Skip to content

ujjalcal/title-review

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 

Repository files navigation

Mortgage Title Rule Pipeline - Dataset Pack

Finalized dataset pack for mortgage title rule compliance pipeline.

Directory Structure

title-review/
├── datasets/
│   ├── inputs/                      # T0 - Authoritative source data
│   │   ├── loan.json                # Core loan data (LOS/MISMO/ULAD)
│   │   ├── title_commitment_document.json  # PDF metadata
│   │   └── selling_guide_requirement.json  # Selling guide requirements
│   │
│   ├── rules/                       # Rule definitions and execution
│   │   ├── rule_proposal.json       # T1 - Rule definitions
│   │   ├── rule_run.json            # T4 - Batch run metadata
│   │   └── rule_result.json         # T4 - Per-loan rule results
│   │
│   ├── extracts/                    # Parsed title commitment data
│   │   ├── title_commitment_extract.json     # T2 - Raw OCR extract
│   │   └── title_commitment_structured.json  # T3 - Normalized data
│   │
│   └── outputs/                     # Pipeline outputs
│       ├── title_compliance_exception.json  # T5A - Compliance findings
│       └── title_cure.json                  # T5B - Cure requests
│
├── schemas/
│   └── dataset_schemas.json         # Schema definitions for all datasets
│
└── docs/
    └── lineage.md                   # End-to-end data lineage documentation

Datasets Overview

Input Datasets (T0)

Dataset PK Description
Loan loan_id Core loan data from LOS/MISMO/ULAD
TitleCommitmentDocument commitment_doc_id Uploaded title commitment PDF metadata
SellingGuideRequirement requirement_id Curated selling guide requirement chunks

Rule Datasets

Dataset PK Description
RuleProposal rule_id Rule definitions with logic and applicability
RuleRun rule_run_id Batch metadata for rule execution
RuleResult (rule_run_id, rule_id, loan_id) Per-loan per-rule evaluation results

Extract Datasets

Dataset PK Description
TitleCommitmentExtract extract_id Raw extracted fields with confidence scores
TitleCommitmentStructured commitment_id Normalized rule-ready commitment data

Output Datasets

Dataset PK Description
TitleComplianceException exception_id Auditable compliance exceptions
TitleCure cure_id Operational cure requests

Key Design Principles

  • Stable Keys: loan_id, commitment_id, rule_id enable easy joins
  • Owner vs Loan Policy Separated: Mortgage-correct policy distinction
  • Outputs Split: Compliance (audit) and Cure (operations) separated
  • Scalable: Design supports 100+ rules without schema changes

Pipeline Flow

Loan + Parsed Title Commitment → RuleResult → TitleComplianceException → TitleCure

See docs/lineage.md for complete lineage documentation.

Sample Rule

The included sample demonstrates the "Title Insurance Coverage >= Original Principal Amount" rule:

  • Loan: $100,000 original principal amount
  • Title Commitment: $90,000 loan policy amount
  • Result: FAIL (delta: -$10,000)
  • Exception: HIGH severity, TitleInsuranceCoverage category
  • Cure: Request to increase loan policy to $100,000

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors