Page Import Table

Overview

Backend-owned lifecycle and source metadata for one public URL capture. Detailed DOM, computed-style, asset, and screenshot evidence is stored in S3; PostgreSQL stores the source request, lifecycle fences, and artifact pointers used to start Code2Des.

Table Definition

Logical Name	Physical Name	Column Name	Data Type	Primary Key	Relation	Unique	Nullable	Default Value	Remarks
Page Import	page_import	id	string	◯					Backend UUIDv7 (varchar(255)); API pageImportId and worker page_import_id
		organization_id	number		organization:id				Organization scope; cascade on update/delete
		project_id	number		project:id				Project scope; cascade on update/delete
		source_kind	string					public_url	Capture source discriminator (varchar(32)); currently always public_url
		source_url	string						Normalized public HTTP(S) URL (text); URL fragments are removed before persistence
		viewport_width	number						Browser viewport width in CSS pixels (integer)
		viewport_height	number						Browser viewport height in CSS pixels (integer)
		device_scale_factor	number					1	Browser device scale factor (real)
		ready_selector	string				◯		Optional readiness selector (varchar(512))
		status	enum					'0'	page_import_status; see Type Codes
		attempt	number					1	Worker-attempt fence (integer); fixed at 1 until a retry path is introduced
		nonce	string						Per-job callback/manifest correlation fence (varchar(255)); not exposed by the user API
		result_url	string				◯		Successful capture artifact or failed artifact pointer (varchar(1000))
		capture_hash	string				◯		SHA-256 identity for the exact completed capture JSON bytes (varchar(80))
		screenshot_url	string				◯		Full-page screenshot artifact pointer (varchar(1000))
		error_message	string				◯		Safe terminal failure detail (text)
		created_at	datetime					CURRENT_TIMESTAMP	Shared audit field
		updated_at	datetime					CURRENT_TIMESTAMP	Shared audit field
		deleted_at	datetime				◯		Shared soft-delete field
		created_by	string						Cognito subject or backend actor (varchar(50))
		updated_by	string						Cognito subject or backend actor (varchar(50))
		deleted_by	string				◯		Cognito subject or backend actor (varchar(50))

Relations

• organization_id → organization.id with update/delete cascade
• project_id → project.id with update/delete cascade
• code2des.page_import_id → page_import.id through a non-unique foreign key, so one completed capture can be reused by multiple Code2Des jobs

Indexes

• PRIMARY KEY (id)
• INDEX page_import_project_status_idx (project_id, status)
• INDEX page_import_org_project_idx (organization_id, project_id)

Type Codes

Database enum: page_import_status.

Value	Meaning
'0'	processing
'1'	completed and eligible as a Code2Des source
'2'	failed

Notes

• The backend inserts the processing row before publishing the SQS message. The worker receives the source request and correlation fields but never receives PostgreSQL credentials.
• The webhook must match the persisted job id, attempt, and nonce; the manifest must repeat the same identity. The database update also requires the row to remain processing and non-deleted.
• A successful callback stores result_url, capture_hash, and screenshot_url. A failed callback stores a safe error_message and may store a failure result_url.
• Code2Des requires the Page Import to be completed with non-null result_url and capture_hash. Triggering Code2Des does not modify or consume the Page Import row.