Federal Patent Docket, Vol. V: K.E.N.T-Ranker Pipeline Source Code and Entity Catalog

This textbook serves as the complete, transparent, and reproducible documentation of the K.E.N.T-Ranker Pipeline, a bespoke software systems engineering framework built to autonomously catalog, normalize, score, and typeset the semi-structured litigation outreach corpus presented in Volume IV. In the modern cognitive era, software engineering dockets submitted as patent priority and whistleblower materials represent a sovereign defense utility. By laying out the complete source code, architectural commentary, and entity metrics of this pipeline on the permanent physical records of the United States Patent and Trademark Office (USPTO), we ensure that this technology remains indestructible, accessible, and referenceable for decades. The Philosophy of the K.E.N.T-Ranker Named in honor of cognitive sovereignty, the K.E.N.T-Ranker is a multi-dimensional natural language pro- cessing and heuristic prioritization engine. It resolves unstructured litigation outreach and whistleblower dockets by: 1. Extracting Unstructured Text Layers: Combining PyPDF and OCR text-layer parsing to capture raw email, receipt, and communication logs Case-Insensitively. 2. Live/Offline NLP Hydration: Mapping standard entity schemas to live Google Knowledge Graph registries, Wikipedia summaries, and Google Natural Language NER (AnalyzeEntities) salience met- rics. 3. Dynamic Priority Scoring: Calibrating score weighting based on psychological and national defense relevance (such as +10,000,000 hardcoded rhetorical points for critical defense actors, x4 multipliers for defense/military organizations, and +9,000 points for General Counsel designations). 4. Multi-dimensional Jaccard Deduplication: Halting exact receipt overlaps by checking page-size, creation dates, and recipient Jaccard indexes simultaneously.