completed local setup with compose

apps/svc_ocr/main.py

@@ -7,13 +7,14 @@ import os
 
 # Import shared libraries
 import sys
+from contextlib import asynccontextmanager
 from datetime import datetime
 from typing import Any, cast
 
 import pytesseract
 import structlog
 import ulid
-from fastapi import BackgroundTasks, Depends, HTTPException, Request
+from fastapi import BackgroundTasks, Depends, FastAPI, HTTPException, Request
 from fastapi.responses import JSONResponse
 from pdf2image import convert_from_bytes
 from PIL import Image
@@ -78,6 +79,8 @@ settings: OCRSettings
 async def init_dependencies(app_settings: OCRSettings) -> None:
     """Initialize service dependencies"""
     global storage_client, document_storage, event_bus, settings, vision_processor
+    # Larger delay to ensure NATS is fully ready before attempting connection
+    await asyncio.sleep(10)
 
     settings = app_settings
     logger.info("Starting OCR service")
@@ -89,17 +92,35 @@ async def init_dependencies(app_settings: OCRSettings) -> None:
     minio_client = create_minio_client(settings)
     storage_client = StorageClient(minio_client)
     document_storage = DocumentStorage(storage_client)
-    # Initialize event bus
-    event_bus = create_event_bus(settings)
-    if not event_bus:
-        raise HTTPException(status_code=500, detail="Event bus not initialized")
-
-    eb = event_bus
-    # mypy: event_bus is Optional, so use local alias after check
-    await eb.start()
-
-    # Subscribe to document ingestion events
-    await eb.subscribe(EventTopics.DOC_INGESTED, _handle_document_ingested)
+    # Initialize event bus with retry logic
+    max_retries = 20
+    delay = 5
+    for attempt in range(1, max_retries + 1):
+        logger.info(
+            "Attempting NATS connection", url=settings.nats_servers, attempt=attempt
+        )
+        event_bus = create_event_bus(settings)
+        if not event_bus:
+            raise HTTPException(status_code=500, detail="Event bus not initialized")
+        eb = event_bus
+        try:
+            # Attempt to start and subscribe
+            await eb.start()
+            await eb.subscribe(EventTopics.DOC_INGESTED, _handle_document_ingested)
+            logger.info("NATS connection established on attempt", attempt=attempt)
+            break
+        except Exception as e:
+            logger.error(
+                "Failed to connect to NATS, retrying",
+                attempt=attempt,
+                error=str(e),
+            )
+            if attempt == max_retries:
+                raise HTTPException(
+                    status_code=500, detail="Failed to connect to NATS after retries"
+                )
+            await asyncio.sleep(delay)
+            delay *= 2  # exponential backoff
 
     # Initialize shared OCRProcessor for vision strategy
     try:
@@ -114,7 +135,26 @@ async def init_dependencies(app_settings: OCRSettings) -> None:
     logger.info("OCR service started successfully")
 
 
-# Create app and settings
+async def shutdown_dependencies() -> None:
+    """Shutdown service dependencies"""
+    logger.info("Shutting down OCR service")
+    eb = event_bus
+    if eb is not None:
+        await eb.stop()
+    logger.info("OCR service shutdown complete")
+
+
+@asynccontextmanager
+async def lifespan(app: FastAPI):  # type: ignore
+    """FastAPI lifespan event handler"""
+    # Startup
+    await init_dependencies(cast(OCRSettings, _settings))
+    yield
+    # Shutdown
+    await shutdown_dependencies()
+
+
+# Create app and settings with lifespan
 app, _settings = create_app(
     service_name="svc-ocr",
     title="Tax Agent OCR Service",
@@ -122,8 +162,8 @@ app, _settings = create_app(
     settings_class=OCRSettings,
 )  # fmt: skip
 
-# Initialize dependencies immediately
-asyncio.run(init_dependencies(cast(OCRSettings, _settings)))
+# Override app's lifespan
+app.router.lifespan_context = lifespan
 
 tracer = get_tracer("svc-ocr")
 metrics = get_metrics()

apps/svc_ocr/requirements.txt

@@ -14,3 +14,12 @@ opencv-python-headless>=4.12.0.88  # Headless version is smaller
 
 # Computer vision (torchvision not in base-ml)
 torchvision>=0.23.0
+
+# OpenTelemetry (required by libs/observability)
+opentelemetry-api>=1.21.0
+opentelemetry-sdk>=1.21.0
+opentelemetry-exporter-otlp-proto-grpc>=1.21.0
+opentelemetry-instrumentation-fastapi>=0.42b0
+opentelemetry-instrumentation-httpx>=0.42b0
+opentelemetry-instrumentation-psycopg2>=0.42b0
+opentelemetry-instrumentation-redis>=0.42b0