full ingestion -> OCR -> extraction flow is now working correctly.

libs/app_factory.py

@@ -2,7 +2,7 @@
 
 # FILE: libs/app_factory.py
 
-from collections.abc import AsyncIterator
+from collections.abc import AsyncIterator, Awaitable, Callable
 from contextlib import asynccontextmanager
 from typing import Any
 
@@ -36,6 +36,8 @@ def create_app(  # pylint: disable=too-many-arguments,too-many-positional-argume
     version: str = "1.0.0",
     settings_class: type[BaseAppSettings] = BaseAppSettings,
     custom_settings: dict[str, Any] | None = None,
+    startup_hooks: list[Callable[[], Awaitable[None]]] | None = None,
+    shutdown_hooks: list[Callable[[], Awaitable[None]]] | None = None,
 ) -> tuple[FastAPI, BaseAppSettings]:
     """Create a FastAPI application with standard configuration"""
 
@@ -56,8 +58,14 @@ def create_app(  # pylint: disable=too-many-arguments,too-many-positional-argume
     ) -> AsyncIterator[None]:  # pylint: disable=unused-argument
         # Startup
         setup_observability(settings)
+        if startup_hooks:
+            for hook in startup_hooks:
+                await hook()
         yield
         # Shutdown
+        if shutdown_hooks:
+            for hook in shutdown_hooks:
+                await hook()
 
     # Create FastAPI app
     app = FastAPI(

libs/events/topics.py

@@ -4,15 +4,15 @@
 class EventTopics:  # pylint: disable=too-few-public-methods
     """Standard event topic names"""
 
-    DOC_INGESTED = "doc.ingested"
-    DOC_OCR_READY = "doc.ocr_ready"
-    DOC_EXTRACTED = "doc.extracted"
-    KG_UPSERT_READY = "kg.upsert.ready"
-    KG_UPSERTED = "kg.upserted"
-    RAG_INDEXED = "rag.indexed"
-    CALC_SCHEDULE_READY = "calc.schedule_ready"
-    FORM_FILLED = "form.filled"
-    HMRC_SUBMITTED = "hmrc.submitted"
-    REVIEW_REQUESTED = "review.requested"
-    REVIEW_COMPLETED = "review.completed"
-    FIRM_SYNC_COMPLETED = "firm.sync.completed"
+    DOC_INGESTED = "doc_ingested"
+    DOC_OCR_READY = "doc_ocr_ready"
+    DOC_EXTRACTED = "doc_extracted"
+    KG_UPSERT_READY = "kg_upsert_ready"
+    KG_UPSERTED = "kg_upserted"
+    RAG_INDEXED = "rag_indexed"
+    CALC_SCHEDULE_READY = "calc_schedule_ready"
+    FORM_FILLED = "form_filled"
+    HMRC_SUBMITTED = "hmrc_submitted"
+    REVIEW_REQUESTED = "review_requested"
+    REVIEW_COMPLETED = "review_completed"
+    FIRM_SYNC_COMPLETED = "firm_sync_completed"

libs/requirements-base.txt

@@ -11,7 +11,7 @@ psycopg2-binary>=2.9.11
 neo4j>=6.0.2
 redis[hiredis]>=6.4.0
 
-minio>=7.2.18
+minio==7.2.18
 boto3>=1.34.0
 qdrant-client>=1.15.1
 

libs/schemas/events.py

@@ -72,22 +72,23 @@ class DocumentExtractedEventData(BaseEventData):
     """Event emitted when field extraction is complete."""
 
     doc_id: str = Field(..., description="Document identifier")
+    tenant_id: str = Field(..., description="Tenant identifier")
     extraction_id: str = Field(..., description="Unique extraction run identifier")
     strategy: Literal["llm", "rules", "hybrid"] = Field(
         ..., description="Extraction strategy used"
     )
-    fields_extracted: int = Field(..., ge=0, description="Number of fields extracted")
-    confidence_avg: float = Field(
-        ..., ge=0.0, le=1.0, description="Average extraction confidence"
+    field_count: int = Field(..., ge=0, description="Number of fields extracted")
+    confidence: float = Field(
+        ..., ge=0.0, le=1.0, description="Extraction confidence score"
     )
-    calibrated_confidence: float = Field(
-        ..., ge=0.0, le=1.0, description="Calibrated confidence score"
+    extraction_results: dict[str, Any] = Field(
+        ..., description="Full extraction results including provenance"
     )
     model_name: str | None = Field(None, description="LLM model used (if applicable)")
-    processing_time_ms: int = Field(
-        ..., ge=0, description="Processing time in milliseconds"
+    processing_time_ms: int | None = Field(
+        None, ge=0, description="Processing time in milliseconds"
     )
-    storage_path: str = Field(..., description="Path to extraction results")
+    storage_path: str | None = Field(None, description="Path to extraction results")
 
 
 # Knowledge Graph events

libs/security/dependencies.py

@@ -41,6 +41,11 @@ def get_current_tenant(request: Request) -> str | None:
         if role.startswith("tenant:"):
             return str(role.split(":", 1)[1])
 
+    # Check for explicit tenant header (useful for testing/API keys)
+    tenant_header = request.headers.get("X-Tenant-ID")
+    if tenant_header:
+        return tenant_header
+
     # Default tenant for development
     return "default"
 

libs/storage/client.py

@@ -19,17 +19,13 @@ class StorageClient:
     async def ensure_bucket(self, bucket_name: str, region: str = "us-east-1") -> bool:
         """Ensure bucket exists, create if not"""
         try:
-            # Check if bucket exists
-            if self.client.bucket_exists(bucket_name):
-                logger.debug("Bucket already exists", bucket=bucket_name)
-                return True
-
-            # Create bucket
-            self.client.make_bucket(bucket_name, location=region)
+            self.client.make_bucket(bucket_name=bucket_name, location=region)
             logger.info("Created bucket", bucket=bucket_name, region=region)
             return True
-
         except S3Error as e:
+            if e.code in ("BucketAlreadyOwnedByYou", "BucketAlreadyExists"):
+                logger.debug("Bucket already exists", bucket=bucket_name)
+                return True
             logger.error("Failed to ensure bucket", bucket=bucket_name, error=str(e))
             return False