How nuix/ocr-candidates works
Project type and purpose
• This is not a multi‐tier web app but a single Ruby script that plugs into the Nuix forensics/review platform. Its sole purpose is to identify “OCR candidates” within a Nuix case—documents that may need OCR—based on word counts, image sizes, and empty pages.
File structure
• Top-level:
• .gitignore, LICENSE-2.0.txt, README.md – standard repo files
• Ruby/
– OCR Processing (OCR Candidate identification).rb – the only functional component, a Nuix script.
• There are no separate “frontend” or “backend” folders; this is a single plugin module.
README for insights
• Confirms the script runs inside Nuix 8.8+
• Describes UI invocation (“Scripts → OCR Processing”)
• Outlines output: a tag hierarchy for various thresholds (image size, page word-counts, “must” empty docs)
Main components and their interactions
a. Components:
1. Nuix Application Host (Nuix Desktop/Engine)
2. Ruby Script Plugin (“OCR Processing”)
3. Nuix API / SDK
4. Data Source (Nuix Case with documents)
5. UI Dialog (parameter collection)
6. Tag Store / Metadata Repository (within the case)
b. Relationships & data flow:
– User launches Nuix → selects the “OCR Processing” script.
– The script presents a UI dialog to gather thresholds.
– After confirmation, the script calls Nuix APIs to iterate over case items, measuring word counts per page, image sizes, or empty pages.
– Based on configured boundaries, the script writes tags back into the case’s metadata store.
c. Architectural patterns
– Plugin architecture: the Ruby script is dynamically loaded into the Nuix host.
– Event‐driven UI: user action triggers processing.
– Layering:
• UI layer (dialog parameters)
• Processing layer (metrics computation)
• Persistence layer (tagging via API)
d. Key technologies
– Ruby language
– Nuix Scripting API/SDK
– Nuix case file system and metadata tagging