ratio1 3.4.109__tar.gz → 3.4.110__tar.gz

{ratio1-3.4.109 → ratio1-3.4.110}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ratio1
-Version: 3.4.109
+Version: 3.4.110
 Summary: `ratio1` or Ration1 SDK is the Python SDK required for client app development for the Ratio1 ecosystem
 Project-URL: Homepage, https://github.com/Ratio1/ratio1_sdk
 Project-URL: Bug Tracker, https://github.com/Ratio1/ratio1_sdk/issues

{ratio1-3.4.109 → ratio1-3.4.110}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "ratio1"
-version = "3.4.109"
+version = "3.4.110"
 authors = [
   { name="Andrei Ionut Damian", email="andrei.damian@ratio1.ai" },
   { name="Cristan Bleotiu", email="cristian.bleotiu@ratio1.ai" },

{ratio1-3.4.109 → ratio1-3.4.110}/ratio1/_ver.py

@@ -1,4 +1,4 @@
-__VER__ = "3.4.109"
+__VER__ = "3.4.110"
 
 if __name__ == "__main__":
   with open("pyproject.toml", "rt") as fd:

ratio1-3.4.110/tutorials/ex26_cerviguard_war_loopback.py

@@ -0,0 +1,374 @@
+"""
+ex26_cerviguard_war_loopback.py
+---------------------------------
+
+CerviGuard WAR - Loopback Pipeline Testing
+
+This tutorial demonstrates the complete CerviGuard WAR (Web Application Resource)
+architecture using the loopback mechanism to test image processing pipelines.
+
+Architecture:
+  1. LOCAL_SERVING_API (FastAPI backend, IS_LOOPBACK_PLUGIN=True)
+     - Receives images from UI via REST endpoints
+     - Writes requests to loopback queue
+     - Reads from loopback queue in process() loop
+     - Calls CERVIGUARD_IMAGE_ANALYZER serving plugin directly
+     - Caches results for polling
+
+  2. Loopback DCT (Data Capture Thread)
+     - Reads from shared memory queue
+     - Feeds data back into pipeline
+
+  3. CERVIGUARD_IMAGE_ANALYZER (Serving Plugin - runs in isolated process)
+     - Decodes base64 images
+     - Extracts dimensions and quality metrics
+     - Mockup for future AI model integration
+
+  4. CerviGuard UI (worker_app_runner - to be added by you)
+     - Frontend that interacts with the API
+
+Flow:
+  UI → POST /cerviguard_submit_image → Write to loopback queue
+  → Loopback DCT feeds back → process() reads from loopback
+  → Calls serving plugin via dataapi_inferences()
+  → Caches result → UI polls via GET /cerviguard_get_result
+
+This setup allows you to verify the loopback mechanism works correctly
+before integrating the actual AI model for cervical cancer detection.
+"""
+import json
+import time
+import requests
+import base64
+from ratio1 import Session
+import numpy as np
+from PIL import Image
+import io
+
+def encode_image_to_base64(image_path: str) -> str:
+  """
+  Encode an image file to base64 string
+
+  Parameters
+  ----------
+  image_path : str
+      Path to the image file
+
+  Returns
+  -------
+  str
+      Base64 encoded image string
+  """
+  with open(image_path, 'rb') as f:
+    img_bytes = f.read()
+    b64_string = base64.b64encode(img_bytes).decode('utf-8')
+    return b64_string
+
+
+def test_cerviguard_api(base_url: str, session):
+  """
+  Test the CerviGuard API endpoints
+
+  Parameters
+  ----------
+  base_url : str
+      Base URL of the API (e.g., http://localhost:5002)
+  session : Session
+      The ratio1 session for logging
+  """
+
+  session.P("\n" + "="*70)
+  session.P("Testing CerviGuard WAR Loopback Pipeline")
+  session.P("="*70 + "\n")
+
+  # Test 1: Check system status
+  session.P("1. Checking CerviGuard system status...", color='b')
+  try:
+    response = requests.get(f"{base_url}/cerviguard_status")
+    session.P(f"Status: {response.status_code}", color='g' if response.status_code == 200 else 'r')
+    result = response.json()
+    session.P(f"Service: {result['result']['service']}", color='g')
+    session.P(f"Version: {result['result']['version']}", color='g')
+    session.P(f"Status: {result['result']['status']}", color='g')
+  except Exception as e:
+    session.P(f"Error: {e}", color='r')
+    session.P("Make sure the pipeline is deployed and API is running!", color='r')
+    return
+
+  print()
+
+  # Test 2: Create a test image (1x1 pixel red image)
+  session.P("2. Creating test cervical image (mockup)...", color='b')
+
+  # Create a small test image (100x100 red square)
+  test_img = np.zeros((100, 100, 3), dtype=np.uint8)
+  test_img[:, :, 0] = 255  # Red channel
+
+  # Convert to PIL Image
+  pil_img = Image.fromarray(test_img)
+
+  # Convert to base64
+  buffer = io.BytesIO()
+  pil_img.save(buffer, format='PNG')
+  img_base64 = base64.b64encode(buffer.getvalue()).decode('utf-8')
+
+  session.P(f"Created test image: 100x100 pixels, RGB", color='g')
+  session.P(f"Base64 length: {len(img_base64)} characters", color='g')
+
+  print()
+
+  # Test 3: Submit image for processing
+  session.P("3. Submitting image to CerviGuard API...", color='b')
+
+  try:
+    response = requests.post(
+      f"{base_url}/cerviguard_submit_image",
+      json={
+        "image_data": img_base64,
+        "metadata": {
+          "patient_id": "TEST-001",
+          "capture_date": "2025-01-01",
+          "test_mode": True
+        }
+      }
+    )
+
+    session.P(f"Status: {response.status_code}", color='g' if response.status_code == 200 else 'r')
+    result = response.json()
+
+    if result.get('result', {}).get('status') == 'submitted':
+      request_id = result['result']['request_id']
+      session.P(f"Request ID: {request_id}", color='g')
+      session.P(f"Poll endpoint: {result['result']['poll_endpoint']}", color='g')
+
+      print()
+
+      # Test 4: Poll for results
+      session.P("4. Polling for results (loopback processing)...", color='b')
+
+      max_attempts = 10
+      attempt = 0
+      result_data = None
+
+      while attempt < max_attempts:
+        attempt += 1
+        time.sleep(1)  # Wait 1 second between polls
+
+        session.P(f"  Polling attempt {attempt}/{max_attempts}...", color='y')
+
+        response = requests.get(
+          f"{base_url}/cerviguard_get_result",
+          params={"request_id": request_id}
+        )
+
+        if response.status_code == 200:
+          poll_result = response.json()
+          status = poll_result.get('result', {}).get('status')
+
+          if status == 'completed':
+            result_data = poll_result['result']['result']
+            session.P(f"  ✓ Processing completed!", color='g')
+            break
+          elif status == 'processing':
+            session.P(f"  ⏳ Still processing...", color='y')
+          elif status == 'error':
+            session.P(f"  ✗ Error: {poll_result.get('result', {}).get('error')}", color='r')
+            break
+          else:
+            session.P(f"  Unknown status: {status}", color='y')
+        else:
+          session.P(f"  HTTP error: {response.status_code}", color='r')
+          break
+
+      print()
+
+      # Test 5: Display results
+      if result_data:
+        session.P("5. Processing Results (via Loopback Pipeline):", color='b')
+        session.P("="*70, color='g')
+
+        if result_data.get('status') == 'completed':
+          img_info = result_data.get('image_info', {})
+          session.P(f"  Image Dimensions:", color='g')
+          session.P(f"    Width:  {img_info.get('width')} pixels", color='w')
+          session.P(f"    Height: {img_info.get('height')} pixels", color='w')
+          session.P(f"    Channels: {img_info.get('channels')}", color='w')
+          session.P(f"    Total Pixels: {img_info.get('total_pixels'):,}", color='w')
+          session.P(f"    Size: {img_info.get('size_mb')} MB", color='w')
+          session.P(f"    Data Type: {img_info.get('dtype')}", color='w')
+          session.P(f"    Shape: {img_info.get('shape')}", color='w')
+          session.P("", color='w')
+          session.P(f"  Processing Info:", color='g')
+          session.P(f"    Processor Version: {result_data.get('processor_version')}", color='w')
+          session.P(f"    Processed At: {result_data.get('processed_at')}", color='w')
+
+          session.P("="*70, color='g')
+          session.P("\n✓ Loopback pipeline working correctly!", color='g', boxed=True)
+          session.P("  The image went through the complete flow:", color='g')
+          session.P("    1. Submitted to LOCAL_SERVING_API endpoint", color='w')
+          session.P("    2. Written to loopback queue (IS_LOOPBACK_PLUGIN=True)", color='w')
+          session.P("    3. Loopback DCT feeds data back to pipeline", color='w')
+          session.P("    4. LOCAL_SERVING_API process() reads from loopback", color='w')
+          session.P("    5. Calls CERVIGUARD_IMAGE_ANALYZER serving plugin", color='w')
+          session.P("    6. Result cached in LOCAL_SERVING_API", color='w')
+          session.P("    7. Retrieved via API polling endpoint", color='w')
+        else:
+          session.P(f"  Status: {result_data.get('status')}", color='y')
+          session.P(f"  Error: {result_data.get('error', 'Unknown')}", color='r')
+      else:
+        session.P("5. ✗ No results received after polling", color='r')
+        session.P("   Check edge node logs for issues", color='y')
+
+    else:
+      session.P(f"Failed to submit image: {result}", color='r')
+
+  except Exception as e:
+    session.P(f"Error during testing: {e}", color='r')
+    import traceback
+    session.P(traceback.format_exc(), color='r')
+
+  print()
+  session.P("="*70)
+  session.P("CerviGuard WAR Testing Complete")
+  session.P("="*70 + "\n")
+
+
+if __name__ == "__main__":
+  # Create a session
+  session = Session(silent=False)
+
+  # Get target node from environment or use default
+  # node = os.environ.get("EE_TARGET_NODE", "INSERT_YOUR_NODE_ADDRESS_HERE")
+  node = "0xai_Avvuy6USRwVfbbxEG2HPiCz85mSJle3zo2MbDh5kBD-g"
+
+  session.P(f"Deploying CerviGuard WAR Loopback Pipeline to {node}...", color='b', boxed=True)
+  session.wait_for_node(node)
+
+  # Pipeline configuration with TWO components:
+  # 1. Loopback DCT (reads from queue and feeds back)
+  # 2. LOCAL_SERVING_API (writes to queue, reads from queue, calls serving plugin)
+  # Note: CERVIGUARD_IMAGE_ANALYZER serving plugin is called via AI_ENGINE config
+
+  pipeline_config = {
+    "NAME": "cerviguard_demo",
+    "TYPE": "Loopback",  # CRITICAL: Use Loopback DCT
+    "LOOPBACK_QUEUE_SIZE": 32,
+
+    "PLUGINS": [
+      # Backend API (receives images, writes to loopback, reads from loopback, calls serving)
+      {
+        "SIGNATURE": "LOCAL_SERVING_API",
+        "INSTANCES": [
+          {
+            "PORT": 5002,
+            "INSTANCE_ID": "cerviguard_api",
+            "AI_ENGINE": "CERVIGUARD_IMAGE_ANALYZER",  # Serving plugin to use
+            "LOG_REQUESTS": True,
+            "RESPONSE_FORMAT": "WRAPPED",
+            "RESULT_CACHE_TTL": 300,  # Keep results for 5 minutes
+          }
+        ]
+      }
+
+      # TODO: Add your CerviGuard UI plugin here
+      # {
+      #   "SIGNATURE": "YOUR_CERVIGUARD_UI_WORKER_APP",
+      #   "INSTANCES": [{"INSTANCE_ID": "ui_01"}]
+      # }
+    ]
+  }
+
+  session.P("\nPipeline Configuration:", color='b')
+  session.P(json.dumps(pipeline_config, indent=2), color='w')
+
+  # Deploy the pipeline
+  session.P("\nDeploying pipeline...", color='b')
+
+  # Create the pipeline with Loopback data source
+  pipeline = session.create_pipeline(
+    node=node,
+    name=pipeline_config["NAME"],
+    data_source=pipeline_config["TYPE"],  # "Loopback"
+    config={
+      "LOOPBACK_QUEUE_SIZE": pipeline_config["LOOPBACK_QUEUE_SIZE"]
+    }
+  )
+
+  # Add the LOCAL_SERVING_API plugin
+  # This plugin handles everything: API endpoints, loopback, and serving plugin calls
+  session.P("Adding LOCAL_SERVING_API plugin...", color='b')
+  api_instance = pipeline.create_plugin_instance(
+    signature='LOCAL_SERVING_API',
+    instance_id='cerviguard_api',
+    log_requests=True,
+    response_format='WRAPPED',
+    port=5082,
+    ai_engine='CERVIGUARD_IMAGE_ANALYZER',  # This serving plugin runs in isolated process
+    result_cache_ttl=300,
+  )
+
+  # Deploy the pipeline
+  session.P("Deploying pipeline to node...", color='b')
+  pipeline.deploy()
+
+  # Wait for deployment
+  session.P("Waiting for pipeline to be fully deployed...", color='b')
+  time.sleep(8)
+
+  # The API should be accessible on localhost
+  # Port is configured in plugin config as 5082
+  base_url = "http://localhost:5082"
+
+  session.P(f"\n" + "="*70, color='g')
+  session.P(f"CerviGuard WAR Backend Deployed!", color='g', boxed=True)
+  session.P(f"="*70, color='g')
+  session.P(f"API URL: {base_url}", color='g')
+  session.P(f"Pipeline: {pipeline_config['NAME']}", color='g')
+  session.P(f"Loopback: ENABLED", color='g')
+  session.P(f"="*70 + "\n", color='g')
+
+  # Give the server a bit more time to start
+  session.P("Waiting for API server to be ready...", color='b')
+  time.sleep(30)
+
+  # Run tests
+  test_cerviguard_api(base_url, session)
+
+  # Show API endpoints
+  session.P("\n" + "="*70)
+  session.P("Available CerviGuard API Endpoints:")
+  session.P("="*70)
+  session.P(f"  POST {base_url}/cerviguard_submit_image", color='g')
+  session.P(f"    - Submit cervical image for analysis", color='w')
+  session.P(f"    - Body: {{'image_data': '<base64>', 'metadata': {{...}}}}", color='w')
+  session.P("")
+  session.P(f"  GET  {base_url}/cerviguard_get_result?request_id=<id>", color='g')
+  session.P(f"    - Poll for processing results", color='w')
+  session.P("")
+  session.P(f"  GET  {base_url}/cerviguard_status", color='g')
+  session.P(f"    - Check system status", color='w')
+  session.P("="*70 + "\n")
+
+  # Integration instructions
+  session.P("Next Steps for CerviGuard UI Integration:", color='y', boxed=True)
+  session.P("1. Add your CerviGuard UI worker_app to the pipeline", color='w')
+  session.P("2. From UI JavaScript, call the API endpoints:", color='w')
+  session.P("   - Submit image: POST /cerviguard_submit_image", color='w')
+  session.P("   - Poll results: GET /cerviguard_get_result?request_id=<id>", color='w')
+  session.P("3. Display the image analysis results in your UI", color='w')
+  session.P("4. The serving plugin runs in isolated process for fault isolation", color='w')
+  session.P("5. Later: Enhance CERVIGUARD_IMAGE_ANALYZER with actual AI model", color='w')
+  session.P("")
+
+  # Keep running
+  session.P("Press Ctrl+C to stop and cleanup...", color='y')
+
+  try:
+    session.run(
+      wait=True,
+      close_pipelines=True
+    )
+  except KeyboardInterrupt:
+    session.P("\nStopping and cleaning up...", color='y')
+
+  session.P("\nCerviGuard WAR tutorial completed!", color='g', boxed=True)

{ratio1-3.4.109 → ratio1-3.4.110}/xperimental/_checks/dauth_check.py

@@ -23,8 +23,9 @@ if __name__ == '__main__' :
   
   eng.reset_network("mainnet")
   
-  for _ in range(50):
+  for i in range(50):
     sleep(1)
+    l.P(f"Running dauth autocomplete check #{i+1}...", show=True)
     d = eng.dauth_autocomplete(
       # dauth_endp='N/Adhstrgredshtfnfnhgm',
       add_env=False,