VortexSurveillance Dev Environment — Home+HITH Setup

Architecture Overview

OneDrive /HomHITH-Dev/01-Input
      │  (Power Automate trigger: When a file is created)
      ▼
Power Automate Dev Flow          (VortexSurveillance tenant)
      │  HTTP POST (base64 file content + metadata)
      ▼
Azure Function  func-homehith-dev  (Python 3.11, Consumption)
      │
      ├─ SHA-256 hash check → skip if duplicate
      ├─ Azure AI Document Intelligence  doc-intel-homehith-dev  (prebuilt-read OCR)
      ├─ INSERT → SQL: Documents table   (local: DAVID-HP\MSSQLDEVSERVER)
      ├─ Chunk text → INSERT → SQL: DocumentChunks table
      └─ Return JSON result to Power Automate
            │
            ├─ status: processed → move to /02-Output/Processed
            └─ status: error    → move to /02-Output/Failed + email alert

Sign in to portal.azure.com with your VortexSurveillance account.

1.1 Resource Group

Resource Groups → + Create

1.2 Storage Account

Storage accounts → + Create

After creation: Access keys → key1 → Connection string — copy it.

1.3 Azure AI Document Intelligence

Cognitive Services → Document Intelligence → + Create

After creation: Keys and Endpoint — copy Key 1 and Endpoint URL.

1.4 Azure Function App

Function App → + Create

Microsoft Entra ID → App registrations → + New registration

After registering, note the Application (client) ID → referred to as <dev-client-id> below.

2.1 Create Client Secret

Certificates & secrets → + New client secret

Copy the Value immediately — it is shown only once.

2.2 Expose an API

Expose an API → Set (accept default api://<dev-client-id>)

+ Add a scope:

func-homehith-dev → Authentication → + Add identity provider

func-homehith-dev → Configuration → Application settings → + New application setting

Click Save after adding all settings.

File: Ingest/AzureFunction/local.settings.json

{
  "IsEncrypted": false,
  "Values": {
    "AzureWebJobsStorage": "<sthomehithdev-connection-string>",
    "FUNCTIONS_WORKER_RUNTIME": "python",
    "SQL_CONNECTION_STRING": "DRIVER={ODBC Driver 17 for SQL Server};SERVER=DAVID-HP\\MSSQLDEVSERVER;DATABASE=HomeHITH;Trusted_Connection=yes;",
    "DOCUMENT_INTELLIGENCE_ENDPOINT": "https://doc-intel-homehith-dev.cognitiveservices.azure.com/",
    "DOCUMENT_INTELLIGENCE_KEY": "<key1-from-doc-intel-homehith-dev>"
  }
}

Open a terminal in Ingest/AzureFunction/ and run:

# Authenticate to VortexSurveillance tenant
az login --tenant 24528d2a-bbba-4be5-8de3-1bb70cb01654

# Deploy
func azure functionapp publish func-homehith-dev --python

After deployment, retrieve the Function Key:
func-homehith-dev → Functions → process-document → Function Keys → default → Copy

1. Go to make.powerautomate.com — sign in with VortexSurveillance account
2. + Create → Automated cloud flow — Name: HomeHITH Dev Test
3. Trigger: OneDrive for Business — When a file is created
   Folder: /HomHITH-Dev/01-Input
4. Get file content: OneDrive for Business — Get file content
   File: ID from trigger dynamic content
5. HTTP action:

HTTP Body:

{
  "filename":      "@{triggerOutputs()?['body/Name']}",
  "filecontent":   "@{base64(body('Get_file_content'))}",
  "uploaded_by":   "dev@VortexSurveillance.onmicrosoft.com",
  "document_type": "CFS"
}

6. Parse JSON — use this schema:

   {
     "status": "processed",
     "document_id": 42,
     "chunks": 12,
     "extraction_method": "AzureDocumentIntelligence"
   }

7. Condition: body('Parse_JSON')?['status'] is equal to processed
8. Yes branch: OneDrive — Move file → /HomHITH-Dev/02-Output/Processed
9. No branch: OneDrive — Move file → /HomHITH-Dev/02-Output/Failed + send email alert
10. Wrap steps 5–9 in a Scope and add a Run After error branch for timeouts/failures

1. Drop a sample PDF into OneDrive /HomHITH-Dev/01-Input
2. Watch the Power Automate run history — should complete with "status": "processed"
3. Check DAVID-HP\MSSQLDEVSERVER → HomeHITH → Documents and DocumentChunks tables for new rows
4. Confirm the file was moved to /02-Output/Processed

To save as PDF: File → Print → Destination: Save as PDF (or Microsoft Print to PDF) → Save