Upload files to "/"

block.py

@@ -1,4 +1,6 @@
 import logging
+import json
+import os
 from pre_processing import process_record
 from processing import processing
 from post_processing import post_processing, post_processing_duplicate
@@ -10,12 +12,23 @@ logging.basicConfig(
 )
 logger = logging.getLogger(__name__)
 
-# API URLs
+def load_schema(path: str):
-# base_url = "http://localhost:8080/api/v1/clusters"
+    """Helper to read the JSON config file."""
-base_url = "http://centurion-mlg.default.svc.cluster.local:8080/api/v1/clusters"
+    with open(path, "r") as f:
-cluster_name = "cluster_deviceid_email_fuzzydevice_direct_new"
+        return json.load(f)
-url_post = f"{base_url}/{cluster_name}/records"
+
-url_get = f"{base_url}/{cluster_name}/record"
+def get_cluster_name(namespace: str) -> str:
+    """
+    Fetches the clusterName for the given namespace from config.json.
+    Expects config.json to contain a list of { "namespace": "...", "clusterName": "..." }.
+    """
+    response_schema = load_schema("/app/config.json")  # Update path if needed
+    for item in response_schema:
+        if item.get("namespace") == namespace:
+            logger.info("Got the clusterName for namespace '%s'", namespace)
+            return item.get("clusterName")
+    logger.error("Provided Namespace '%s' not found in config.json.", namespace)
+    raise ValueError(f"Namespace '{namespace}' not found")
 
 def __main__(
         application_key: str,
@@ -26,6 +39,12 @@ def __main__(
         hd_score_m1: float
     ) -> dict:
 
+    namespace = os.getenv("NAMESPACE", "staging")
+    base_url = "http://centurion-mlg.default.svc.cluster.local:8080/api/v1/clusters"
+    cluster_name = get_cluster_name(namespace)
+    url_post = f"{base_url}/{cluster_name}/records"
+    url_get  = f"{base_url}/{cluster_name}/record"
+
     data = {
         "application_key": application_key,
         "application_timestamp": application_timestamp,

config.json

@@ -0,0 +1,10 @@
+[
+  {
+    "namespace": "staging",
+    "clusterName": "cluster_deviceid_email_fuzzydevice_direct_new"
+  },
+  {
+    "namespace": "production",
+    "clusterName": "cluster_deviceid_email_fuzzydevice_direct_new_prod"
+  }
+]

pre_processing.py

@@ -20,6 +20,21 @@ def process_record(url, record):
             output = response.json()
             logger.info(f"Pre Processed record: {record['application_key']} - Response: {output}")
             return output
+        elif response.status_code == 404:
+            # Special-case handling for "CLUSTER_NOT_FOUND" (or any 404).
+            # Return a benign structure that includes 'hd_key' and 'matches'
+            # so post_processing does not fail when it does:
+            #   record_id = data.get('hd_key')
+            #   connected_records = data.get('matches', []) + [record_id]
+            logger.warning(
+                f"Ignoring 404 for record {record['application_key']}. "
+                f"Status: 404, Response: {response.text}"
+            )
+            return {
+                "application_key": record.get("application_key"),
+                "hd_key": None,
+                "matches": []
+            }
         else:
             logger.error(f"Failed to process record {record['application_key']}. Status: {response.status_code}, Response: {response.text}")
             return {"error": response.text}