Submit Worker Deployment Guide

Overview

The Submit Worker (submit.michaeljwright.com.au) is a Cloudflare Worker that provides a user-friendly interface for non-technical users to ingest content into the Michael J Wright Digital Archive.

Features:

• Multi-file upload (up to 100 files, 50MB each)
• Automatic catalog ID generation (MJW-{TYPE}-{YEAR}-{NNN})
• Bulk metadata for entire submission (applies to all files)
• Per-file title and metadata overrides
• Dublin Core metadata mapping
• Direct Fedora 6 integration
• Cloudflare Zero Trust SSO support (managed via Cloudflare dashboard)

Location: /workers/submit-ingest/

Project Structure

workers/
└── submit-ingest/
    ├── wrangler.toml          # Cloudflare Worker configuration
    └── src/
        └── index.js           # Worker logic + form UI

Setup & Deployment

Step 1: Install Dependencies

In the submit-ingest directory:

cd workers/submit-ingest
npm install wrangler@latest --save-dev

Step 2: Configure Wrangler

The wrangler.toml is already configured for:

• Route: submit.michaeljwright.com.au/*
• Zone: michaeljwright.com.au
• Environment variables for Fedora URL and file size limits

Step 3: Set Secrets

Before deployment, configure the Fedora authentication secret:

# Navigate to submit-ingest directory
cd workers/submit-ingest

# Set the FEDORA_BASIC_AUTH secret (base64 encoded username:password)
npx wrangler secret put FEDORA_BASIC_AUTH
# When prompted, paste: Basic base64(curator1:password)
# Example: Basic Y3VyYXRvcjE6Q2hhbmdlVGhpcyBDdXJhdG9yUGFzc3dvcmQ3ODkh

To generate the base64 string locally:

# In PowerShell
$credentials = "curator1:ChangeThisCuratorPassword789!"
$bytes = [System.Text.Encoding]::UTF8.GetBytes($credentials)
$base64 = [Convert]::ToBase64String($bytes)
Write-Host "Basic $base64"

Then paste the output into the secret prompt.

Step 4: Deploy Worker

npx wrangler deploy

Expected output:

✓ Published mjw-submit-ingest@1.0.0
✓ Deployment successful!
- https://submit.michaeljwright.com.au/

Step 5: Configure Cloudflare Zero Trust (Optional)

To protect the submit interface with SSO:

1. Go to Cloudflare Zero Trust dashboard
2. Navigate to Access → Applications
3. Create new application:
   • Domain: submit.michaeljwright.com.au
   • Type: Public hostname
   • Policy: Allow authenticated users (your preferred IdP: Google, GitHub, etc.)
4. Save and apply

The worker will automatically respect Cloudflare Access headers.

Usage Flow

For Non-Technical Users

1. Navigate to https://submit.michaeljwright.com.au/
2. Step 1: Select content type (Painting, Photograph, Poem, Notebook), series name, and year
3. Step 2: Upload files via drag-drop or file browser (max 100 files, 50MB each)
4. Step 3: Review auto-generated catalog IDs (MJW-P-2024-001, MJW-P-2024-002, etc.)
5. Step 4: Enter bulk metadata (description, keywords, medium, dimensions, condition)
6. Step 5: Add unique title for each file; optionally override bulk metadata per-file
7. Step 6: Review and submit

For Curators

Files are ingested directly to Fedora at:

/fcrepo/rest/{type}s/{catalog-id}/
├── Metadata (Dublin Core JSON-LD)
└── Binary file attachment

All metadata is Dublin Core compliant and immediately queryable via Fedora REST API.

API Endpoints

GET /

Returns the submission form HTML interface.

curl https://submit.michaeljwright.com.au/

POST /api/submit

Ingests files and metadata to Fedora.

Request:

• Content-Type: multipart/form-data
• Fields:
  • files (multiple): File objects to upload
  • contentType: P, PH, PM, or NB
  • creationYear: 4-digit year
  • seriesName: Collection/series name (optional)
  • startingNumber: Sequential number start (001-999)
  • bulkDescription, bulkKeywords, bulkMedium, bulkDimensions, bulkCondition
  • individualMetadata: JSON object with per-file overrides

Response:

{
  "ingested": 40,
  "total": 40,
  "errors": null
}

Fedora Integration

The worker creates Fedora containers with the following Dublin Core mapping:

File Validation

Accepted File Types:

• Images: .jpg, .jpeg, .png, .gif, .tiff, .webp
• Documents: .pdf, .doc, .docx, .txt
• Any type up to 50MB

Client-side validation:

• Max 100 files per submission
• Max 50MB per file
• Prevents duplicate file names in single submission

Server-side validation:

• Fedora authentication check
• Catalog ID conflict detection (future)
• MIME type validation

Monitoring & Troubleshooting

Check Deployment Status

npx wrangler deployments list

View Live Logs

npx wrangler tail

Test File Upload Locally

npx wrangler dev
# Opens http://localhost:8787/

Common Issues

Issue: "Fedora authentication not configured"

• Fix: Ensure FEDORA_BASIC_AUTH secret is set (npx wrangler secret put FEDORA_BASIC_AUTH)

Issue: Files uploaded but not appearing in Fedora

• Fix: Check Fedora logs and verify catalog ID matches path structure
• Ensure Cloudflare Tunnel is connected and healthy

Issue: "Maximum 100 files allowed" error

• Fix: Submit in batches of 100 or fewer

Updating the Worker

To make changes to the submission form or ingestion logic:

1. Edit /workers/submit-ingest/src/index.js
2. Test locally: npx wrangler dev
3. Deploy: npx wrangler deploy

Integration with Pages

The homepage at https://data.michaeljwright.com.au/ includes:

• CTA Button: "Submit Content" → https://submit.michaeljwright.com.au/
• Card: "Submit Content" with description
• Back Link: Submit form includes link back to main site

Users can seamlessly navigate between documentation, archive browse, and content submission.

Next Steps

• Deploy worker with npx wrangler deploy
• Configure Fedora BASIC_AUTH secret
• Set up Cloudflare Zero Trust Access policy (if desired)
• Test end-to-end: upload test files, verify in Fedora
• Document for non-technical users
• Monitor initial submissions via Wrangler logs

Questions?

Refer to:

• Worker code: /workers/submit-ingest/src/index.js
• Main archive docs: /docs/
• Fedora API docs: https://wiki.lyrasis.org/display/FEDORA6x/