docs/data-models.md

# Data Models: ResearchLink AI

All data models are Pydantic v2 BaseModel subclasses.

## IngestionInput (`schemas/inputs.py`)

User-provided inputs. Requires at least one source.

Key fields: `paper_pdf_path`, `paper_pdf_url`, `paper_url`, `github_url`, `bibtex`, `author_notes`, `target_output_dir`, `force_overwrite`

## ResolvedSources (`schemas/inputs.py`)

After source resolution. Tracks what's available.

Key fields: `local_pdf_path`, `pdf_available`, `paper_url_reachable`, `github_url_reachable`

## PaperExtraction (`schemas/paper.py`)

Raw text and structure from PDF.

Key fields: `full_text`, `title_candidates`, `abstract`, `section_headings`, `references_raw`, `extraction_complete`, `confidence`

## PaperMetadata (`schemas/paper.py`)

Structured paper metadata.

Key fields: `id`, `title`, `year`, `status`, `venue`, `authors_provisional`, `areas`, `tags`, `slug`, `verification`, `confidence`

## CitationEntry + CitationReport (`schemas/citations.py`)

Per-reference and aggregate citation analysis.

Key enum: `CitationStatus` — verified, needs-verification, placeholder, incomplete, conflict

## RepoAnalysis (`schemas/repository.py`)

GitHub repository inspection results.

Key fields: `owner`, `repo_name`, `readme_content`, `detected_language`, `detected_frameworks`, `setup_commands`, `is_likely_official_impl`

## GenerationReport (`schemas/outputs.py`)

Final pipeline summary.

Key fields: `slug`, `output_dir`, `files_generated`, `errors`, `warnings`, `quality_flags`, `success`, `duration_seconds`