-- Migration: 026_DocumentContentHashes
-- Description: Create table for tracking document content hashes to prevent duplicate processing

CREATE TABLE IF NOT EXISTS document_content_hashes (
    id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
    company_id TEXT NOT NULL,
    content_hash TEXT NOT NULL,  -- SHA-256 hash of file content
    original_filename TEXT NOT NULL,
    attachment_id TEXT,  -- Reference to attachment_read_models
    draft_id TEXT,  -- Reference to journal_entry_draft_read_models
    created_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),

    -- Ensure unique hash per company (same document can exist in different companies)
    CONSTRAINT uq_document_hash_company UNIQUE(company_id, content_hash)
);

-- Index for fast lookups by company
CREATE INDEX IF NOT EXISTS idx_document_content_hashes_company_id
    ON document_content_hashes(company_id);

-- Index for hash lookups
CREATE INDEX IF NOT EXISTS idx_document_content_hashes_content_hash
    ON document_content_hashes(content_hash);

COMMENT ON TABLE document_content_hashes IS 'Tracks document content hashes to prevent duplicate processing of the same document';
COMMENT ON COLUMN document_content_hashes.content_hash IS 'SHA-256 hash of the file content';
COMMENT ON COLUMN document_content_hashes.draft_id IS 'Reference to the journal entry draft created from this document';