Data storage

DynamoDB tables and S3 buckets used in PayPulse Cloud.

PayPulse Cloud uses two storage technologies: DynamoDB for structured invoice and user data, and S3 for raw invoice files and Lambda deployment artifacts.

DynamoDB tables

There are 11 DynamoDB tables in total. All tables have server-side encryption enabled.

RentalInvoices

Stores parsed rental invoice data. This is the primary table for rental invoice records.

Property	Value
Partition key	UserID (String)
Sort key	InvoiceID (String)
Billing mode	Provisioned (5 read / 5 write capacity units, autoscaling 1–10 at 70% target)
Stream	Enabled — NEW_AND_OLD_IMAGES
Point-in-time recovery	Disabled

GSI: due_date_year-due_date_month-index

• Hash key: due_date_year (String)
• Range key: due_date_month (String)
• Projection: ALL

The DynamoDB stream on RentalInvoices is what triggers send_invoice_notification on every new INSERT. See Event flow for the full pipeline.

Users

Stores user account information including a last_retail_invoice_fetch timestamp used for incremental retail invoice ingestion.

Property	Value
Partition key	UserID (String)
Billing mode	Pay per request
Stream	Disabled

GSI: Email-index

• Hash key: Email (String)
• Projection: ALL

RetailInvoices (base table)

The base table for all retail invoice records, regardless of category. Detail tables (below) store category-specific fields keyed by InvoiceID.

Property	Value
Partition key	UserID (String)
Sort key	InvoiceID (String)
Billing mode	Pay per request
Stream	Enabled — NEW_AND_OLD_IMAGES
Point-in-time recovery	Enabled

GSI-1: invoice_date-index

• Hash key: UserID (String)
• Range key: invoice_date (String)
• Projection: ALL
• Use: date-range queries for a single user

GSI-2: sub_type-invoice_date-index

• Hash key: UserID_SubType (String, composite of UserID + _ + sub_type, e.g. user_abc123_food-delivery)
• Range key: invoice_date (String)
• Projection: ALL
• Use: category + date queries (e.g., “all grocery invoices in March”)

Common fields include: vendor_name, sub_type, total_amount, currency, invoice_date, s3_path.

Retail invoice detail tables

Each retail category has its own detail table. All eight share the same schema.

Table name	Category
FoodDeliveryInvoices	Restaurants and food delivery orders
ClothingInvoices	Clothing and fashion purchases
TechnologyInvoices	Electronics and tech products
SubscriptionInvoices	Streaming services and memberships
GroceryInvoices	Supermarket purchases
MiscellaneousUtilityInvoices	Utility bills (electricity, water, internet)
MiscellaneousInvoices	Other retail purchases
TravelInvoices	Transportation invoices (flights, trains, buses)

All detail tables use:

• Partition key: InvoiceID (String)
• Billing mode: Pay per request
• Server-side encryption: Enabled

VendorConfig

Drives automated retail invoice fetching. Adding a row here is all that is needed to start fetching invoices from a new vendor — no code changes required.

Property	Value
Partition key	vendor_id (String)
Billing mode	Pay per request

Common fields include: vendor_name, invoice_sub_type, email_patterns, subject_keywords, parser_type, active (boolean).

S3 buckets

rental-invoices-bucket

Stores all invoice files fetched from Gmail. Versioning is enabled and all public access is blocked. Encryption uses AES-256 (SSE-S3).

Rental invoices (PDF):

rental-invoices-bucket/invoices/{user_id}/rental/{filename}.pdf

Example:

invoices/user_a1b2c3d4/rental/invoice_0001.pdf

Retail invoices (HTML):

rental-invoices-bucket/invoices/{user_id}/retail/{sub_type}/{vendor}_{date}_{hash}.html

The {sub_type} path segment is one of:

sub_type	Description
food-delivery	Restaurant and food delivery orders
clothing	Clothing and fashion purchases
technology	Electronics and tech products
subscriptions	Streaming services and memberships
grocery	Grocery store purchases
utility	Utility bills (electricity, water, etc.)
miscellaneous	Other retail purchases
travel	Transportation invoices (flights, trains, buses, etc.)

{user_id} is a UUID prefixed with user_, generated at signup.

S3 event notifications:

Filter	Trigger
Prefix invoices/, suffix .pdf	Invokes parse_invoice Lambda
Prefix invoices/, suffix .html	Invokes parse_retail_invoice Lambda

Lambda functions bucket

Contains the deployment ZIP archives for all ZIP-based Lambda functions. Versioning is enabled. Lambda has s3:GetObject permission via a bucket policy.

Functions deployed from this bucket:

• fetch_invoices
• fetch_latest_invoice
• fetch_retail_invoices
• send_invoice_notification
• get_rental_invoice
• get_rental_invoices
• delete_user
• get_user_profile
• login_user
• signup_user
• gmail_store_tokens

When you upload a new ZIP version, Terraform detects the updated S3 object version and re-deploys the function automatically on the next terraform apply.