not name.replace("_", "").replace(".", "").isalnum(): msg = f"Invalid identifier: {name}" raise ValueError(msg) # Quote the identifier return f'"{name}"' Type Safety Issue The function signature indicates it can return either a single tuple or a list of tuples, but the implementation only returns single tuples. This inconsistency could lead to runtime errors when callers expect the list format. ) -> tuple[str, list] | list[tuple[str, list]]: """ Parameter Handling The build_limit_offset method manually calculates parameter indices instead of using the class's parameter counter, which could lead to parameter numbering conflicts when used with other methods. def build_limit_offset( self, limit: int | None = None, offset: int | None = None ) -> tuple[str, list[Any]]: """ Build LIMIT/OFFSET clause. Args: limit: Maximum number of rows offset: Number of rows to skip Returns: Tuple of (LIMIT/OFFSET clause, parameters) """ clause_parts = [] params = [] param_idx = len(self.params) + 1 if limit is not None: clause_parts.append(f"LIMIT ${param_idx}") params.append(limit) param_idx += 1 if offset is not None: clause_parts.append(f"OFFSET ${param_idx}") params.append(offset) return " " + " ".join(clause_parts) if clause_parts else "", params

[qodo-merge-for-open-source]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
Security	Validate SQL operator parameter The operator parameter is not validated, allowing potential SQL injection through malicious operators. Only specific operators like 'AND' and 'OR' should be allowed. agents-api/agents_api/queries/sql_builder.py [86-114] def build_where_clause( self, conditions: dict[str, Any], operator: str = "AND" ) -> tuple[str, list[Any]]: """ Build WHERE clause from conditions. Args: conditions: Dictionary of field_name -> value operator: Logical operator (AND/OR) Returns: Tuple of (WHERE clause, parameters list) """ + # Validate operator + if operator.upper() not in ("AND", "OR"): + msg = f"Invalid operator: {operator}" + raise ValueError(msg) + where_parts = [] for field_name, value in conditions.items(): if value is None: where_parts.append(f"{self.safe_identifier(field_name)} IS NULL") else: placeholder = self.add_param(value) where_parts.append( f"{self.safe_identifier(field_name)} = {placeholder}" ) if not where_parts: return "", [] - return f" WHERE {f' {operator} '.join(where_parts)}", self.params + return f" WHERE {f' {operator.upper()} '.join(where_parts)}", self.params [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 10 __ Why: This suggestion correctly points out a critical SQL injection vulnerability. The operator parameter is used directly in the query string without validation, allowing arbitrary SQL to be injected. The proposed fix of whitelisting allowed operators ('AND', 'OR') is the correct approach to mitigate this high-severity security risk.	High
	Strengthen SQL identifier validation The identifier validation is insufficient and could allow SQL injection. The current regex allows dots which could be exploited, and the validation doesn't handle edge cases like empty strings or reserved keywords. agents-api/agents_api/queries/sql_builder.py [39-57] def safe_identifier(self, name: str) -> str: """ Safely quote an SQL identifier (table/column name). Note: This is a simple implementation. In production, consider using a more robust identifier escaping mechanism. """ # Special case for SELECT * if name == "*": return name - # Basic validation - only allow alphanumeric, underscore, dot - if not name.replace("_", "").replace(".", "").isalnum(): + # Validate identifier format more strictly + if not name or not re.match(r'^[a-zA-Z_][a-zA-Z0-9_]*$', name): msg = f"Invalid identifier: {name}" raise ValueError(msg) # Quote the identifier return f'"{name}"' [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 9 __ Why: The suggestion correctly identifies a security weakness in the safe_identifier function. The existing validation allows dots (.), which could be exploited for unauthorized schema or table access. The proposed regex provides a much stricter and safer validation for SQL identifiers, significantly improving the security of the dynamic query builder.	High
More

[ellipsis-dev]

Caution

Changes requested ❌

Reviewed everything up to d3888cf in 3 minutes and 35 seconds. Click for details.

• Reviewed 6325 lines of code in 127 files
• Skipped 0 files when reviewing.
• Skipped posting 9 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. agents-api/agents_api/common/utils/expressions.py:15

• Draft comment:
  Constants were relocated in the file. Ensure the new position (after some imports) is intentional and that evaluation functions using these values are not affected by initialization order.
• Reason this comment was not posted:
  Comment was on unchanged code.

2. agents-api/agents_api/queries/agents/list_agents.py:45

• Draft comment:
  The ORDER BY clause and parameter placeholders have been re-indexed. Verify that the merged metadata filter (via build_metadata_filter_conditions) preserves the expected parameter order, so that $2–$5 still map correctly to sort_by, direction, limit, and offset.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% The comment is asking the PR author to verify the behavior of the code, which violates the rule against asking the author to confirm or verify intentions. It does not provide a specific suggestion or point out a specific issue with the code.

3. agents-api/agents_api/queries/executions/list_execution_transitions.py:126

• Draft comment:
  Scope filter condition uses placeholder $7 when scope_id is provided. Confirm that appending scope_id maintains consistent parameter indexing compared to the base params list.
• Reason this comment was not posted:
  Comment looked like it was already resolved.

4. agents-api/agents_api/queries/sql_builder.py:39

• Draft comment:
  The safe_identifier function uses a basic alphanumeric check for SQL identifiers. Consider a more robust escaping mechanism to safely handle identifiers that may include reserved words or special characters.
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable: usefulness confidence = 0% vs. threshold = 50% The comment raises a valid security concern, but the code's docstring already explicitly acknowledges this limitation with "Note: This is a simple implementation. In production, consider using a more robust identifier escaping mechanism." The comment is essentially repeating what's already documented in the code itself. The comment could be valuable if it provided specific suggestions for better escaping mechanisms or pointed out concrete vulnerabilities in the current approach. Since the code already acknowledges this limitation and the comment doesn't provide additional actionable information beyond what's in the docstring, it's not adding value. Delete the comment since it's redundant with the existing docstring and doesn't provide additional actionable guidance.

5. agents-api/agents_api/queries/sql_builder.py:315

• Draft comment:
  build_patch_query falls back to a SELECT query when there are no fields to update. Ensure this behavior is well documented and that downstream code properly handles the SELECT as an alternative to an UPDATE.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% The comment suggests ensuring that a specific behavior is documented and handled properly in downstream code. This falls under asking the PR author to ensure something is done, which is against the rules. The comment does not provide a specific code suggestion or ask for a specific test to be written.

6. agents-api/agents_api/queries/users/patch_user.py:146

• Draft comment:
  Similar to patch_agent, metadata is merged via a string replace in the update query. This approach is brittle; consider integrating the merge operator directly when building the query to avoid unintended replacements.
• Reason this comment was not posted:
  Marked as duplicate.

7. agents-api/agents_api/queries/users/patch_user.py:156

• Draft comment:
  The final parameters combine developer_id and the project name (defaulting to 'default' if not provided). Verify that this defaulting behavior is consistent with other parts of the codebase, ensuring a uniform handling of missing project information.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% The comment is asking the PR author to verify the consistency of defaulting behavior across the codebase, which is against the rules. It doesn't provide a specific suggestion or point out a specific issue with the code.

8. agents-api/agents_api/common/utils/expressions.py:98

• Draft comment:
  The comment reads "frozen_box doesn't work coz we need some mutability in the values." Consider using a more formal alternative such as "because" instead of "coz" to improve clarity and readability.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% The comment is focused on suggesting a change in language for clarity and readability, which is not directly related to the functionality or correctness of the code. It doesn't provide a specific code suggestion or highlight a potential issue with the code itself. Therefore, it doesn't align with the rules for useful comments.

9. agents-api/tests/test_prepare_for_step.py:176

• Draft comment:
  It looks like there's an unintended use of backticks in the string literal for the workflow on this line: "subworkflow[0].if_else.then". Should this be "subworkflow[0].if_else.then" instead?
• Reason this comment was not posted:
  Comment was not on a location in the diff, so it can't be submitted as a review comment.

Workflow ID: wflow_fazGq7TclCMtpS8F

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[ellipsis-dev]

The update query is now built using build_patch_query with a param_offset of 2, and the metadata field is modified by a string replace. This string substitution for the JSONB merge operator is fragile; consider a more robust mechanism to ensure only the intended clause is altered.

[ellipsis-dev]

In build_limit_offset, parameters are managed manually. For consistency and to avoid potential parameter index conflicts, consider using the builder's add_param method instead of manual index tracking.