SanaAI Source

The SanaAI source extracts insight reports from Sana’s learning platform using the query-based reporting API.

Installation

pip install bizon[sana_ai]

Quick Start

name: sana-pipeline

source:
  name: sana_ai
  stream: insight_report
  domain: your-company
  query: "SELECT * FROM users WHERE active = true"
  authentication:
    type: oauth
    params:
      client_id: your-client-id
      client_secret: BIZON_ENV_SANA_SECRET
      token_url: https://your-company.sana.ai/oauth/token

destination:
  name: bigquery
  config:
    project_id: my-project
    dataset_id: learning
    gcs_buffer_bucket: my-bucket

Available Streams

Stream	Description	Incremental
insight_report	Query-based insight reports	No

Configuration

Source Configuration

Field	Type	Required	Description
domain	string	Yes	Sana instance domain (e.g., company for company.sana.ai)
query	string	Yes	SQL-like query for the insight report

Authentication

SanaAI uses OAuth2 authentication:

source:
  name: sana_ai
  stream: insight_report
  domain: your-company
  query: "SELECT * FROM users"
  authentication:
    type: oauth
    params:
      client_id: your-client-id
      client_secret: BIZON_ENV_SANA_SECRET
      token_url: https://your-company.sana.ai/oauth/token

To get OAuth credentials:

1. Contact your Sana administrator
2. Request API access with OAuth credentials
3. You’ll receive a client_id and client_secret

How It Works

The SanaAI source uses an asynchronous job pattern:

1. Create Job: Submits a query to create an insight report job
2. Poll Status: Polls the job status until completion
3. Download Results: Downloads the CSV results when ready
4. Parse Data: Parses CSV into records

This pattern handles large reports that may take time to generate.

Example Configuration

User Activity Report

name: sana-user-activity

source:
  name: sana_ai
  stream: insight_report
  domain: acme-corp
  query: |
    SELECT
      user_id,
      email,
      course_completions,
      last_activity_date
    FROM user_activity
    WHERE last_activity_date >= '2024-01-01'
  authentication:
    type: oauth
    params:
      client_id: your-client-id
      client_secret: BIZON_ENV_SANA_SECRET
      token_url: https://acme-corp.sana.ai/oauth/token

destination:
  name: bigquery
  config:
    project_id: my-project
    dataset_id: learning
    dataset_location: US
    gcs_buffer_bucket: my-staging-bucket

Query Syntax

The query field accepts Sana’s insight query syntax. Consult your Sana documentation for available tables and fields.

Example queries:

-- All users
SELECT * FROM users

-- Active users with completions
SELECT user_id, email, course_completions
FROM users
WHERE active = true

-- Course progress
SELECT course_id, user_id, progress, completed_at
FROM course_progress
WHERE completed_at IS NOT NULL

Data Structure

Records are returned as CSV rows converted to dictionaries. The schema depends on your query’s SELECT columns.

Example record for SELECT user_id, email, name FROM users:

Field	Description
user_id	User ID
email	User email
name	User name

Transforms

Process insight data:

transforms:
  - label: normalize-fields
    python: |
      data = {
        'user_id': data.get('user_id'),
        'email': data.get('email', '').lower(),
        'name': data.get('name'),
        'is_active': data.get('active', '').lower() == 'true'
      }

Next Steps

• Sources Overview - Learn about source connectors
• Authentication - Configure OAuth
• Transforms - Transform extracted data