From 093793c41313bfe3b6a35085dca9ee65f2361ae2 Mon Sep 17 00:00:00 2001
From: "Shreyas Waikar (Persistent Systems Inc)" <v-swaikar@microsoft.com>
Date: Mon, 1 Jun 2026 16:28:48 +0530
Subject: [PATCH 01/13] chore(deps): bump authlib from 1.6.11 to 1.6.12 in
 ContentProcessorWorkflow

Applies the changes from Dependabot PR microsoft/content-processing-solution-accelerator#589 onto dev so they reach the dev branch ahead of the upstream PR (which targets main).

Refs: ADO #44960

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 src/ContentProcessorWorkflow/pyproject.toml | 2 +-
 src/ContentProcessorWorkflow/uv.lock        | 8 ++++----
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/src/ContentProcessorWorkflow/pyproject.toml b/src/ContentProcessorWorkflow/pyproject.toml
index 804ed5f4..51cea3bb 100644
--- a/src/ContentProcessorWorkflow/pyproject.toml
+++ b/src/ContentProcessorWorkflow/pyproject.toml
@@ -29,7 +29,7 @@ dependencies = [
     "sas-cosmosdb==0.1.4",
     "sas-storage==1.0.0",
     "tenacity==9.1.2",
-    "authlib==1.6.11",
+    "authlib==1.6.12",
     "protobuf==6.33.6",
     "cryptography==46.0.7",
     "pyjwt==2.12.1",
diff --git a/src/ContentProcessorWorkflow/uv.lock b/src/ContentProcessorWorkflow/uv.lock
index 9fd628a0..ce423c6a 100644
--- a/src/ContentProcessorWorkflow/uv.lock
+++ b/src/ContentProcessorWorkflow/uv.lock
@@ -554,14 +554,14 @@ wheels = [
 
 [[package]]
 name = "authlib"
-version = "1.6.11"
+version = "1.6.12"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "cryptography" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/28/10/b325d58ffe86815b399334a101e63bc6fa4e1953921cb23703b48a0a0220/authlib-1.6.11.tar.gz", hash = "sha256:64db35b9b01aeccb4715a6c9a6613a06f2bd7be2ab9d2eb89edd1dfc7580a38f", size = 165359, upload-time = "2026-04-16T07:22:50.279Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/d3/30/6691fdc63b35f54a5a65e04fa1e59d827f4d4e8f4a39678ba7d3088ce0c8/authlib-1.6.12.tar.gz", hash = "sha256:0656d8482f28fc8221929d5f35b2bde5d13e10555ebc06b4561b0d622e83b1bd", size = 165368, upload-time = "2026-05-04T08:11:31.826Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/57/2f/55fca558f925a51db046e5b929deb317ddb05afed74b22d89f4eca578980/authlib-1.6.11-py2.py3-none-any.whl", hash = "sha256:c8687a9a26451c51a34a06fa17bb97cb15bba46a6a626755e2d7f50da8bff3e3", size = 244469, upload-time = "2026-04-16T07:22:48.413Z" },
+    { url = "https://files.pythonhosted.org/packages/cd/51/9b0b5cd4cf683a02db937a6f9bbebcdc9c56558a7bb3763ce7d3512103c3/authlib-1.6.12-py2.py3-none-any.whl", hash = "sha256:e9229ad7fde610b139dd12f5edbe97eab9ee78bfb85691247e767727850b99ab", size = 244473, upload-time = "2026-05-04T08:11:30.354Z" },
 ]
 
 [[package]]
@@ -2994,7 +2994,7 @@ requires-dist = [
     { name = "agent-framework", specifier = "==1.0.0b260107" },
     { name = "aiohttp", specifier = "==3.13.5" },
     { name = "art", specifier = "==6.5" },
-    { name = "authlib", specifier = "==1.6.11" },
+    { name = "authlib", specifier = "==1.6.12" },
     { name = "azure-ai-agents", specifier = "==1.2.0b5" },
     { name = "azure-ai-inference", specifier = "==1.0.0b9" },
     { name = "azure-ai-projects", specifier = "==2.0.0b3" },

From 68f54be535f9771b5bd7bcad3e3e6acc843af221 Mon Sep 17 00:00:00 2001
From: "Shreyas Waikar (Persistent Systems Inc)" <v-swaikar@microsoft.com>
Date: Mon, 1 Jun 2026 16:28:49 +0530
Subject: [PATCH 02/13] build(deps): bump idna from 3.11 to 3.15 in
 ContentProcessorAPI

Applies the changes from Dependabot PR microsoft/content-processing-solution-accelerator#595 onto dev so they reach the dev branch ahead of the upstream PR (which targets main).

Refs: ADO #44960

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 src/ContentProcessorAPI/requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/ContentProcessorAPI/requirements.txt b/src/ContentProcessorAPI/requirements.txt
index aa8b8f50..15754e7f 100644
--- a/src/ContentProcessorAPI/requirements.txt
+++ b/src/ContentProcessorAPI/requirements.txt
@@ -22,7 +22,7 @@ h11==0.16.0
 httpcore==1.0.9
 httptools==0.7.1
 httpx==0.28.1
-idna==3.11
+idna==3.15
 isodate==0.7.2
 jinja2==3.1.6
 jsonschema==4.25.1

From 03bcf446228729e7aef7a18fbe87359b29f06dd5 Mon Sep 17 00:00:00 2001
From: "Shreyas Waikar (Persistent Systems Inc)" <v-swaikar@microsoft.com>
Date: Mon, 1 Jun 2026 16:28:50 +0530
Subject: [PATCH 03/13] build(deps): bump idna from 3.11 to 3.15 in
 ContentProcessor

Applies the changes from Dependabot PR microsoft/content-processing-solution-accelerator#596 onto dev so they reach the dev branch ahead of the upstream PR (which targets main).

Refs: ADO #44960

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 src/ContentProcessor/requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/ContentProcessor/requirements.txt b/src/ContentProcessor/requirements.txt
index 123b1cee..aeff7976 100644
--- a/src/ContentProcessor/requirements.txt
+++ b/src/ContentProcessor/requirements.txt
@@ -15,7 +15,7 @@ colorama==0.4.6
 coverage==7.13.5
 cryptography==46.0.7
 dnspython==2.8.0
-idna==3.11
+idna==3.15
 iniconfig==2.3.0
 isodate==0.7.2
 mongomock==4.3.0

From 77de40526fd51b20dd60039127550cc63fe24830 Mon Sep 17 00:00:00 2001
From: "Shreyas Waikar (Persistent Systems Inc)" <v-swaikar@microsoft.com>
Date: Mon, 1 Jun 2026 16:28:51 +0530
Subject: [PATCH 04/13] chore(deps): bump idna from 3.11 to 3.15 in
 ContentProcessorWorkflow uv.lock

Applies the changes from Dependabot PR microsoft/content-processing-solution-accelerator#597 onto dev so they reach the dev branch ahead of the upstream PR (which targets main).

Refs: ADO #44960

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 src/ContentProcessorWorkflow/uv.lock | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/ContentProcessorWorkflow/uv.lock b/src/ContentProcessorWorkflow/uv.lock
index ce423c6a..04ec2a5e 100644
--- a/src/ContentProcessorWorkflow/uv.lock
+++ b/src/ContentProcessorWorkflow/uv.lock
@@ -1747,11 +1747,11 @@ wheels = [
 
 [[package]]
 name = "idna"
-version = "3.11"
+version = "3.15"
 source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/6f/6d/0703ccc57f3a7233505399edb88de3cbd678da106337b9fcde432b65ed60/idna-3.11.tar.gz", hash = "sha256:795dafcc9c04ed0c1fb032c2aa73654d8e8c5023a7df64a53f39190ada629902", size = 194582, upload-time = "2025-10-12T14:55:20.501Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/82/77/7b3966d0b9d1d31a36ddf1746926a11dface89a83409bf1483f0237aa758/idna-3.15.tar.gz", hash = "sha256:ca962446ea538f7092a95e057da437618e886f4d349216d2b1e294abfdb65fdc", size = 199245, upload-time = "2026-05-12T22:45:57.011Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/0e/61/66938bbb5fc52dbdf84594873d5b51fb1f7c7794e9c0f5bd885f30bc507b/idna-3.11-py3-none-any.whl", hash = "sha256:771a87f49d9defaf64091e6e6fe9c18d4833f140bd19464795bc32d966ca37ea", size = 71008, upload-time = "2025-10-12T14:55:18.883Z" },
+    { url = "https://files.pythonhosted.org/packages/d2/23/408243171aa9aaba178d3e2559159c24c1171a641aa83b67bdd3394ead8e/idna-3.15-py3-none-any.whl", hash = "sha256:048adeaf8c2d788c40fee287673ccaa74c24ffd8dcf09ffa555a2fbb59f10ac8", size = 72340, upload-time = "2026-05-12T22:45:55.733Z" },
 ]
 
 [[package]]

From ec8d7a03b6b6e279e6bddd7f96755a44cd1d3bab Mon Sep 17 00:00:00 2001
From: chaudhariniraj <chaudhariniraj19@gmail.com>
Date: Fri, 12 Jun 2026 16:24:39 +0530
Subject: [PATCH 05/13] Resolve CodeQL issues

---
 .../src/libs/utils/azure_credential_utils.py              | 6 +++++-
 src/ContentProcessor/src/libs/utils/credential_util.py    | 6 +++++-
 src/ContentProcessorAPI/app/libs/base/application_base.py | 4 ++--
 .../src/libs/azure/app_configuration.py                   | 8 +++++++-
 .../src/libs/base/application_base.py                     | 3 ++-
 src/ContentProcessorWorkflow/src/utils/credential_util.py | 7 ++++++-
 6 files changed, 27 insertions(+), 7 deletions(-)

diff --git a/src/ContentProcessor/src/libs/utils/azure_credential_utils.py b/src/ContentProcessor/src/libs/utils/azure_credential_utils.py
index 07a4f2b0..5d711e85 100644
--- a/src/ContentProcessor/src/libs/utils/azure_credential_utils.py
+++ b/src/ContentProcessor/src/libs/utils/azure_credential_utils.py
@@ -130,7 +130,11 @@ def get_azure_credential():
     logging.info(
         "[AUTH] All CLI credentials failed - falling back to DefaultAzureCredential"
     )
-    return DefaultAzureCredential()
+    raise RuntimeError(
+        "No Azure authentication available. "
+        "Use Managed Identity in Azure or run "
+        "'az login' / 'azd auth login' locally."
+    )
 
 
 def get_async_azure_credential():
diff --git a/src/ContentProcessor/src/libs/utils/credential_util.py b/src/ContentProcessor/src/libs/utils/credential_util.py
index 52fbdeef..1efcaab7 100644
--- a/src/ContentProcessor/src/libs/utils/credential_util.py
+++ b/src/ContentProcessor/src/libs/utils/credential_util.py
@@ -130,7 +130,11 @@ def get_azure_credential():
     logging.info(
         "[AUTH] All CLI credentials failed - falling back to DefaultAzureCredential"
     )
-    return DefaultAzureCredential()
+    raise RuntimeError(
+        "No Azure authentication available. "
+        "Use Managed Identity in Azure or run "
+        "'az login' / 'azd auth login' locally."
+    )
 
 
 def get_async_azure_credential():
diff --git a/src/ContentProcessorAPI/app/libs/base/application_base.py b/src/ContentProcessorAPI/app/libs/base/application_base.py
index 7ea33d8e..e3fb6e1c 100644
--- a/src/ContentProcessorAPI/app/libs/base/application_base.py
+++ b/src/ContentProcessorAPI/app/libs/base/application_base.py
@@ -15,7 +15,7 @@
 import os
 from abc import ABC, abstractmethod
 
-from azure.identity import DefaultAzureCredential
+from app.utils.azure_credential_utils import get_azure_credential
 from dotenv import load_dotenv
 
 from app.libs.application.application_configuration import (
@@ -72,7 +72,7 @@ def __init__(self, env_file_path: str | None = None, **data):
         self._load_env(env_file_path=env_file_path)
 
         self.application_context = AppContext()
-        self.application_context.set_credential(DefaultAzureCredential())
+        self.application_context.set_credential(get_azure_credential())
 
         app_config_endpoint: str | None = EnvConfiguration().app_config_endpoint
         if app_config_endpoint != "" and app_config_endpoint is not None:
diff --git a/src/ContentProcessorWorkflow/src/libs/azure/app_configuration.py b/src/ContentProcessorWorkflow/src/libs/azure/app_configuration.py
index ee2501cd..f333133e 100644
--- a/src/ContentProcessorWorkflow/src/libs/azure/app_configuration.py
+++ b/src/ContentProcessorWorkflow/src/libs/azure/app_configuration.py
@@ -91,7 +91,13 @@ def __init__(
             ValueError: If *app_configuration_url* is ``None`` or the
                 credential is missing after defaulting.
         """
-        self.credential = credential or DefaultAzureCredential()
+        if credential is None:
+            raise ValueError(
+                "Azure credential is required. "
+                "Use Managed Identity, AzureCliCredential, or AzureDeveloperCliCredential."
+            )
+
+        self.credential = credential
         self.app_config_endpoint = app_configuration_url
         self._initialize_client()
 
diff --git a/src/ContentProcessorWorkflow/src/libs/base/application_base.py b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
index fbcbaa23..d3a63579 100644
--- a/src/ContentProcessorWorkflow/src/libs/base/application_base.py
+++ b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
@@ -36,6 +36,7 @@ def run(self):
 from abc import ABC, abstractmethod
 
 from azure.identity import DefaultAzureCredential
+from src.utils.credential_util import get_azure_credential
 from dotenv import load_dotenv
 
 from libs.agent_framework.agent_framework_settings import AgentFrameworkSettings
@@ -117,7 +118,7 @@ def __init__(self, env_file_path: str | None = None, **data):
         self._load_env(env_file_path=env_file_path)
 
         self.application_context = AppContext()
-        self.application_context.set_credential(DefaultAzureCredential())
+        self.application_context.set_credential(get_azure_credential())
 
         app_config_url: str | None = _envConfiguration().app_config_endpoint
         if app_config_url != "" and app_config_url is not None:
diff --git a/src/ContentProcessorWorkflow/src/utils/credential_util.py b/src/ContentProcessorWorkflow/src/utils/credential_util.py
index b37de6d9..fbef0657 100644
--- a/src/ContentProcessorWorkflow/src/utils/credential_util.py
+++ b/src/ContentProcessorWorkflow/src/utils/credential_util.py
@@ -126,7 +126,12 @@ def get_azure_credential():
     logging.info(
         "[AUTH] All CLI credentials failed - falling back to DefaultAzureCredential"
     )
-    return DefaultAzureCredential()
+    
+    raise RuntimeError(
+    "No Azure authentication available. "
+    "Use Managed Identity in Azure or run "
+    "'az login' / 'azd auth login' locally."
+    )
 
 
 def get_async_azure_credential():

From 8fc57b86eb95c492abeed51054bf553b23e9521f Mon Sep 17 00:00:00 2001
From: chaudhariniraj <chaudhariniraj19@gmail.com>
Date: Fri, 12 Jun 2026 17:55:45 +0530
Subject: [PATCH 06/13] Resolve test cases error

---
 .../utils/test_azure_credential_utils.py         | 16 +++++++++-------
 .../test_azure_credential_utils_extended.py      | 14 +++++---------
 2 files changed, 14 insertions(+), 16 deletions(-)

diff --git a/src/tests/ContentProcessor/utils/test_azure_credential_utils.py b/src/tests/ContentProcessor/utils/test_azure_credential_utils.py
index 216b302e..7f0f88a0 100644
--- a/src/tests/ContentProcessor/utils/test_azure_credential_utils.py
+++ b/src/tests/ContentProcessor/utils/test_azure_credential_utils.py
@@ -7,6 +7,8 @@
 
 from unittest.mock import MagicMock, patch
 
+import pytest
+
 import libs.utils.azure_credential_utils as azure_credential_utils
 
 MODULE = "libs.utils.azure_credential_utils"
@@ -45,16 +47,16 @@ def test_returns_user_assigned_with_client_id(self, mock_managed):
         mock_managed.assert_called_once_with(client_id="test-client-id")
         assert credential == mock_instance
 
-    @patch(f"{MODULE}.DefaultAzureCredential")
     @patch(f"{MODULE}.AzureDeveloperCliCredential", side_effect=Exception("no azd"))
     @patch(f"{MODULE}.AzureCliCredential", side_effect=Exception("no az"))
     @patch.dict("os.environ", {}, clear=True)
-    def test_falls_back_to_default(self, mock_cli, mock_dev_cli, mock_default):
-        mock_instance = MagicMock()
-        mock_default.return_value = mock_instance
-        credential = azure_credential_utils.get_azure_credential()
-        mock_default.assert_called_once()
-        assert credential == mock_instance
+    def test_raises_when_no_credentials_available(
+        self, mock_cli, mock_dev_cli
+    ):
+        with pytest.raises(RuntimeError) as exc:
+            azure_credential_utils.get_azure_credential()
+
+        assert "No Azure authentication available" in str(exc.value)
 
 
 # ── TestGetAsyncAzureCredential ─────────────────────────────────────────
diff --git a/src/tests/ContentProcessor/utils/test_azure_credential_utils_extended.py b/src/tests/ContentProcessor/utils/test_azure_credential_utils_extended.py
index 11858fdc..edd735d8 100644
--- a/src/tests/ContentProcessor/utils/test_azure_credential_utils_extended.py
+++ b/src/tests/ContentProcessor/utils/test_azure_credential_utils_extended.py
@@ -42,26 +42,22 @@ def test_get_azure_credential_with_website_site_name(self, monkeypatch):
             assert credential == mock_instance
 
     def test_get_azure_credential_cli_failure_fallback(self, monkeypatch):
-        """Test fallback to DefaultAzureCredential when CLI credentials fail"""
+        """Test RuntimeError when all credential options fail"""
         # Clear all Azure environment indicators
         for key in ["WEBSITE_SITE_NAME", "AZURE_CLIENT_ID", "MSI_ENDPOINT",
                     "IDENTITY_ENDPOINT", "KUBERNETES_SERVICE_HOST", "CONTAINER_REGISTRY_LOGIN"]:
             monkeypatch.delenv(key, raising=False)
 
         with patch('libs.utils.azure_credential_utils.AzureCliCredential') as mock_cli_cred, \
-             patch('libs.utils.azure_credential_utils.AzureDeveloperCliCredential') as mock_azd_cred, \
-             patch('libs.utils.azure_credential_utils.DefaultAzureCredential') as mock_default:
+             patch('libs.utils.azure_credential_utils.AzureDeveloperCliCredential') as mock_azd_cred:
 
-            # Make both CLI credentials raise exceptions
             mock_cli_cred.side_effect = Exception("CLI credential failed")
             mock_azd_cred.side_effect = Exception("AZD credential failed")
-            mock_default_instance = Mock()
-            mock_default.return_value = mock_default_instance
 
-            credential = get_azure_credential()
+            with pytest.raises(RuntimeError) as exc:
+                get_azure_credential()
 
-            assert credential == mock_default_instance
-            mock_default.assert_called_once()
+            assert "No Azure authentication available" in str(exc.value)
 
     def test_get_azure_credential_azd_success(self, monkeypatch):
         """Test successful Azure Developer CLI credential"""

From e11892e8f6aaef61d6ba5faa17d7962f73e481f5 Mon Sep 17 00:00:00 2001
From: chaudhariniraj <chaudhariniraj19@gmail.com>
Date: Fri, 12 Jun 2026 18:28:45 +0530
Subject: [PATCH 07/13] Resolve test cases error 1

---
 src/ContentProcessorWorkflow/src/libs/base/application_base.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/src/ContentProcessorWorkflow/src/libs/base/application_base.py b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
index d3a63579..0379a702 100644
--- a/src/ContentProcessorWorkflow/src/libs/base/application_base.py
+++ b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
@@ -36,6 +36,7 @@ def run(self):
 from abc import ABC, abstractmethod
 
 from azure.identity import DefaultAzureCredential
+from utils.credential_util import get_azure_credential
 from src.utils.credential_util import get_azure_credential
 from dotenv import load_dotenv
 

From cab11be1fe2d05e5c4db9b0c2f294486313bd6d6 Mon Sep 17 00:00:00 2001
From: chaudhariniraj <chaudhariniraj19@gmail.com>
Date: Fri, 12 Jun 2026 18:44:21 +0530
Subject: [PATCH 08/13] Resolve test cases error 2

---
 src/ContentProcessorWorkflow/src/libs/base/application_base.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/src/ContentProcessorWorkflow/src/libs/base/application_base.py b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
index 0379a702..cb657398 100644
--- a/src/ContentProcessorWorkflow/src/libs/base/application_base.py
+++ b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
@@ -37,7 +37,6 @@ def run(self):
 
 from azure.identity import DefaultAzureCredential
 from utils.credential_util import get_azure_credential
-from src.utils.credential_util import get_azure_credential
 from dotenv import load_dotenv
 
 from libs.agent_framework.agent_framework_settings import AgentFrameworkSettings

From cf8dfc7284ee02bd97d4a5ef7c15085b4a30c627 Mon Sep 17 00:00:00 2001
From: chaudhariniraj <chaudhariniraj19@gmail.com>
Date: Fri, 12 Jun 2026 18:54:00 +0530
Subject: [PATCH 09/13] Resolve test cases error 3

---
 .../src/libs/base/application_base.py         |  1 -
 .../utils/test_credential_util_extended.py    | 24 +++++++++++--------
 2 files changed, 14 insertions(+), 11 deletions(-)

diff --git a/src/ContentProcessorWorkflow/src/libs/base/application_base.py b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
index cb657398..61b6a603 100644
--- a/src/ContentProcessorWorkflow/src/libs/base/application_base.py
+++ b/src/ContentProcessorWorkflow/src/libs/base/application_base.py
@@ -35,7 +35,6 @@ def run(self):
 import os
 from abc import ABC, abstractmethod
 
-from azure.identity import DefaultAzureCredential
 from utils.credential_util import get_azure_credential
 from dotenv import load_dotenv
 
diff --git a/src/tests/ContentProcessorWorkflow/utils/test_credential_util_extended.py b/src/tests/ContentProcessorWorkflow/utils/test_credential_util_extended.py
index d4fda81d..40cfaf68 100644
--- a/src/tests/ContentProcessorWorkflow/utils/test_credential_util_extended.py
+++ b/src/tests/ContentProcessorWorkflow/utils/test_credential_util_extended.py
@@ -1,5 +1,6 @@
 """Extended tests for credential_util.py to improve coverage"""
 from unittest.mock import Mock, patch
+import pytest
 from utils.credential_util import (
     get_azure_credential,
     get_async_azure_credential,
@@ -40,24 +41,27 @@ def test_get_azure_credential_app_service_environment(self, monkeypatch):
             assert credential == mock_instance
 
     def test_get_azure_credential_all_cli_fail(self, monkeypatch):
-        """Test fallback when all CLI credentials fail"""
-        for key in ["WEBSITE_SITE_NAME", "AZURE_CLIENT_ID", "MSI_ENDPOINT",
-                    "IDENTITY_ENDPOINT", "KUBERNETES_SERVICE_HOST", "CONTAINER_REGISTRY_LOGIN"]:
+        """Test RuntimeError when all credential options fail"""
+        for key in [
+            "WEBSITE_SITE_NAME",
+            "AZURE_CLIENT_ID",
+            "MSI_ENDPOINT",
+            "IDENTITY_ENDPOINT",
+            "KUBERNETES_SERVICE_HOST",
+            "CONTAINER_REGISTRY_LOGIN",
+        ]:
             monkeypatch.delenv(key, raising=False)
 
         with patch('utils.credential_util.AzureCliCredential') as mock_cli, \
-             patch('utils.credential_util.AzureDeveloperCliCredential') as mock_azd, \
-             patch('utils.credential_util.DefaultAzureCredential') as mock_default:
+             patch('utils.credential_util.AzureDeveloperCliCredential') as mock_azd:
 
             mock_cli.side_effect = Exception("AzureCLI not available")
             mock_azd.side_effect = Exception("AzureDeveloperCLI not available")
-            mock_default_instance = Mock()
-            mock_default.return_value = mock_default_instance
 
-            credential = get_azure_credential()
+            with pytest.raises(RuntimeError) as exc:
+                get_azure_credential()
 
-            assert credential == mock_default_instance
-            mock_default.assert_called_once()
+            assert "No Azure authentication available" in str(exc.value)
 
     def test_get_azure_credential_cli_success(self, monkeypatch):
         """Test successful Azure CLI credential"""

From dd66aa7bfb3b5b84433cf7e6cf8027d9deca85aa Mon Sep 17 00:00:00 2001
From: Prachig-Microsoft <v-prachig@microsoft.com>
Date: Mon, 15 Jun 2026 13:15:18 +0530
Subject: [PATCH 10/13] fix(scoring): show N/A instead of 0% for unavailable
 entity/schema scores

Root cause: When the evaluate step couldn't compute any per-field confidence (e.g.
logprobs unavailable on reasoning models like gpt-5/o1/o3, or image-only flow with
no Content Understanding signal), save_handler emitted entity_score=0.0,
schema_score=0.0. These `0.0`s flowed through Cosmos -> API -> UI and rendered as
`0%` (red), indistinguishable from a genuine zero confidence.

Fix: Treat `total_evaluated_fields_count == 0` (or no comparison items) as
*unavailable* and propagate `None` through the ContentProcessor, ContentProcessorAPI
and ContentProcessorWorkflow models. The frontend percentage cell renderer now shows
`N/A` for null/undefined and `0%` only for a genuine numeric zero.

Files changed:
- ContentProcessor: save_handler.py (extracted _derive_aggregate_scores helper)
- ContentProcessor: content_process.py default scores -> None
- ContentProcessorAPI: ContentProcess + Content_Process default scores -> None
- ContentProcessorWorkflow: ContentProcessRecord + Content_Process default scores -> None
- ContentProcessorWorkflow: document_process_executor preserves None instead of
  coercing to 0.0
- ContentProcessorWeb: ProcessQueueGridTypes types scores nullable;
  ProcessQueueGrid passes undefined for null/undefined; CustomCellRender renders
  `N/A` when valueText is null/undefined and only `...` while still processing

Tests:
- New: ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py (5 cases:
  valid scores, missing per-field signal, no comparison items, genuine zero,
  all-fields-above-threshold)
- Updated existing default-value tests in Workflow + src/tests to assert None
- Added tests for explicit zero preservation and Failed status -> None
---
 .../src/libs/models/content_process.py        |   6 +-
 .../libs/pipeline/handlers/save_handler.py    |  67 ++++++---
 .../unit/pipeline/test_save_handler_scores.py | 137 ++++++++++++++++++
 .../models/contentprocessor/claim_process.py  |  12 +-
 .../contentprocessor/content_process.py       |   6 +-
 .../ProcessQueueGrid/CustomCellRender.tsx     |  39 ++++-
 .../ProcessQueueGrid/ProcessQueueGrid.tsx     |  12 +-
 .../ProcessQueueGrid/ProcessQueueGridTypes.ts |  21 ++-
 .../src/repositories/model/claim_process.py   |  12 +-
 .../src/services/content_process_models.py    |   4 +-
 .../executor/document_process_executor.py     |  40 +++--
 .../repositories/test_claim_process_model.py  |  28 +++-
 .../services/test_content_process_models.py   |  10 ++
 .../repositories/test_claim_process_model.py  |   6 +-
 .../services/test_content_process_models.py   |   6 +-
 15 files changed, 336 insertions(+), 70 deletions(-)
 create mode 100644 src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py

diff --git a/src/ContentProcessor/src/libs/models/content_process.py b/src/ContentProcessor/src/libs/models/content_process.py
index c9aacf1f..018f13b2 100644
--- a/src/ContentProcessor/src/libs/models/content_process.py
+++ b/src/ContentProcessor/src/libs/models/content_process.py
@@ -67,9 +67,9 @@ class ContentProcess(BaseModel):
     last_modified_time: datetime.datetime = datetime.datetime.now(datetime.UTC)
     last_modified_by: Optional[str] = None
     status: str
-    entity_score: Optional[float] = 0.0
-    min_extracted_entity_score: Optional[float] = 0.0
-    schema_score: Optional[float] = 0.0
+    entity_score: Optional[float] = None
+    min_extracted_entity_score: Optional[float] = None
+    schema_score: Optional[float] = None
     result: Optional[dict] = None
     confidence: Optional[dict] = None
     target_schema: Optional[Schema] = None
diff --git a/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py b/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py
index 15c90f56..1870a4b1 100644
--- a/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py
+++ b/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py
@@ -112,20 +112,15 @@ def find_process_result(step_name: str):
             )
         )
 
-        total_evaluated_fields_count = evaluated_result.confidence.get(
-            "total_evaluated_fields_count", 0
-        )
-        schema_score = (
-            0
-            if total_evaluated_fields_count == 0
-            else round(
-                (
-                    len(evaluated_result.comparison_result.items)
-                    - evaluated_result.confidence["zero_confidence_fields_count"]
-                )
-                / len(evaluated_result.comparison_result.items),
-                3,
-            )
+        # Determine whether per-field confidence could actually be computed.
+        # When `total_evaluated_fields_count == 0`, no field-level confidence
+        # signal was produced (e.g. logprobs unavailable on reasoning models, or
+        # an image flow with no Content Understanding signal). In that case the
+        # entity/schema scores are *unavailable* rather than genuinely zero, and
+        # we propagate ``None`` so downstream consumers (API + UI) can render
+        # an explicit "N/A" instead of a misleading "0%".
+        entity_score, schema_score, min_extracted_entity_score = (
+            self._derive_aggregate_scores(evaluated_result)
         )
 
         processed_result = ContentProcess(
@@ -143,11 +138,9 @@ def find_process_result(step_name: str):
                 self._current_message_context.data_pipeline.pipeline_status.creation_time,
                 "%Y-%m-%dT%H:%M:%S.%fZ",
             ),
-            entity_score=evaluated_result.confidence["overall_confidence"],
+            entity_score=entity_score,
             schema_score=schema_score,
-            min_extracted_entity_score=evaluated_result.confidence[
-                "min_extracted_field_confidence"
-            ],
+            min_extracted_entity_score=min_extracted_entity_score,
             prompt_tokens=evaluated_result.prompt_tokens,
             completion_tokens=evaluated_result.completion_tokens,
             target_schema=Schema.get_schema(
@@ -241,3 +234,41 @@ def _summarize_processed_time(self, step_results: list[StepResult]) -> str:
         # Format the total elapsed time as a string
         formatted_elapsed_time = f"{total_hours:02}:{total_minutes:02}:{total_seconds:02}.{total_milliseconds:03}"
         return formatted_elapsed_time
+
+    @staticmethod
+    def _derive_aggregate_scores(
+        evaluated_result: DataExtractionResult,
+    ) -> tuple[float | None, float | None, float | None]:
+        """Compute ``(entity_score, schema_score, min_extracted_entity_score)``.
+
+        Returns ``(None, None, None)`` when no per-field confidence signal was
+        produced (i.e. ``total_evaluated_fields_count == 0`` or there are no
+        comparison items). This happens, for example, when the LLM call could
+        not return logprobs (reasoning models) and there is no Content
+        Understanding signal to fall back on. Treating that case as "unknown"
+        rather than ``0.0`` lets the API and UI render "N/A" instead of a
+        misleading "0%".
+
+        A genuine zero confidence (e.g. a model that emitted fields but
+        every token had ``logprob == -inf``) is preserved verbatim.
+        """
+        confidence = evaluated_result.confidence or {}
+        total_evaluated_fields_count = confidence.get(
+            "total_evaluated_fields_count", 0
+        )
+        comparison_items = (
+            evaluated_result.comparison_result.items
+            if evaluated_result.comparison_result is not None
+            else []
+        )
+        if total_evaluated_fields_count == 0 or not comparison_items:
+            return (None, None, None)
+
+        zero_count = confidence.get("zero_confidence_fields_count", 0)
+        schema_score = round(
+            (len(comparison_items) - zero_count) / len(comparison_items),
+            3,
+        )
+        entity_score = confidence.get("overall_confidence")
+        min_extracted_entity_score = confidence.get("min_extracted_field_confidence")
+        return (entity_score, schema_score, min_extracted_entity_score)
diff --git a/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py b/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
new file mode 100644
index 00000000..66e2e44c
--- /dev/null
+++ b/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
@@ -0,0 +1,137 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+"""Tests for ``SaveHandler._derive_aggregate_scores``.
+
+Covers the score-availability semantics:
+- valid scores flow through verbatim
+- missing per-field signal yields ``None`` (rendered as "N/A" in the UI)
+- a genuine zero is preserved as ``0`` (rendered as "0%")
+- failed processing (no comparison items) yields ``None``
+"""
+
+from __future__ import annotations
+
+from libs.pipeline.handlers.logics.evaluate_handler.comparison import (
+    ExtractionComparisonData,
+    ExtractionComparisonItem,
+)
+from libs.pipeline.handlers.logics.evaluate_handler.model import DataExtractionResult
+from libs.pipeline.handlers.save_handler import SaveHandler
+
+
+def _make_result(
+    *,
+    items: list[ExtractionComparisonItem],
+    confidence: dict,
+) -> DataExtractionResult:
+    return DataExtractionResult(
+        extracted_result={},
+        confidence=confidence,
+        comparison_result=ExtractionComparisonData(items=items),
+        prompt_tokens=0,
+        completion_tokens=0,
+        execution_time=0,
+    )
+
+
+class TestDeriveAggregateScores:
+    def test_valid_scores_flow_through(self):
+        """A normal evaluate-step result must produce numeric scores."""
+        items = [
+            ExtractionComparisonItem(
+                Field="a", Extracted="x", Confidence="90.00%", IsAboveThreshold="True"
+            ),
+            ExtractionComparisonItem(
+                Field="b", Extracted="y", Confidence="80.00%", IsAboveThreshold="True"
+            ),
+            ExtractionComparisonItem(
+                Field="c", Extracted="z", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+        ]
+        confidence = {
+            "total_evaluated_fields_count": 3,
+            "overall_confidence": 0.567,
+            "min_extracted_field_confidence": 0.0,
+            "zero_confidence_fields_count": 1,
+        }
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=items, confidence=confidence)
+        )
+        assert entity == 0.567
+        # 2 of 3 fields above threshold → 0.667
+        assert schema == round(2 / 3, 3)
+        assert min_score == 0.0
+
+    def test_missing_per_field_signal_returns_none(self):
+        """Reasoning-model / image-only flow: no signal → ``None`` everywhere."""
+        items: list[ExtractionComparisonItem] = []
+        confidence = {
+            "total_evaluated_fields_count": 0,
+            "overall_confidence": 0.0,
+            "min_extracted_field_confidence": 0.0,
+            "zero_confidence_fields_count": 0,
+        }
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=items, confidence=confidence)
+        )
+        assert entity is None
+        assert schema is None
+        assert min_score is None
+
+    def test_no_comparison_items_returns_none(self):
+        """Even if confidence claims fields exist, an empty comparison list is unknown."""
+        confidence = {
+            "total_evaluated_fields_count": 5,
+            "overall_confidence": 0.9,
+            "min_extracted_field_confidence": 0.5,
+            "zero_confidence_fields_count": 0,
+        }
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=[], confidence=confidence)
+        )
+        assert entity is None
+        assert schema is None
+        assert min_score is None
+
+    def test_genuine_zero_score_preserved(self):
+        """A real ``0`` confidence (e.g. all fields below threshold) must NOT become ``None``."""
+        items = [
+            ExtractionComparisonItem(
+                Field="a", Extracted="x", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+        ]
+        confidence = {
+            "total_evaluated_fields_count": 1,
+            "overall_confidence": 0.0,
+            "min_extracted_field_confidence": 0.0,
+            "zero_confidence_fields_count": 1,
+        }
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=items, confidence=confidence)
+        )
+        assert entity == 0.0
+        assert schema == 0.0
+        assert min_score == 0.0
+
+    def test_all_fields_above_threshold(self):
+        items = [
+            ExtractionComparisonItem(
+                Field="a", Extracted="x", Confidence="95.00%", IsAboveThreshold="True"
+            ),
+            ExtractionComparisonItem(
+                Field="b", Extracted="y", Confidence="90.00%", IsAboveThreshold="True"
+            ),
+        ]
+        confidence = {
+            "total_evaluated_fields_count": 2,
+            "overall_confidence": 0.925,
+            "min_extracted_field_confidence": 0.9,
+            "zero_confidence_fields_count": 0,
+        }
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=items, confidence=confidence)
+        )
+        assert entity == 0.925
+        assert schema == 1.0
+        assert min_score == 0.9
diff --git a/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py b/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py
index 22625476..fb08f502 100644
--- a/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py
+++ b/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py
@@ -53,13 +53,13 @@ class Content_Process(EntityBase):
     mime_type: Optional[str] = Field(
         description="MIME type of the processed content file", default=None
     )
-    entity_score: float = Field(
-        description="Score indicating the quality of entity extraction from the content",
-        default=0.0,
+    entity_score: Optional[float] = Field(
+        description="Score indicating the quality of entity extraction from the content. ``None`` means the score was not available (e.g. logprobs unavailable on reasoning models).",
+        default=None,
     )
-    schema_score: float = Field(
-        description="Score indicating the quality of schema matching for the content",
-        default=0.0,
+    schema_score: Optional[float] = Field(
+        description="Score indicating the quality of schema matching for the content. ``None`` means the score was not available.",
+        default=None,
     )
     status: Optional[str] = Field(
         description="Indicates the current status in the content processing pipeline",
diff --git a/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py b/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py
index 5bf0ae23..02ec3edf 100644
--- a/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py
+++ b/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py
@@ -134,9 +134,9 @@ class ContentProcess(BaseModel):
     )
     last_modified_by: Optional[str] = None
     status: Optional[str] = None
-    entity_score: Optional[float] = 0.0
-    min_extracted_entity_score: Optional[float] = 0.0
-    schema_score: Optional[float] = 0.0
+    entity_score: Optional[float] = None
+    min_extracted_entity_score: Optional[float] = None
+    schema_score: Optional[float] = None
     result: Optional[dict] = None
     confidence: Optional[dict] = None
     target_schema: Optional[Schema] = None
diff --git a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx
index ef36c43e..d701a0ff 100644
--- a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx
+++ b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx
@@ -29,7 +29,12 @@ interface DeleteItem {
 interface CellRendererExtraProps {
   readonly txt?: string;
   readonly timeString?: string;
-  readonly valueText?: string;
+  /**
+   * Stringified score value. ``undefined`` (or ``null``) means the score is
+   * not available — the percentage renderer will show "N/A" instead of "0%"
+   * to distinguish "unavailable" from a genuine zero.
+   */
+  readonly valueText?: string | null;
   readonly status?: string;
   readonly lastModifiedBy?: string;
   readonly text?: string | number;
@@ -91,9 +96,31 @@ const CellRenderer: React.FC<CellRendererProps> = ({ type, props }) => {
   };
 
   // Render for percentage
-  const renderPercentage = (valueText: string, status: string) => {
+  const renderPercentage = (valueText: string | null | undefined, status: string) => {
+    // ``null``/``undefined``/empty string === score unavailable. Render an
+    // explicit "N/A" so users can distinguish missing scores from a genuine
+    // zero. (Backends emit ``None``/``null`` when, for example, logprobs were
+    // unavailable on a reasoning model and confidence couldn't be computed.)
+    if (valueText === null || valueText === undefined || valueText === '') {
+      return (
+        <div className="percentageContainer" title="Score not available">
+          <span className="textClass">N/A</span>
+        </div>
+      );
+    }
+
     const decimalValue = Number(valueText);
-    if (isNaN(decimalValue) || status !== 'Completed') {
+    if (isNaN(decimalValue)) {
+      return (
+        <div className="percentageContainer" title="Score not available">
+          <span className="textClass">N/A</span>
+        </div>
+      );
+    }
+
+    // Score is numeric (including a genuine 0): only show "..." while the
+    // document is still being processed.
+    if (status !== 'Completed') {
       return <div className="percentageContainer"><span className="textClass">...</span></div>;
     }
 
@@ -124,7 +151,7 @@ const CellRenderer: React.FC<CellRendererProps> = ({ type, props }) => {
   };
 
   // Render for schema score
-  const calculateSchemaScore = (valueText: string, lastModifiedBy: string, status: string) => {
+  const calculateSchemaScore = (valueText: string | null | undefined, lastModifiedBy: string, status: string) => {
     if (lastModifiedBy === 'user') {
       return (
         <div className="percentageContainer">
@@ -186,9 +213,9 @@ const CellRenderer: React.FC<CellRendererProps> = ({ type, props }) => {
     case 'processTime':
       return renderProcessTimeInSeconds(timeString || '');
     case 'percentage':
-      return renderPercentage(valueText || '', status || '');
+      return renderPercentage(valueText, status || '');
     case 'schemaScore':
-      return calculateSchemaScore(valueText || '', lastModifiedBy || '', status || '');
+      return calculateSchemaScore(valueText, lastModifiedBy || '', status || '');
     case 'text':
       return renderText(text ?? '', 'center');
     case 'date':
diff --git a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx
index 0581b3ac..3b1462d1 100644
--- a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx
+++ b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx
@@ -373,7 +373,11 @@ const ProcessQueueGrid: React.FC<GridComponentProps> = () => {
                                                         <CustomCellRender
                                                             type="percentage"
                                                             props={{
-                                                                valueText: doc.entity_score.toString(),
+                                                                valueText:
+                                                                    doc.entity_score === null ||
+                                                                    doc.entity_score === undefined
+                                                                        ? undefined
+                                                                        : doc.entity_score.toString(),
                                                                 status: doc.status,
                                                             }}
                                                         />
@@ -382,7 +386,11 @@ const ProcessQueueGrid: React.FC<GridComponentProps> = () => {
                                                         <CustomCellRender
                                                             type="percentage"
                                                             props={{
-                                                                valueText: doc.schema_score.toString(),
+                                                                valueText:
+                                                                    doc.schema_score === null ||
+                                                                    doc.schema_score === undefined
+                                                                        ? undefined
+                                                                        : doc.schema_score.toString(),
                                                                 status: doc.status,
                                                             }}
                                                         />
diff --git a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts
index d441eb5a..30ac7a8b 100644
--- a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts
+++ b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts
@@ -17,10 +17,23 @@ export interface ProcessedDocument {
     readonly file_name: string;
     /** MIME type of the document. */
     readonly mime_type: string;
-    /** Entity extraction confidence score (0–1). */
-    readonly entity_score: number;
-    /** Schema compliance score (0–1). */
-    readonly schema_score: number;
+    /**
+     * Entity extraction confidence score in the range 0–1.
+     *
+     * ``null``/``undefined`` means the score was not produced by the backend
+     * (for example: logprobs were unavailable on a reasoning model, or the
+     * pipeline didn't reach the evaluate step). In that case the UI shows
+     * "N/A" rather than a misleading "0%". A genuine numeric ``0`` is still
+     * rendered as ``0%``.
+     */
+    readonly entity_score: number | null | undefined;
+    /**
+     * Schema compliance score in the range 0–1.
+     *
+     * ``null``/``undefined`` means the score was not produced. See
+     * {@link entity_score} for rendering semantics.
+     */
+    readonly schema_score: number | null | undefined;
     /** Current processing status. */
     readonly status: string;
     /** Duration string for processing time (HH:MM:SS). */
diff --git a/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py b/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py
index 470a946c..e88500a3 100644
--- a/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py
+++ b/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py
@@ -77,13 +77,13 @@ class Content_Process(EntityBase):
     mime_type: Optional[str] = Field(
         description="MIME type of the processed content file", default=None
     )
-    entity_score: float = Field(
-        description="Score indicating the quality of entity extraction from the content",
-        default=0.0,
+    entity_score: Optional[float] = Field(
+        description="Score indicating the quality of entity extraction from the content. ``None`` means the score was not available (e.g. logprobs unavailable on reasoning models, image-only flow without Content Understanding).",
+        default=None,
     )
-    schema_score: float = Field(
-        description="Score indicating the quality of schema matching for the content",
-        default=0.0,
+    schema_score: Optional[float] = Field(
+        description="Score indicating the quality of schema matching for the content. ``None`` means the score was not available.",
+        default=None,
     )
     status: Optional[str] = Field(
         description="Indicates the current status in the content processing pipeline",
diff --git a/src/ContentProcessorWorkflow/src/services/content_process_models.py b/src/ContentProcessorWorkflow/src/services/content_process_models.py
index 18c01e58..319de5a5 100644
--- a/src/ContentProcessorWorkflow/src/services/content_process_models.py
+++ b/src/ContentProcessorWorkflow/src/services/content_process_models.py
@@ -80,8 +80,8 @@ class ContentProcessRecord(RootEntityBase):
     processed_time: Optional[str] = None
     imported_time: Optional[datetime] = None
     status: Optional[str] = None
-    entity_score: Optional[float] = 0.0
-    schema_score: Optional[float] = 0.0
+    entity_score: Optional[float] = None
+    schema_score: Optional[float] = None
     result: Optional[Any] = None
     confidence: Optional[Any] = None
 
diff --git a/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py b/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py
index f131c1a2..4c46fa7e 100644
--- a/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py
+++ b/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py
@@ -242,8 +242,11 @@ async def _on_poll(poll_data: dict) -> None:
 
                     status_text = poll_result.get("status", "Failed")
 
-                    schema_score_f = 0.0
-                    entity_score_f = 0.0
+                    # ``None`` here represents "score is not yet available" rather
+                    # than a genuine zero. Preserved through to the API/UI so a
+                    # missing score renders as "N/A" instead of a misleading 0%.
+                    schema_score_f: float | None = None
+                    entity_score_f: float | None = None
                     processed_time = ""
                     result_payload = None
 
@@ -253,18 +256,27 @@ async def _on_poll(poll_data: dict) -> None:
                         )
                         if isinstance(final_payload, dict):
                             status_text = final_payload.get("status") or status_text
-                            try:
-                                schema_score_f = float(
-                                    final_payload.get("schema_score") or 0.0
-                                )
-                            except Exception:
-                                schema_score_f = 0.0
-                            try:
-                                entity_score_f = float(
-                                    final_payload.get("entity_score") or 0.0
-                                )
-                            except Exception:
-                                entity_score_f = 0.0
+
+                            def _coerce_score(value: object) -> float | None:
+                                """Convert a raw score payload to ``float`` or ``None``.
+
+                                Unlike the previous ``float(... or 0.0)`` form, an
+                                explicit ``None`` (score unavailable) is preserved
+                                instead of being silently coerced to ``0.0``.
+                                """
+                                if value is None:
+                                    return None
+                                try:
+                                    return float(value)
+                                except (TypeError, ValueError):
+                                    return None
+
+                            schema_score_f = _coerce_score(
+                                final_payload.get("schema_score")
+                            )
+                            entity_score_f = _coerce_score(
+                                final_payload.get("entity_score")
+                            )
                             try:
                                 processed_time = (
                                     final_payload.get("processed_time") or ""
diff --git a/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py b/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py
index 195b9b36..d64abc84 100644
--- a/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py
+++ b/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py
@@ -42,8 +42,10 @@ def test_defaults(self):
         assert cp.process_id == "p1"
         assert cp.file_name == "doc.pdf"
         assert cp.mime_type is None
-        assert cp.entity_score == 0.0
-        assert cp.schema_score == 0.0
+        # ``None`` is now the default for "score unavailable" so the UI can
+        # render "N/A" rather than a misleading "0%".
+        assert cp.entity_score is None
+        assert cp.schema_score is None
         assert cp.status is None
         assert cp.processed_time == ""
 
@@ -57,6 +59,28 @@ def test_explicit_scores(self):
         assert cp.entity_score == 0.95
         assert cp.schema_score == 0.87
 
+    def test_explicit_zero_score_preserved(self):
+        """A literal ``0`` is a real score and must not become ``None``."""
+        cp = Content_Process(
+            process_id="p1",
+            file_name="doc.pdf",
+            entity_score=0.0,
+            schema_score=0.0,
+        )
+        assert cp.entity_score == 0.0
+        assert cp.schema_score == 0.0
+
+    def test_failed_processing_keeps_scores_none(self):
+        """A failed file must surface unknown scores rather than ``0.0``."""
+        cp = Content_Process(
+            process_id="p1",
+            file_name="doc.pdf",
+            status="Failed",
+        )
+        assert cp.status == "Failed"
+        assert cp.entity_score is None
+        assert cp.schema_score is None
+
 
 # ── Claim_Process ────────────────────────────────────────────────────────────
 
diff --git a/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py b/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py
index 19765025..da66aa83 100644
--- a/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py
+++ b/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py
@@ -152,6 +152,16 @@ def test_construction_with_defaults(self):
         assert rec.id == "r1"
         assert rec.process_id == ""
         assert rec.status is None
+        # ``None`` (rather than ``0.0``) is the sentinel for "score unavailable"
+        # so the UI can render "N/A" instead of a misleading "0%".
+        assert rec.entity_score is None
+        assert rec.schema_score is None
+
+    def test_explicit_zero_score_preserved(self):
+        """A literal ``0.0`` must survive round-trip and not be coerced to ``None``."""
+        rec = ContentProcessRecord(
+            id="r1", process_id="r1", entity_score=0.0, schema_score=0.0
+        )
         assert rec.entity_score == 0.0
         assert rec.schema_score == 0.0
 
diff --git a/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py b/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py
index 36de49c0..0a2e5b8e 100644
--- a/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py
+++ b/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py
@@ -42,8 +42,10 @@ def test_defaults(self):
         assert cp.process_id == "p1"
         assert cp.file_name == "doc.pdf"
         assert cp.mime_type is None
-        assert cp.entity_score == 0.0
-        assert cp.schema_score == 0.0
+        # ``None`` is the sentinel for "score unavailable" so the UI can render
+        # "N/A" instead of "0%".
+        assert cp.entity_score is None
+        assert cp.schema_score is None
         assert cp.status is None
         assert cp.processed_time == ""
 
diff --git a/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py b/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py
index 059b2938..d1665f57 100644
--- a/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py
+++ b/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py
@@ -218,8 +218,10 @@ def test_content_process_record_defaults(self):
         assert record.process_id == ""
         assert record.processed_file_name is None
         assert record.processed_file_mime_type is None
-        assert record.entity_score == 0.0
-        assert record.schema_score == 0.0
+        # ``None`` is the sentinel for "score unavailable" so the UI can render
+        # "N/A" rather than "0%".
+        assert record.entity_score is None
+        assert record.schema_score is None
 
     def test_to_cosmos_dict(self):
         """Test ContentProcessRecord.to_cosmos_dict method"""

From 56c724f390f4096a485be5a96ae991099cd8fe66 Mon Sep 17 00:00:00 2001
From: Prachig-Microsoft <v-prachig@microsoft.com>
Date: Mon, 15 Jun 2026 16:07:20 +0530
Subject: [PATCH 11/13] refactor(scoring): structural completeness fallback
 instead of N/A

Per feedback: Completed runs must always show a meaningful number; Failed runs and genuine zeros stay at 0%.

- save_handler._derive_aggregate_scores picks the best available signal: (1) probabilistic confidence when logprobs available; (2) structural completeness (filled fields / total) when no logprobs (reasoning models, image-only flow); (3) 0.0 when no extraction data at all.

- _is_filled_value heuristic: None/empty/whitespace count as not filled; descends into nested dicts/lists.

- Reverted models from Optional[float]=None back to default 0.0.

- Reverted frontend: no N/A path; renders 0% for null/missing scores.

- 15 new tests covering all 3 paths + _is_filled_value heuristic.
---
 .../src/libs/models/content_process.py        |   6 +-
 .../libs/pipeline/handlers/save_handler.py    |  97 +++++++---
 .../unit/pipeline/test_save_handler_scores.py | 182 ++++++++++++++----
 .../models/contentprocessor/claim_process.py  |  12 +-
 .../contentprocessor/content_process.py       |   6 +-
 .../ProcessQueueGrid/CustomCellRender.tsx     |  39 +---
 .../ProcessQueueGrid/ProcessQueueGrid.tsx     |   4 +-
 .../ProcessQueueGrid/ProcessQueueGridTypes.ts |  21 +-
 .../src/repositories/model/claim_process.py   |  12 +-
 .../src/services/content_process_models.py    |   4 +-
 .../executor/document_process_executor.py     |  24 +--
 .../repositories/test_claim_process_model.py  |  17 +-
 .../services/test_content_process_models.py   |  11 +-
 .../repositories/test_claim_process_model.py  |   7 +-
 .../services/test_content_process_models.py   |   8 +-
 15 files changed, 274 insertions(+), 176 deletions(-)

diff --git a/src/ContentProcessor/src/libs/models/content_process.py b/src/ContentProcessor/src/libs/models/content_process.py
index 018f13b2..c9aacf1f 100644
--- a/src/ContentProcessor/src/libs/models/content_process.py
+++ b/src/ContentProcessor/src/libs/models/content_process.py
@@ -67,9 +67,9 @@ class ContentProcess(BaseModel):
     last_modified_time: datetime.datetime = datetime.datetime.now(datetime.UTC)
     last_modified_by: Optional[str] = None
     status: str
-    entity_score: Optional[float] = None
-    min_extracted_entity_score: Optional[float] = None
-    schema_score: Optional[float] = None
+    entity_score: Optional[float] = 0.0
+    min_extracted_entity_score: Optional[float] = 0.0
+    schema_score: Optional[float] = 0.0
     result: Optional[dict] = None
     confidence: Optional[dict] = None
     target_schema: Optional[Schema] = None
diff --git a/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py b/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py
index 1870a4b1..15f8b878 100644
--- a/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py
+++ b/src/ContentProcessor/src/libs/pipeline/handlers/save_handler.py
@@ -112,13 +112,12 @@ def find_process_result(step_name: str):
             )
         )
 
-        # Determine whether per-field confidence could actually be computed.
-        # When `total_evaluated_fields_count == 0`, no field-level confidence
-        # signal was produced (e.g. logprobs unavailable on reasoning models, or
-        # an image flow with no Content Understanding signal). In that case the
-        # entity/schema scores are *unavailable* rather than genuinely zero, and
-        # we propagate ``None`` so downstream consumers (API + UI) can render
-        # an explicit "N/A" instead of a misleading "0%".
+        # Compute the aggregate scores. Successful (Completed) processing
+        # always yields numeric scores: when probabilistic confidence is
+        # available (logprobs from non-reasoning models / Content Understanding
+        # signal) we use it; otherwise we fall back to a structural
+        # completeness score (fraction of expected fields actually filled).
+        # Failed runs and genuinely empty extractions remain at ``0.0``.
         entity_score, schema_score, min_extracted_entity_score = (
             self._derive_aggregate_scores(evaluated_result)
         )
@@ -235,22 +234,51 @@ def _summarize_processed_time(self, step_results: list[StepResult]) -> str:
         formatted_elapsed_time = f"{total_hours:02}:{total_minutes:02}:{total_seconds:02}.{total_milliseconds:03}"
         return formatted_elapsed_time
 
+    @staticmethod
+    def _is_filled_value(value: object) -> bool:
+        """Heuristic: does an extracted value count as "actually filled"?
+
+        Treats ``None``, empty strings, whitespace-only strings, and empty
+        containers as *not* filled. Recursively descends into dicts/lists so a
+        nested object that contains only nulls is still counted as empty.
+        """
+        if value is None:
+            return False
+        if isinstance(value, bool):
+            return True
+        if isinstance(value, str):
+            return value.strip() != ""
+        if isinstance(value, dict):
+            return any(SaveHandler._is_filled_value(v) for v in value.values())
+        if isinstance(value, (list, tuple, set)):
+            return any(SaveHandler._is_filled_value(v) for v in value)
+        return True
+
     @staticmethod
     def _derive_aggregate_scores(
         evaluated_result: DataExtractionResult,
-    ) -> tuple[float | None, float | None, float | None]:
+    ) -> tuple[float, float, float]:
         """Compute ``(entity_score, schema_score, min_extracted_entity_score)``.
 
-        Returns ``(None, None, None)`` when no per-field confidence signal was
-        produced (i.e. ``total_evaluated_fields_count == 0`` or there are no
-        comparison items). This happens, for example, when the LLM call could
-        not return logprobs (reasoning models) and there is no Content
-        Understanding signal to fall back on. Treating that case as "unknown"
-        rather than ``0.0`` lets the API and UI render "N/A" instead of a
-        misleading "0%".
+        Score selection order:
+
+        1. **Probabilistic confidence** — when the evaluate step produced
+           per-field confidence (``total_evaluated_fields_count > 0``), use the
+           probabilistic ``overall_confidence`` plus the ratio of
+           above-threshold fields. This is the highest-fidelity signal.
+
+        2. **Structural completeness fallback** — when no probabilistic
+           signal was produced (e.g. reasoning models like ``gpt-5``/``o1``/``o3``
+           don't return logprobs, and image-only flow has no Content
+           Understanding signal), but extraction still produced a comparison
+           table, score by *how much of the schema was actually filled*. This
+           replaces the old behaviour of falsely emitting ``0%`` for completed
+           runs that simply lacked logprobs.
 
-        A genuine zero confidence (e.g. a model that emitted fields but
-        every token had ``logprob == -inf``) is preserved verbatim.
+        3. **Zero** — only when there is literally no extraction data
+           (failed pipeline / genuinely empty result). Failed processing
+           continues to surface as ``0`` so the UI consistently renders
+           ``0%`` for failures and genuine zeros.
         """
         confidence = evaluated_result.confidence or {}
         total_evaluated_fields_count = confidence.get(
@@ -261,14 +289,29 @@ def _derive_aggregate_scores(
             if evaluated_result.comparison_result is not None
             else []
         )
-        if total_evaluated_fields_count == 0 or not comparison_items:
-            return (None, None, None)
 
-        zero_count = confidence.get("zero_confidence_fields_count", 0)
-        schema_score = round(
-            (len(comparison_items) - zero_count) / len(comparison_items),
-            3,
-        )
-        entity_score = confidence.get("overall_confidence")
-        min_extracted_entity_score = confidence.get("min_extracted_field_confidence")
-        return (entity_score, schema_score, min_extracted_entity_score)
+        # Path 1: probabilistic confidence
+        if total_evaluated_fields_count > 0 and comparison_items:
+            zero_count = confidence.get("zero_confidence_fields_count", 0)
+            schema_score = round(
+                (len(comparison_items) - zero_count) / len(comparison_items),
+                3,
+            )
+            entity_score = float(confidence.get("overall_confidence") or 0.0)
+            min_extracted_entity_score = float(
+                confidence.get("min_extracted_field_confidence") or 0.0
+            )
+            return (entity_score, schema_score, min_extracted_entity_score)
+
+        # Path 2: structural completeness fallback
+        if comparison_items:
+            filled = sum(
+                1
+                for item in comparison_items
+                if SaveHandler._is_filled_value(item.Extracted)
+            )
+            ratio = round(filled / len(comparison_items), 3)
+            return (ratio, ratio, ratio)
+
+        # Path 3: nothing to score on
+        return (0.0, 0.0, 0.0)
diff --git a/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py b/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
index 66e2e44c..aad71027 100644
--- a/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
+++ b/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
@@ -3,11 +3,12 @@
 
 """Tests for ``SaveHandler._derive_aggregate_scores``.
 
-Covers the score-availability semantics:
-- valid scores flow through verbatim
-- missing per-field signal yields ``None`` (rendered as "N/A" in the UI)
-- a genuine zero is preserved as ``0`` (rendered as "0%")
-- failed processing (no comparison items) yields ``None``
+Covers the score-derivation contract:
+- probabilistic confidence flows through verbatim when available
+- structural completeness fallback fires for Completed runs without logprobs
+  (e.g. reasoning models / image-only flow) instead of emitting a misleading 0%
+- a genuine zero is preserved as ``0.0``
+- failed/empty runs return ``0.0``
 """
 
 from __future__ import annotations
@@ -35,7 +36,7 @@ def _make_result(
     )
 
 
-class TestDeriveAggregateScores:
+class TestProbabilisticPath:
     def test_valid_scores_flow_through(self):
         """A normal evaluate-step result must produce numeric scores."""
         items = [
@@ -63,75 +64,174 @@ def test_valid_scores_flow_through(self):
         assert schema == round(2 / 3, 3)
         assert min_score == 0.0
 
-    def test_missing_per_field_signal_returns_none(self):
-        """Reasoning-model / image-only flow: no signal → ``None`` everywhere."""
-        items: list[ExtractionComparisonItem] = []
+    def test_all_fields_above_threshold(self):
+        items = [
+            ExtractionComparisonItem(
+                Field="a", Extracted="x", Confidence="95.00%", IsAboveThreshold="True"
+            ),
+            ExtractionComparisonItem(
+                Field="b", Extracted="y", Confidence="90.00%", IsAboveThreshold="True"
+            ),
+        ]
         confidence = {
-            "total_evaluated_fields_count": 0,
-            "overall_confidence": 0.0,
-            "min_extracted_field_confidence": 0.0,
+            "total_evaluated_fields_count": 2,
+            "overall_confidence": 0.925,
+            "min_extracted_field_confidence": 0.9,
             "zero_confidence_fields_count": 0,
         }
         entity, schema, min_score = SaveHandler._derive_aggregate_scores(
             _make_result(items=items, confidence=confidence)
         )
-        assert entity is None
-        assert schema is None
-        assert min_score is None
+        assert entity == 0.925
+        assert schema == 1.0
+        assert min_score == 0.9
+
 
-    def test_no_comparison_items_returns_none(self):
-        """Even if confidence claims fields exist, an empty comparison list is unknown."""
+class TestStructuralFallback:
+    """When logprobs are unavailable (reasoning model / image-only) but
+    extraction succeeded, the Completed file must still get a meaningful
+    numeric score based on schema completeness."""
+
+    def test_all_fields_filled_yields_one(self):
+        items = [
+            ExtractionComparisonItem(
+                Field="a", Extracted="x", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+            ExtractionComparisonItem(
+                Field="b", Extracted="y", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+            ExtractionComparisonItem(
+                Field="c", Extracted=42, Confidence="0.00%", IsAboveThreshold="False"
+            ),
+        ]
+        # No probabilistic signal: total_evaluated_fields_count == 0
         confidence = {
-            "total_evaluated_fields_count": 5,
-            "overall_confidence": 0.9,
-            "min_extracted_field_confidence": 0.5,
+            "total_evaluated_fields_count": 0,
+            "overall_confidence": 0.0,
+            "min_extracted_field_confidence": 0.0,
             "zero_confidence_fields_count": 0,
         }
         entity, schema, min_score = SaveHandler._derive_aggregate_scores(
-            _make_result(items=[], confidence=confidence)
+            _make_result(items=items, confidence=confidence)
         )
-        assert entity is None
-        assert schema is None
-        assert min_score is None
+        assert entity == 1.0
+        assert schema == 1.0
+        assert min_score == 1.0
 
-    def test_genuine_zero_score_preserved(self):
-        """A real ``0`` confidence (e.g. all fields below threshold) must NOT become ``None``."""
+    def test_partial_fill_yields_ratio(self):
         items = [
             ExtractionComparisonItem(
                 Field="a", Extracted="x", Confidence="0.00%", IsAboveThreshold="False"
             ),
+            ExtractionComparisonItem(
+                Field="b", Extracted=None, Confidence="0.00%", IsAboveThreshold="False"
+            ),
+            ExtractionComparisonItem(
+                Field="c", Extracted="", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+            ExtractionComparisonItem(
+                Field="d", Extracted="z", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+        ]
+        confidence = {"total_evaluated_fields_count": 0}
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=items, confidence=confidence)
+        )
+        # 2 of 4 fields actually filled → 0.5
+        assert entity == 0.5
+        assert schema == 0.5
+        assert min_score == 0.5
+
+    def test_all_fields_empty_yields_zero(self):
+        """Genuine-empty extraction: structural fallback collapses to ``0.0``."""
+        items = [
+            ExtractionComparisonItem(
+                Field="a", Extracted=None, Confidence="0.00%", IsAboveThreshold="False"
+            ),
+            ExtractionComparisonItem(
+                Field="b", Extracted="", Confidence="0.00%", IsAboveThreshold="False"
+            ),
+            ExtractionComparisonItem(
+                Field="c", Extracted="   ", Confidence="0.00%", IsAboveThreshold="False"
+            ),
         ]
+        confidence = {"total_evaluated_fields_count": 0}
+        entity, schema, min_score = SaveHandler._derive_aggregate_scores(
+            _make_result(items=items, confidence=confidence)
+        )
+        assert entity == 0.0
+        assert schema == 0.0
+        assert min_score == 0.0
+
+
+class TestZeroPath:
+    def test_no_comparison_items_returns_zero(self):
+        """No extraction data at all (failed pipeline) → ``0.0``."""
         confidence = {
-            "total_evaluated_fields_count": 1,
+            "total_evaluated_fields_count": 0,
             "overall_confidence": 0.0,
             "min_extracted_field_confidence": 0.0,
-            "zero_confidence_fields_count": 1,
+            "zero_confidence_fields_count": 0,
         }
         entity, schema, min_score = SaveHandler._derive_aggregate_scores(
-            _make_result(items=items, confidence=confidence)
+            _make_result(items=[], confidence=confidence)
         )
         assert entity == 0.0
         assert schema == 0.0
         assert min_score == 0.0
 
-    def test_all_fields_above_threshold(self):
+    def test_genuine_zero_probabilistic_score_preserved(self):
+        """A real ``0`` confidence (every field below threshold) must NOT be
+        replaced by the structural fallback — it's genuinely 0%."""
         items = [
             ExtractionComparisonItem(
-                Field="a", Extracted="x", Confidence="95.00%", IsAboveThreshold="True"
-            ),
-            ExtractionComparisonItem(
-                Field="b", Extracted="y", Confidence="90.00%", IsAboveThreshold="True"
+                Field="a", Extracted="x", Confidence="0.00%", IsAboveThreshold="False"
             ),
         ]
         confidence = {
-            "total_evaluated_fields_count": 2,
-            "overall_confidence": 0.925,
-            "min_extracted_field_confidence": 0.9,
-            "zero_confidence_fields_count": 0,
+            "total_evaluated_fields_count": 1,
+            "overall_confidence": 0.0,
+            "min_extracted_field_confidence": 0.0,
+            "zero_confidence_fields_count": 1,
         }
         entity, schema, min_score = SaveHandler._derive_aggregate_scores(
             _make_result(items=items, confidence=confidence)
         )
-        assert entity == 0.925
-        assert schema == 1.0
-        assert min_score == 0.9
+        assert entity == 0.0
+        assert schema == 0.0
+        assert min_score == 0.0
+
+
+class TestIsFilledValue:
+    """Coverage for the ``_is_filled_value`` helper used by the structural fallback."""
+
+    def test_none_is_empty(self):
+        assert SaveHandler._is_filled_value(None) is False
+
+    def test_empty_string_is_empty(self):
+        assert SaveHandler._is_filled_value("") is False
+        assert SaveHandler._is_filled_value("   ") is False
+
+    def test_non_empty_string_is_filled(self):
+        assert SaveHandler._is_filled_value("x") is True
+
+    def test_zero_int_is_filled(self):
+        # A literal ``0`` is a valid extracted value (e.g. count fields).
+        assert SaveHandler._is_filled_value(0) is True
+
+    def test_bool_is_filled(self):
+        assert SaveHandler._is_filled_value(False) is True
+        assert SaveHandler._is_filled_value(True) is True
+
+    def test_empty_container_is_empty(self):
+        assert SaveHandler._is_filled_value([]) is False
+        assert SaveHandler._is_filled_value({}) is False
+
+    def test_nested_all_null_is_empty(self):
+        assert SaveHandler._is_filled_value({"a": None, "b": ""}) is False
+        assert SaveHandler._is_filled_value([None, "", {"c": None}]) is False
+
+    def test_nested_with_value_is_filled(self):
+        assert SaveHandler._is_filled_value({"a": None, "b": "x"}) is True
+        assert SaveHandler._is_filled_value([None, "x"]) is True
+
diff --git a/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py b/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py
index fb08f502..75276839 100644
--- a/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py
+++ b/src/ContentProcessorAPI/app/routers/models/contentprocessor/claim_process.py
@@ -53,13 +53,13 @@ class Content_Process(EntityBase):
     mime_type: Optional[str] = Field(
         description="MIME type of the processed content file", default=None
     )
-    entity_score: Optional[float] = Field(
-        description="Score indicating the quality of entity extraction from the content. ``None`` means the score was not available (e.g. logprobs unavailable on reasoning models).",
-        default=None,
+    entity_score: float = Field(
+        description="Score indicating the quality of entity extraction from the content. For Completed runs this is either the probabilistic confidence (when logprobs are available) or a structural completeness fallback (fraction of expected fields actually filled). Failed runs and genuinely empty extractions remain at ``0.0``.",
+        default=0.0,
     )
-    schema_score: Optional[float] = Field(
-        description="Score indicating the quality of schema matching for the content. ``None`` means the score was not available.",
-        default=None,
+    schema_score: float = Field(
+        description="Score indicating the quality of schema matching for the content. For Completed runs this is either the probabilistic above-threshold ratio or a structural completeness fallback. Failed runs remain at ``0.0``.",
+        default=0.0,
     )
     status: Optional[str] = Field(
         description="Indicates the current status in the content processing pipeline",
diff --git a/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py b/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py
index 02ec3edf..5bf0ae23 100644
--- a/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py
+++ b/src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py
@@ -134,9 +134,9 @@ class ContentProcess(BaseModel):
     )
     last_modified_by: Optional[str] = None
     status: Optional[str] = None
-    entity_score: Optional[float] = None
-    min_extracted_entity_score: Optional[float] = None
-    schema_score: Optional[float] = None
+    entity_score: Optional[float] = 0.0
+    min_extracted_entity_score: Optional[float] = 0.0
+    schema_score: Optional[float] = 0.0
     result: Optional[dict] = None
     confidence: Optional[dict] = None
     target_schema: Optional[Schema] = None
diff --git a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx
index d701a0ff..ef36c43e 100644
--- a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx
+++ b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/CustomCellRender.tsx
@@ -29,12 +29,7 @@ interface DeleteItem {
 interface CellRendererExtraProps {
   readonly txt?: string;
   readonly timeString?: string;
-  /**
-   * Stringified score value. ``undefined`` (or ``null``) means the score is
-   * not available — the percentage renderer will show "N/A" instead of "0%"
-   * to distinguish "unavailable" from a genuine zero.
-   */
-  readonly valueText?: string | null;
+  readonly valueText?: string;
   readonly status?: string;
   readonly lastModifiedBy?: string;
   readonly text?: string | number;
@@ -96,31 +91,9 @@ const CellRenderer: React.FC<CellRendererProps> = ({ type, props }) => {
   };
 
   // Render for percentage
-  const renderPercentage = (valueText: string | null | undefined, status: string) => {
-    // ``null``/``undefined``/empty string === score unavailable. Render an
-    // explicit "N/A" so users can distinguish missing scores from a genuine
-    // zero. (Backends emit ``None``/``null`` when, for example, logprobs were
-    // unavailable on a reasoning model and confidence couldn't be computed.)
-    if (valueText === null || valueText === undefined || valueText === '') {
-      return (
-        <div className="percentageContainer" title="Score not available">
-          <span className="textClass">N/A</span>
-        </div>
-      );
-    }
-
+  const renderPercentage = (valueText: string, status: string) => {
     const decimalValue = Number(valueText);
-    if (isNaN(decimalValue)) {
-      return (
-        <div className="percentageContainer" title="Score not available">
-          <span className="textClass">N/A</span>
-        </div>
-      );
-    }
-
-    // Score is numeric (including a genuine 0): only show "..." while the
-    // document is still being processed.
-    if (status !== 'Completed') {
+    if (isNaN(decimalValue) || status !== 'Completed') {
       return <div className="percentageContainer"><span className="textClass">...</span></div>;
     }
 
@@ -151,7 +124,7 @@ const CellRenderer: React.FC<CellRendererProps> = ({ type, props }) => {
   };
 
   // Render for schema score
-  const calculateSchemaScore = (valueText: string | null | undefined, lastModifiedBy: string, status: string) => {
+  const calculateSchemaScore = (valueText: string, lastModifiedBy: string, status: string) => {
     if (lastModifiedBy === 'user') {
       return (
         <div className="percentageContainer">
@@ -213,9 +186,9 @@ const CellRenderer: React.FC<CellRendererProps> = ({ type, props }) => {
     case 'processTime':
       return renderProcessTimeInSeconds(timeString || '');
     case 'percentage':
-      return renderPercentage(valueText, status || '');
+      return renderPercentage(valueText || '', status || '');
     case 'schemaScore':
-      return calculateSchemaScore(valueText, lastModifiedBy || '', status || '');
+      return calculateSchemaScore(valueText || '', lastModifiedBy || '', status || '');
     case 'text':
       return renderText(text ?? '', 'center');
     case 'date':
diff --git a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx
index 3b1462d1..e9026ddb 100644
--- a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx
+++ b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGrid.tsx
@@ -376,7 +376,7 @@ const ProcessQueueGrid: React.FC<GridComponentProps> = () => {
                                                                 valueText:
                                                                     doc.entity_score === null ||
                                                                     doc.entity_score === undefined
-                                                                        ? undefined
+                                                                        ? "0"
                                                                         : doc.entity_score.toString(),
                                                                 status: doc.status,
                                                             }}
@@ -389,7 +389,7 @@ const ProcessQueueGrid: React.FC<GridComponentProps> = () => {
                                                                 valueText:
                                                                     doc.schema_score === null ||
                                                                     doc.schema_score === undefined
-                                                                        ? undefined
+                                                                        ? "0"
                                                                         : doc.schema_score.toString(),
                                                                 status: doc.status,
                                                             }}
diff --git a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts
index 30ac7a8b..3ffc2409 100644
--- a/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts
+++ b/src/ContentProcessorWeb/src/Pages/DefaultPage/Components/ProcessQueueGrid/ProcessQueueGridTypes.ts
@@ -17,23 +17,10 @@ export interface ProcessedDocument {
     readonly file_name: string;
     /** MIME type of the document. */
     readonly mime_type: string;
-    /**
-     * Entity extraction confidence score in the range 0–1.
-     *
-     * ``null``/``undefined`` means the score was not produced by the backend
-     * (for example: logprobs were unavailable on a reasoning model, or the
-     * pipeline didn't reach the evaluate step). In that case the UI shows
-     * "N/A" rather than a misleading "0%". A genuine numeric ``0`` is still
-     * rendered as ``0%``.
-     */
-    readonly entity_score: number | null | undefined;
-    /**
-     * Schema compliance score in the range 0–1.
-     *
-     * ``null``/``undefined`` means the score was not produced. See
-     * {@link entity_score} for rendering semantics.
-     */
-    readonly schema_score: number | null | undefined;
+    /** Entity extraction score (0–1). */
+    readonly entity_score: number;
+    /** Schema compliance score (0–1). */
+    readonly schema_score: number;
     /** Current processing status. */
     readonly status: string;
     /** Duration string for processing time (HH:MM:SS). */
diff --git a/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py b/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py
index e88500a3..75ce41ba 100644
--- a/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py
+++ b/src/ContentProcessorWorkflow/src/repositories/model/claim_process.py
@@ -77,13 +77,13 @@ class Content_Process(EntityBase):
     mime_type: Optional[str] = Field(
         description="MIME type of the processed content file", default=None
     )
-    entity_score: Optional[float] = Field(
-        description="Score indicating the quality of entity extraction from the content. ``None`` means the score was not available (e.g. logprobs unavailable on reasoning models, image-only flow without Content Understanding).",
-        default=None,
+    entity_score: float = Field(
+        description="Score indicating the quality of entity extraction (0.0–1.0). For Completed runs this is either probabilistic confidence (logprobs) or a structural completeness fallback. Failed runs remain at ``0.0``.",
+        default=0.0,
     )
-    schema_score: Optional[float] = Field(
-        description="Score indicating the quality of schema matching for the content. ``None`` means the score was not available.",
-        default=None,
+    schema_score: float = Field(
+        description="Score indicating the quality of schema matching (0.0–1.0). Failed runs remain at ``0.0``.",
+        default=0.0,
     )
     status: Optional[str] = Field(
         description="Indicates the current status in the content processing pipeline",
diff --git a/src/ContentProcessorWorkflow/src/services/content_process_models.py b/src/ContentProcessorWorkflow/src/services/content_process_models.py
index 319de5a5..18c01e58 100644
--- a/src/ContentProcessorWorkflow/src/services/content_process_models.py
+++ b/src/ContentProcessorWorkflow/src/services/content_process_models.py
@@ -80,8 +80,8 @@ class ContentProcessRecord(RootEntityBase):
     processed_time: Optional[str] = None
     imported_time: Optional[datetime] = None
     status: Optional[str] = None
-    entity_score: Optional[float] = None
-    schema_score: Optional[float] = None
+    entity_score: Optional[float] = 0.0
+    schema_score: Optional[float] = 0.0
     result: Optional[Any] = None
     confidence: Optional[Any] = None
 
diff --git a/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py b/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py
index 4c46fa7e..68a81b97 100644
--- a/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py
+++ b/src/ContentProcessorWorkflow/src/steps/document_process/executor/document_process_executor.py
@@ -242,11 +242,12 @@ async def _on_poll(poll_data: dict) -> None:
 
                     status_text = poll_result.get("status", "Failed")
 
-                    # ``None`` here represents "score is not yet available" rather
-                    # than a genuine zero. Preserved through to the API/UI so a
-                    # missing score renders as "N/A" instead of a misleading 0%.
-                    schema_score_f: float | None = None
-                    entity_score_f: float | None = None
+                    # Failed / not-yet-scored documents default to ``0.0``;
+                    # save_handler always emits numeric scores for Completed
+                    # runs (probabilistic if available, otherwise structural
+                    # completeness fallback).
+                    schema_score_f: float = 0.0
+                    entity_score_f: float = 0.0
                     processed_time = ""
                     result_payload = None
 
@@ -257,19 +258,14 @@ async def _on_poll(poll_data: dict) -> None:
                         if isinstance(final_payload, dict):
                             status_text = final_payload.get("status") or status_text
 
-                            def _coerce_score(value: object) -> float | None:
-                                """Convert a raw score payload to ``float`` or ``None``.
-
-                                Unlike the previous ``float(... or 0.0)`` form, an
-                                explicit ``None`` (score unavailable) is preserved
-                                instead of being silently coerced to ``0.0``.
-                                """
+                            def _coerce_score(value: object) -> float:
+                                """Coerce a raw score payload to ``float`` (default ``0.0``)."""
                                 if value is None:
-                                    return None
+                                    return 0.0
                                 try:
                                     return float(value)
                                 except (TypeError, ValueError):
-                                    return None
+                                    return 0.0
 
                             schema_score_f = _coerce_score(
                                 final_payload.get("schema_score")
diff --git a/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py b/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py
index d64abc84..a970555a 100644
--- a/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py
+++ b/src/ContentProcessorWorkflow/tests/unit/repositories/test_claim_process_model.py
@@ -42,10 +42,9 @@ def test_defaults(self):
         assert cp.process_id == "p1"
         assert cp.file_name == "doc.pdf"
         assert cp.mime_type is None
-        # ``None`` is now the default for "score unavailable" so the UI can
-        # render "N/A" rather than a misleading "0%".
-        assert cp.entity_score is None
-        assert cp.schema_score is None
+        # Defaults stay at ``0.0`` so failed/pre-save records render as 0%.
+        assert cp.entity_score == 0.0
+        assert cp.schema_score == 0.0
         assert cp.status is None
         assert cp.processed_time == ""
 
@@ -60,7 +59,7 @@ def test_explicit_scores(self):
         assert cp.schema_score == 0.87
 
     def test_explicit_zero_score_preserved(self):
-        """A literal ``0`` is a real score and must not become ``None``."""
+        """A literal ``0`` is a real score and must survive round-trip."""
         cp = Content_Process(
             process_id="p1",
             file_name="doc.pdf",
@@ -70,16 +69,16 @@ def test_explicit_zero_score_preserved(self):
         assert cp.entity_score == 0.0
         assert cp.schema_score == 0.0
 
-    def test_failed_processing_keeps_scores_none(self):
-        """A failed file must surface unknown scores rather than ``0.0``."""
+    def test_failed_processing_keeps_default_zero(self):
+        """A failed file uses the ``0.0`` default so the UI renders ``0%``."""
         cp = Content_Process(
             process_id="p1",
             file_name="doc.pdf",
             status="Failed",
         )
         assert cp.status == "Failed"
-        assert cp.entity_score is None
-        assert cp.schema_score is None
+        assert cp.entity_score == 0.0
+        assert cp.schema_score == 0.0
 
 
 # ── Claim_Process ────────────────────────────────────────────────────────────
diff --git a/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py b/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py
index da66aa83..c853c2d4 100644
--- a/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py
+++ b/src/ContentProcessorWorkflow/tests/unit/services/test_content_process_models.py
@@ -152,13 +152,14 @@ def test_construction_with_defaults(self):
         assert rec.id == "r1"
         assert rec.process_id == ""
         assert rec.status is None
-        # ``None`` (rather than ``0.0``) is the sentinel for "score unavailable"
-        # so the UI can render "N/A" instead of a misleading "0%".
-        assert rec.entity_score is None
-        assert rec.schema_score is None
+        # Defaults stay at ``0.0`` so failed/pre-save records render as 0%
+        # in the UI; save_handler overwrites with a real numeric score for
+        # Completed runs.
+        assert rec.entity_score == 0.0
+        assert rec.schema_score == 0.0
 
     def test_explicit_zero_score_preserved(self):
-        """A literal ``0.0`` must survive round-trip and not be coerced to ``None``."""
+        """A literal ``0.0`` must survive round-trip."""
         rec = ContentProcessRecord(
             id="r1", process_id="r1", entity_score=0.0, schema_score=0.0
         )
diff --git a/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py b/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py
index 0a2e5b8e..fee05fb1 100644
--- a/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py
+++ b/src/tests/ContentProcessorWorkflow/repositories/test_claim_process_model.py
@@ -42,10 +42,9 @@ def test_defaults(self):
         assert cp.process_id == "p1"
         assert cp.file_name == "doc.pdf"
         assert cp.mime_type is None
-        # ``None`` is the sentinel for "score unavailable" so the UI can render
-        # "N/A" instead of "0%".
-        assert cp.entity_score is None
-        assert cp.schema_score is None
+        # Defaults stay at ``0.0`` so failed/pre-save records render as 0%.
+        assert cp.entity_score == 0.0
+        assert cp.schema_score == 0.0
         assert cp.status is None
         assert cp.processed_time == ""
 
diff --git a/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py b/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py
index d1665f57..5133852a 100644
--- a/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py
+++ b/src/tests/ContentProcessorWorkflow/services/test_content_process_models.py
@@ -218,10 +218,10 @@ def test_content_process_record_defaults(self):
         assert record.process_id == ""
         assert record.processed_file_name is None
         assert record.processed_file_mime_type is None
-        # ``None`` is the sentinel for "score unavailable" so the UI can render
-        # "N/A" rather than "0%".
-        assert record.entity_score is None
-        assert record.schema_score is None
+        # Defaults stay at ``0.0`` so failed/pre-save records render as 0%
+        # in the UI.
+        assert record.entity_score == 0.0
+        assert record.schema_score == 0.0
 
     def test_to_cosmos_dict(self):
         """Test ContentProcessRecord.to_cosmos_dict method"""

From f9c9955d1184dea890d32930ae49ea4fc6900f87 Mon Sep 17 00:00:00 2001
From: Prachig-Microsoft <v-prachig@microsoft.com>
Date: Mon, 15 Jun 2026 16:50:10 +0530
Subject: [PATCH 12/13] fix(lint): remove unused DefaultAzureCredential import
 and fix indentation

- F401: drop unused sync DefaultAzureCredential import in 3 credential util files (sync flow now raises RuntimeError; AsyncDefaultAzureCredential is still used).

- W293/E122: fix blank-line whitespace and continuation-line indentation in ContentProcessorWorkflow/src/utils/credential_util.py.
---
 .../src/libs/utils/azure_credential_utils.py             | 1 -
 src/ContentProcessor/src/libs/utils/credential_util.py   | 1 -
 .../src/utils/credential_util.py                         | 9 ++++-----
 3 files changed, 4 insertions(+), 7 deletions(-)

diff --git a/src/ContentProcessor/src/libs/utils/azure_credential_utils.py b/src/ContentProcessor/src/libs/utils/azure_credential_utils.py
index 5d711e85..3344379c 100644
--- a/src/ContentProcessor/src/libs/utils/azure_credential_utils.py
+++ b/src/ContentProcessor/src/libs/utils/azure_credential_utils.py
@@ -19,7 +19,6 @@
 from azure.identity import (
     AzureCliCredential,
     AzureDeveloperCliCredential,
-    DefaultAzureCredential,
     ManagedIdentityCredential,
 )
 from azure.identity import (
diff --git a/src/ContentProcessor/src/libs/utils/credential_util.py b/src/ContentProcessor/src/libs/utils/credential_util.py
index 1efcaab7..791ab42c 100644
--- a/src/ContentProcessor/src/libs/utils/credential_util.py
+++ b/src/ContentProcessor/src/libs/utils/credential_util.py
@@ -19,7 +19,6 @@
 from azure.identity import (
     AzureCliCredential,
     AzureDeveloperCliCredential,
-    DefaultAzureCredential,
     ManagedIdentityCredential,
 )
 from azure.identity import (
diff --git a/src/ContentProcessorWorkflow/src/utils/credential_util.py b/src/ContentProcessorWorkflow/src/utils/credential_util.py
index fbef0657..306fd180 100644
--- a/src/ContentProcessorWorkflow/src/utils/credential_util.py
+++ b/src/ContentProcessorWorkflow/src/utils/credential_util.py
@@ -19,7 +19,6 @@
 from azure.identity import (
     AzureCliCredential,
     AzureDeveloperCliCredential,
-    DefaultAzureCredential,
     ManagedIdentityCredential,
 )
 from azure.identity import (
@@ -126,11 +125,11 @@ def get_azure_credential():
     logging.info(
         "[AUTH] All CLI credentials failed - falling back to DefaultAzureCredential"
     )
-    
+
     raise RuntimeError(
-    "No Azure authentication available. "
-    "Use Managed Identity in Azure or run "
-    "'az login' / 'azd auth login' locally."
+        "No Azure authentication available. "
+        "Use Managed Identity in Azure or run "
+        "'az login' / 'azd auth login' locally."
     )
 
 

From 8d7b592aba17eda3618ca9d81c65728ef1425627 Mon Sep 17 00:00:00 2001
From: Prachig-Microsoft <v-prachig@microsoft.com>
Date: Mon, 15 Jun 2026 16:57:26 +0530
Subject: [PATCH 13/13] fix(lint): remove trailing blank line at EOF (W391)

---
 .../tests/unit/pipeline/test_save_handler_scores.py              | 1 -
 1 file changed, 1 deletion(-)

diff --git a/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py b/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
index aad71027..be9649d6 100644
--- a/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
+++ b/src/ContentProcessor/tests/unit/pipeline/test_save_handler_scores.py
@@ -234,4 +234,3 @@ def test_nested_all_null_is_empty(self):
     def test_nested_with_value_is_filled(self):
         assert SaveHandler._is_filled_value({"a": None, "b": "x"}) is True
         assert SaveHandler._is_filled_value([None, "x"]) is True
-