v0.5.13 - Critical Feature 051 API Export Fix

Critical Bug Fix

CRITICAL: v0.5.11 and v0.5.12 did not export the Feature 051 API from the common module's __init__.py file.

This release makes Feature 051 Simplified Connection API actually usable for the first time in PyPI packages.

What Was Broken

In v0.5.11 and v0.5.12, the Feature 051 code existed but was not exported:

• iris_vector_rag/common/__init__.py was empty (only # common module comment)
• from iris_vector_rag.common import get_iris_connection would fail with ImportError
• Users could not access the new simplified connection API

What's Fixed

• ✅ Added proper exports to iris_vector_rag/common/__init__.py
• ✅ Feature 051 API now accessible via standard imports
• ✅ All three API functions exported: get_iris_connection(), detect_iris_edition(), IRISConnectionPool

Changes Since v0.5.12

1. Fixed Module Exports (commit 53af704)
   • Added get_iris_connection() export
   • Added detect_iris_edition() export
   • Added IRISConnectionPool export

Changes Since v0.5.11

1. Removed Non-existent _DBAPI References (v0.5.12 - commit 08f8cd7b)

   • Removed incorrect iris._DBAPI checks from iris_connection.py:66-78
   • Removed incorrect iris._DBAPI checks from iris_dbapi_connector.py:109-123
   • Updated to use official iris.connect() API only (per InterSystems documentation)

2. Fixed Module Exports (v0.5.13 - commit 53af704)

   • Added proper exports to iris_vector_rag/common/__init__.py

Migration from v0.5.11/v0.5.12

If you tried to use Feature 051 in v0.5.11 or v0.5.12 and got import errors:

# This now works in v0.5.13 (was broken in v0.5.11/v0.5.12)
from iris_vector_rag.common import get_iris_connection, detect_iris_edition, IRISConnectionPool

# Get connection (automatic caching, thread-safe)
conn = get_iris_connection()

# Optional: Use connection pooling for high-concurrency scenarios
pool = IRISConnectionPool(max_connections=10)
with pool.acquire() as conn:
    # Use connection
    pass

Verification

Package verified at: https://pypi.org/project/iris-vector-rag/0.5.13/

References

• Feature 051 Documentation: specs/051-simplify-iris-connection/quickstart.md
• Migration Guide: /Users/tdyar/ws/hipporag2-pipeline/MIGRATION_FEATURE_051.md
• Contract Tests: tests/contract/test_iris_connection_contract.py (12 passing, 4 skipped)

v0.5.12: Fix _DBAPI References

Bug Fix Release

Fixed

• Removed non-existent _DBAPI module references from connection code
• Simplified to use official iris.connect() API as documented by InterSystems
• Cleaned up unnecessary module checks in iris_connection.py and iris_dbapi_connector.py

Technical Details

• Files Modified:
  • iris_vector_rag/common/iris_connection.py (lines 66-78)
  • iris_vector_rag/common/iris_dbapi_connector.py (lines 109-123)
• API Used: Official InterSystems IRIS DBAPI via import iris + iris.connect()
• Documentation: https://docs.intersystems.com/irislatest/csp/docbook/DocBook.UI.Page.cls?KEY=BPYNAT_pyapi
• Testing: All contract tests pass (12 passed, 4 skipped)

Installation

pip install iris-vector-rag==0.5.12
# or upgrade from previous version
pip install --upgrade iris-vector-rag

Compatibility

• Fully backward compatible with v0.5.11
• No breaking changes to public API
• Works with intersystems-irispython v5.1.2+ and v5.3.0+

PyPI

https://pypi.org/project/iris-vector-rag/0.5.12/

v0.5.8 - Entity Storage FK Validation Fix

Critical Bug Fix Release

Fixed

• Entity Storage FK Validation: Resolved critical foreign key validation failures in batch entity storage. Relationships now correctly use entity.id (UUID) instead of entity.text (entity names) as foreign keys, eliminating "20 missing entity IDs, 30 orphaned relationships" errors
• Batch Entity Processing: Fixed entity ID mismatch in GraphRAGPipeline relationship creation that caused systematic FK validation failures during knowledge graph construction
• hipporag2 Compatibility: Enhanced batch storage reliability for downstream pipelines like hipporag2-pipeline

Changed

• iris-vector-graph Dependency: Updated minimum version to >=1.1.7 to leverage upstream PPR connection type fixes and improved graph traversal reliability

Impact

This release resolves a critical bug that prevented proper entity-relationship referential integrity during knowledge graph construction. The fix ensures that all relationships reference valid entity UUIDs, eliminating orphaned relationships and FK validation failures.

Upgrade Notes

• Update to iris-vector-graph>=1.1.7
• No breaking changes to existing APIs
• Existing knowledge graphs remain compatible

Related Issues

• Fixes entity storage issues discovered during hipporag2-pipeline integration
• Resolves systematic FK validation failures in GraphRAG workflows