Task 03 — Industries with Their Users

Difficulty: Beginner–Intermediate
Goal: Add an admin endpoint that returns all industries, each with the list of users belonging to that industry, merged into a single response.

---

What you'll learn

• How to make multiple parallel DB calls from a single service
• How to join/merge two datasets in the service layer
• Using ThreadPoolExecutor for concurrent queries

---

Endpoint spec

GET /api/admin/industries/with-users
Authorization: Bearer <admin-token>

Response 200:
{
  "status": "success",
  "status_code": 200,
  "data": [
    {
      "industryId": "I001",
      "industryName": "Industry A",
      "users": [
        { "userId": "I001_1234_alice", "username": "alice", "type": "DEFAULT", "active": true },
        { "userId": "I001_5678_bob",   "username": "bob",   "type": "OTP",     "active": true }
      ]
    }
  ]
}

Real field names from the backend

• Industry documents use industryId (format: I1, I2, I100) and industryName
• User documents use userId, username, type (DEFAULT/OTP/SSO/microsoft), active, email, phoneNo
• There is no name field on users — the display name field is username

---

Files to create / edit

File	What to do
app/routes/admin/industry_users.py	Define namespace + resource
app/services/admin/industry_users_service.py	Fetch industries + users, merge
app/formatters/admin/industry_users_formatter.py	Shape the merged output
app/apis/admin.py	Register the namespace
app/database/queires_list.py	No new queries needed — use existing methods
app/database/database_supporter.py	No new methods needed — use get_workspace_data() and get_all_users_of_industry()

---

Service — key logic

Use ThreadPoolExecutor to fetch users for all industries concurrently instead of making N sequential calls.

The real DB method to get all users for an industry is DatabaseSupporter.get_all_users_of_industry(industry_id) — it returns id, username, type, active fields.

To get workspace data (all industries), use DatabaseSupporter.get_workspace_data() — it already exists and returns industryId, industryName, meta.logo, meta.sector, meta.subSector.

# app/services/admin/industry_users_service.py
from concurrent.futures import ThreadPoolExecutor
from app.database.database_supporter import DatabaseSupporter

class IndustryUsersService:
    def get_all(self):
        # get_workspace_data() already exists — returns all industries with key fields
        industries = DatabaseSupporter.get_workspace_data()

        def fetch_users(industry):
            # get_all_users_of_industry() already exists — returns id, username, type, active
            users = DatabaseSupporter.get_all_users_of_industry(industry['industryId'])
            return {**industry, 'users': users}

        with ThreadPoolExecutor() as executor:
            results = list(executor.map(fetch_users, industries))

        return results

---

Formatter

class IndustryUsersFormatter:
    def __init__(self, data):
        self.data = data

    def format(self):
        return [
            {
                'industryId':   ind['industryId'],
                'industryName': ind.get('industryName'),   # real field is 'industryName', not 'name'
                'users': [
                    {
                        'userId':   u.get('id'),           # get_all_users_of_industry returns 'id'
                        'username': u.get('username'),
                        'type':     u.get('type'),         # 'DEFAULT', 'OTP', 'SSO', 'microsoft'
                        'active':   u.get('active'),
                    }
                    for u in ind.get('users', [])
                ]
            }
            for ind in self.data
        ]

---

Things to watch out for

• Industries with zero users should still appear in the response, with "users": [].
• Keep the user query scoped to one industry at a time — never SELECT * FROM c across the full users container.

---

Done when

• Response contains all industries
• Each industry object has a users array (empty array is fine if no users)
• Response time stays reasonable even with many industries (parallel fetching works)