notes.txt

Here's a summary of the backend design decisions based on the given context:

1. Modularized Lambda Functions
    Each lambda function has a single responsibility, such as:

 - set-reminder-by-text: Handles natural language parsing, schedules reminders in EventBridge or Scheduler, and stores them in DynamoDB.
 - get-reminders: Queries DynamoDB to fetch past/upcoming reminders, optionally retrieving EventBridge schedules for future reminders.

Benefits:
 - Scalability: Each function can scale independently based on demand.
 - Ease of Maintenance: Issues are isolated to specific functions, making debugging and enhancements easier.

2. DynamoDB as the Primary Data Store

Structure:
 - Partition Key (PK): CUSTOMER#{device_id}
 - Sort Key (SK): REMINDER#{reminder_id}

Additional Attributes: start_date, time, repeat_frequency, is_completed, updated_at, eventbridge_expression, etc.

Benefits:

 - Flexibility: DynamoDB's schema-less design allows easy addition of new fields.
 - Performance: Efficient querying for reminders by device_id and filtering by is_completed or dates.

3. EventBridge for Scheduling

EventBridge Rule Types:
 - cron expressions for precise scheduling (e.g., every Monday at 10 AM).
 - rate expressions for recurring intervals (e.g., every 2 days).
 - at expressions for one-time events.
Decisions:

One-Time Reminders: Use at expressions.
Recurring Reminders: Use rate or cron expressions based on repeat frequency.
Event Parsing: EventBridge expressions are parsed using croniter or regex to generate next occurrences.
Benefits:

Reliability: Native AWS service ensures high availability.
Customizability: Supports complex schedules like specific days of the week or month.
4. Scheduler for One-Time Events
AWS Scheduler is used to trigger single-use reminders when a one-time reminder is created.
Why?
Scheduler provides a lightweight alternative for single-event triggering, avoiding unnecessary EventBridge rule management.
5. Handling Natural Language Parsing
Model Used: OpenAI gpt-3.5-turbo for text-to-JSON parsing.
Steps:
Parse text into a structured Reminder object (e.g., task, start date, time, repeat frequency).
Default handling for missing time (11:00 AM) or start date (today).
Pydantic models ensure data integrity and structure validation.
Benefits:

Flexibility: Can handle various natural language formats like "every alternate day" or "next Monday at 5 PM."
Extensibility: The logic can be improved for additional natural language patterns.
6. Error Handling
DynamoDB Writes: If reminders cannot be directly scheduled, they are added to an SQS queue for error handling and retry.
ClientError Handling: Specific AWS exceptions (e.g., EventBridge failures) are logged, and fallback actions (like SQS message enqueue) are performed.
Benefits:

Resilience: Ensures failed operations are retried or logged for manual review.
Transparency: Clear logging of errors for debugging.
7. API Design
GET /get-reminder-list

Query reminders filtered by past or upcoming.
Supports include_schedule to fetch next occurrences using parsed EventBridge expressions.
POST /set-reminder-by-text

Accepts natural language input (reminder_data.text) and processes it into a structured reminder.
Request/Response Structure:

JSON-based, with clear field names (start_date, time, repeat_frequency) and default values for missing inputs.
Example response includes reminder_id, reminder_scheduled_message, and next_occurrences.
Benefits:

Consistency: JSON response formats are easy to consume by the frontend.
Simplicity: Separate endpoints for creating, querying, and managing reminders.
8. Repeat Frequency and Flexibility
Repeat Frequency: Handles both fixed intervals (daily, weekly) and custom schedules (selected_days_of_week).
Custom Day Mapping:
AWS day-of-week (1=Sunday) mapped to croniter (0=Sunday) for accurate cron parsing.
Benefits:

Flexibility: Supports complex user requirements, like reminders on specific days of the week or month.
Standardization: All custom frequencies are converted into AWS-compatible cron/rate expressions.
9. Parsing EventBridge Expressions
Goal: Generate next occurrences of reminders for upcoming schedules.
Steps:
Use croniter for cron expressions.
Use regex for rate expressions (e.g., "rate(2 days)").
Return parsed occurrences to the frontend for display.
Benefits:

User Experience: Users can see exact upcoming occurrences of reminders.
Predictability: Ensures reminders are consistent with their configured schedule.
10. CORS and Security
CORS Enabled: Allows cross-origin requests from the frontend.
Security Measures:
Headers restricted to specific origins/domains.
IAM roles ensure that only authorized services (like Scheduler, EventBridge) can interact with the system.
Overall Benefits
Scalable and Modular: Lambda functions can handle independent tasks and scale individually.
Performance Optimized: DynamoDB and EventBridge ensure low-latency, high-throughput operations.
Future-Proof: Design can easily accommodate new features like more complex scheduling or integration with other AWS services.
These decisions ensure a robust and extensible backend for your reminder system.