Add scotch-parser semantic comparison and steal plan

[GordonLeong]

Motivation

• Provide an implementation-focused comparison between edgartools 5.19 and the analyzed sec-parser heuristics to decide which deterministic rules to reimplement for scotch-parser v0.1.
• Prioritize small, testable enrichment passes (IDs, sentence splitting, table context, footnotes) that run after edgartools parsing and avoid rewriting sanitize/stamp unless strictly needed.
• Produce a concrete integration plan, minimal stamping deltas, and a test/validation blueprint so downstream reader and card pipelines get stable provenance and sidecar output.

Description

• Add a planning document at docs/internal/planning/scotch-parser-semantic-comparison.md that contains an executive comparison, categorized gap & steal matrix, prioritized top-5/next-5 steals, minimal sanitize/stamp delta guidance, a packages/scotch-parser/ folder sketch, pass sequencing, and a test plan.
• The matrix maps each heuristic from the sec-parser analysis to edgartools status (covered/partial/not) and gives a steal decision with proposed pass name, inputs, deterministic rules, outputs, risks, and minimal tests for each YES.
• Recommends v0.1 top steals (assign_semantic_ids_pass, sentence_split_pass, table_context_linking_pass, table_type_refinement_pass, footnote_linking_pass) and minimal, opt-in stamping deltas (--stamp-table-cells, --stamp-footnote-refs) only if required.
• Specifies how to call edgar.documents.parse_html as the base engine, enrichment pass ordering, sidecar schema expectations, and validation metrics for deterministic quality checks.

Testing

• This change is documentation-only and did not modify parser runtime code or unit tests.
• No automated unit tests were added or executed as part of this PR because the change is a planning/spec document.
• The new file presence and contents were validated locally by listing and printing the file to ensure it contains the planned comparison, matrix, implementation plan, and test plan.

---

Codex Task