# Allowed enumerated values

Source: arXiv:2604.05018, App. F.1 Outline Agent prompt.

These are the only values the validator will accept for each enum field.
Anything else will fail `validate_outline.py`.

## `plot_type`

| Value | Use for |
|---|---|
| `"plot"` | Statistical/quantitative charts (line, bar, radar, scatter, box, histogram, heatmap, etc.). The `objective` field MUST also name the specific chart type. |
| `"diagram"` | Conceptual / architectural diagrams (block diagram, flowchart, system overview). Rendered with graphviz or matplotlib patches. |

## `data_source`

| Value | Meaning |
|---|---|
| `"idea.md"` | The figure visualizes a concept or architecture from the idea (typically diagrams). |
| `"experimental_log.md"` | The figure plots numeric values from the experimental log. |
| `"both"` | The figure combines conceptual context with empirical data. |

## `aspect_ratio`

Exactly one of these 12 strings:

```
1:1   1:4   2:3   3:2   3:4   4:1   4:3   4:5   5:4   9:16   16:9   21:9
```

Common conventions:

| Use | Ratio |
|---|---|
| Square ablation grid | `1:1` |
| Single column figure (2-col layout) | `4:3` or `5:4` |
| Wide cross-column figure (2-col layout) | `16:9` or `21:9` |
| Vertical mobile / tall comparison | `9:16` or `2:3` |
| Teaser banner | `21:9` |
| Tall stacked subplots | `3:4` or `4:5` |

The Plotting Agent's `render_matplotlib.py` translates the ratio into pixel
dimensions at 300 DPI.

## `figure_id` rules

- Snake_case: `^[a-z0-9_]+$`
- Convention: prefix with `fig_` (e.g., `fig_main_results`,
  `fig_ablation_temperature_sensitivity`).
- MUST be semantically meaningful (no `fig_1`, `fig_a`, `chart`).
- MUST NOT contain the word `figure` as a substring after the `fig_` prefix.

## Citation hint format

Two acceptable forms (per the prompt):

| You know | Use |
|---|---|
| Exact author and title | `"Author (Exact Paper Title)"` — e.g., `"Vaswani et al. (Attention Is All You Need)"` |
| You don't know either | `"research paper or technical report introducing '[Exact Model/Dataset/Metric Name]'"` — e.g., `"research paper or technical report introducing 'AdamW optimizer'"` |

**Never guess authors.** The Literature Review Agent will use these hints as
search queries — wrong authors lead to wrong citations.

{
  "_comment": "Verbatim example from arXiv:2604.05018, App. F.1, pp. 43-44. Use as a reference for shape and style. Do not copy literally.",
  "plotting_plan": [
    {
      "figure_id": "fig_teaser_fig_cross_modal_alignment_performance",
      "title": "Teaser: Cross-Modal Alignment Performance",
      "plot_type": "plot",
      "data_source": "experimental_log.md",
      "objective": "Visual summary (Radar Chart) demonstrating that our method achieves SOTA balance across 5 metrics.",
      "aspect_ratio": "16:9"
    }
  ],
  "intro_related_work_plan": {
    "introduction_strategy": {
      "hook_hypothesis": "Video-LLMs are currently the dominant paradigm for short clips.",
      "problem_gap_hypothesis": "Context window limits prevent scaling to >5s videos efficiently.",
      "search_directions": [
        "Find highly cited papers establishing the real-world impact of context limits in video generation",
        "Search for published 'long-context video generation' surveys",
        "Identify foundational papers establishing causal video generation"
      ]
    },
    "related_work_strategy": {
      "overview": "Investigate three specific paradigms to build a graph proving the necessity of our Sliding-Window approach.",
      "subsections": [
        {
          "subsection_title": "2.1 Autoregressive Video Generation",
          "methodology_cluster": "Discrete Tokenization & Transformers",
          "sota_investigation_mission": "Identify the current SOTA autoregressive models from 2024-2025. Determine their maximum stable generation length.",
          "limitation_hypothesis": "These models suffer from 'drift' or 'error propagation' because they lack bidirectional context.",
          "limitation_search_queries": [
            "Autoregressive video generation error propagation metrics",
            "Causal masking limitations in temporal video transformers"
          ],
          "bridge_to_our_method": "Our method introduces bidirectional blocks to fix the hypothesized drift issue."
        },
        {
          "subsection_title": "2.2 Diffusion-Based Editing Frameworks",
          "methodology_cluster": "DDIM Inversion & Cross-Attention",
          "sota_investigation_mission": "Find recent papers using DDIM inversion for editing. Identify the standard benchmarks they use.",
          "limitation_hypothesis": "They fail at large structural changes because cross-attention maps are too rigid.",
          "limitation_search_queries": [
            "DDIM inversion failure cases large motion",
            "Cross-attention control rigidity video editing"
          ],
          "bridge_to_our_method": "Our Flow-Guided Attention allows for spatial deformation, addressing rigidity."
        }
      ]
    }
  },
  "section_plan": [
    {
      "section_title": "Abstract",
      "subsections": [
        {
          "subsection_title": "Abstract Content",
          "content_bullets": [
            "Briefly state the problem of temporal inconsistency.",
            "Introduce the proposed method.",
            "Highlight key results."
          ],
          "citation_hints": []
        }
      ]
    },
    {
      "section_title": "3. Methodology",
      "subsections": [
        {
          "subsection_title": "3.1 Temporal-Aware Attention Mechanism",
          "content_bullets": [
            "Define the query-key matching logic",
            "Explain the masking strategy"
          ],
          "citation_hints": [
            "Vaswani et al. (Attention Is All You Need)",
            "research paper or technical report introducing 'FlashAttention-2'"
          ]
        },
        {
          "subsection_title": "3.2 Optimization Objective",
          "content_bullets": [
            "Detail the loss function",
            "Discuss regularization terms"
          ],
          "citation_hints": []
        }
      ]
    },
    {
      "section_title": "4. Experiments",
      "subsections": [
        {
          "subsection_title": "4.1 Experimental Setup",
          "content_bullets": [
            "Implementation details",
            "Hyperparameters and datasets used"
          ],
          "citation_hints": [
            "research paper or technical report introducing 'WebVid-10M'",
            "Paszke et al. (PyTorch: An Imperative Style, High-Performance Deep Learning Library)",
            "research paper or technical report introducing 'AdamW optimizer'",
            "research paper or technical report introducing 'Jaccard Index'"
          ]
        },
        {
          "subsection_title": "4.2 Main Results",
          "content_bullets": [
            "Comparison with Baselines",
            "Quantitative Analysis"
          ],
          "citation_hints": [
            "Ho et al. (Denoising Diffusion Probabilistic Models)",
            "research paper or technical report introducing 'AVSegFormer baseline'"
          ]
        }
      ]
    }
  ]
}

# Outline JSON Schema (prose)

The Outline Agent emits a single JSON object with three top-level keys:
`plotting_plan`, `intro_related_work_plan`, `section_plan`. Source:
arXiv:2604.05018, App. F.1, pp. 43–44.

The machine-readable JSON Schema is in `outline_schema.json` and is enforced
by `scripts/validate_outline.py`. This document is the human-readable view.

## Top-level structure

```json
{
  "plotting_plan": [ <FigureSpec>, ... ],
  "intro_related_work_plan": <IntroRelWorkPlan>,
  "section_plan": [ <SectionSpec>, ... ]
}
```

## FigureSpec

```json
{
  "figure_id":     "fig_framework_overview",            // snake_case, no "Figure"
  "title":         "Framework Overview",
  "plot_type":     "diagram",                           // "plot" or "diagram"
  "data_source":   "idea.md",                           // "idea.md" | "experimental_log.md" | "both"
  "objective":     "Visual summary (Block Diagram) of the proposed pipeline.",
  "aspect_ratio":  "16:9"                               // see allowed-values.md
}
```

If `plot_type` is `"plot"`, the `objective` MUST name the specific chart type
(e.g., "Visual summary (Radar Chart) demonstrating that our method achieves
SOTA balance across 5 metrics.").

## IntroRelWorkPlan

```json
{
  "introduction_strategy": {
    "hook_hypothesis":        "Video-LLMs are currently the dominant paradigm for short clips.",
    "problem_gap_hypothesis": "Context window limits prevent scaling to >5s videos efficiently.",
    "search_directions": [
      "Find highly cited papers establishing the real-world impact of context limits in video generation",
      "Search for published 'long-context video generation' surveys",
      "Identify foundational papers establishing causal video generation"
    ]
  },
  "related_work_strategy": {
    "overview": "Investigate three specific paradigms to build a graph proving the necessity of our Sliding-Window approach.",
    "subsections": [
      {
        "subsection_title":          "2.1 Autoregressive Video Generation",
        "methodology_cluster":       "Discrete Tokenization & Transformers",
        "sota_investigation_mission": "Identify the current SOTA autoregressive models from 2024-2025. Determine their maximum stable generation length.",
        "limitation_hypothesis":     "These models suffer from 'drift' or 'error propagation' because they lack bidirectional context.",
        "limitation_search_queries": [
          "Autoregressive video generation error propagation metrics",
          "Causal masking limitations in temporal video transformers"
        ],
        "bridge_to_our_method": "Our method introduces bidirectional blocks to fix the hypothesized drift issue."
      }
    ]
  }
}
```

## SectionSpec

```json
{
  "section_title": "3. Methodology",
  "subsections": [
    {
      "subsection_title": "3.1 Temporal-Aware Attention Mechanism",
      "content_bullets": [
        "Define the query-key matching logic",
        "Explain the masking strategy"
      ],
      "citation_hints": [
        "Vaswani et al. (Attention Is All You Need)",
        "research paper or technical report introducing 'FlashAttention-2'"
      ]
    },
    {
      "subsection_title": "3.2 Optimization Objective",
      "content_bullets": [
        "Detail the loss function",
        "Discuss regularization terms"
      ],
      "citation_hints": []
    }
  ]
}
```

## Validation rules (enforced by `validate_outline.py`)

1. The top-level object MUST contain exactly the keys `plotting_plan`,
   `intro_related_work_plan`, `section_plan` (extras allowed but reported).
2. `plotting_plan` is a non-empty array.
3. Each FigureSpec has all six fields with correct enum values.
4. `figure_id` matches `^[a-z0-9_]+$` and does not contain `figure` as a
   substring after `fig_`.
5. `intro_related_work_plan.introduction_strategy` and `.related_work_strategy`
   both present.
6. `related_work_strategy.subsections` has 2–4 entries (the paper's
   "2-4 distinct methodology clusters" rule).
7. `section_plan` is a non-empty array. Each SectionSpec has `section_title`
   and `subsections`.
8. **Hierarchy rule**: if a section has any subsection, it must have ≥2
   (no orphans). The validator emits a WARNING for sections with exactly 1
   subsection.
9. `citation_hints` is always an array (may be empty).

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "PaperOrchestra Outline",
  "description": "Schema for outline.json — output of the Outline Agent (Step 1 of arXiv:2604.05018, App. F.1).",
  "type": "object",
  "required": ["plotting_plan", "intro_related_work_plan", "section_plan"],
  "additionalProperties": true,
  "properties": {
    "plotting_plan": {
      "type": "array",
      "minItems": 1,
      "items": { "$ref": "#/$defs/FigureSpec" }
    },
    "intro_related_work_plan": {
      "type": "object",
      "required": ["introduction_strategy", "related_work_strategy"],
      "properties": {
        "introduction_strategy": {
          "type": "object",
          "required": ["hook_hypothesis", "problem_gap_hypothesis", "search_directions"],
          "properties": {
            "hook_hypothesis":        { "type": "string", "minLength": 1 },
            "problem_gap_hypothesis": { "type": "string", "minLength": 1 },
            "search_directions": {
              "type": "array",
              "minItems": 3,
              "maxItems": 8,
              "items": { "type": "string", "minLength": 1 }
            }
          }
        },
        "related_work_strategy": {
          "type": "object",
          "required": ["overview", "subsections"],
          "properties": {
            "overview": { "type": "string", "minLength": 1 },
            "subsections": {
              "type": "array",
              "minItems": 2,
              "maxItems": 4,
              "items": { "$ref": "#/$defs/RelWorkSubsection" }
            }
          }
        }
      }
    },
    "section_plan": {
      "type": "array",
      "minItems": 1,
      "items": { "$ref": "#/$defs/SectionSpec" }
    }
  },
  "$defs": {
    "FigureSpec": {
      "type": "object",
      "required": ["figure_id", "title", "plot_type", "data_source", "objective", "aspect_ratio"],
      "properties": {
        "figure_id": {
          "type": "string",
          "pattern": "^[a-z0-9_]+$",
          "not": { "pattern": "figure(?!_)" }
        },
        "title":     { "type": "string", "minLength": 1 },
        "plot_type": { "enum": ["plot", "diagram"] },
        "data_source": { "enum": ["idea.md", "experimental_log.md", "both"] },
        "objective": { "type": "string", "minLength": 1 },
        "aspect_ratio": {
          "enum": ["1:1", "1:4", "2:3", "3:2", "3:4", "4:1", "4:3", "4:5", "5:4", "9:16", "16:9", "21:9"]
        }
      }
    },
    "RelWorkSubsection": {
      "type": "object",
      "required": [
        "subsection_title", "methodology_cluster", "sota_investigation_mission",
        "limitation_hypothesis", "limitation_search_queries", "bridge_to_our_method"
      ],
      "properties": {
        "subsection_title":           { "type": "string", "minLength": 1 },
        "methodology_cluster":        { "type": "string", "minLength": 1 },
        "sota_investigation_mission": { "type": "string", "minLength": 1 },
        "limitation_hypothesis":      { "type": "string", "minLength": 1 },
        "limitation_search_queries": {
          "type": "array",
          "minItems": 1,
          "items": { "type": "string", "minLength": 1 }
        },
        "bridge_to_our_method":       { "type": "string", "minLength": 1 }
      }
    },
    "SectionSpec": {
      "type": "object",
      "required": ["section_title", "subsections"],
      "properties": {
        "section_title": { "type": "string", "minLength": 1 },
        "subsections": {
          "type": "array",
          "items": {
            "type": "object",
            "required": ["subsection_title", "content_bullets", "citation_hints"],
            "properties": {
              "subsection_title": { "type": "string", "minLength": 1 },
              "content_bullets": {
                "type": "array",
                "minItems": 1,
                "items": { "type": "string", "minLength": 1 }
              },
              "citation_hints": {
                "type": "array",
                "items": { "type": "string", "minLength": 1 }
              }
            }
          }
        }
      }
    }
  }
}

# Outline Agent — verbatim prompt

**Source: arXiv:2604.05018, Appendix F.1, pages 40–44 (verbatim).**

This is the exact system prompt used by the Outline Agent in the paper.
Reproduce it as your system message. Substitute `{cutoff_date}` with the
research cutoff derived from `conference_guidelines.md`.

---

```
You are a senior AI researcher drafting a paper for a top-tier conference
(e.g., NeurIPS, ICML, CVPR, ICLR). Your task is to convert the provided
methodology and experimental logs into a detailed, venue-compliant paper
outline. You must output a single JSON object.

Your inputs are:
  1. idea.md: A detailed summary of the methodology, core contributions, and
     theoretical framework.
  2. experimental_log.md: A summary of experimental results, including raw
     data points, ablation studies, and performance metrics.
  3. template.tex: The template structure. You must use the section commands
     (e.g., \section{...}) found here as your primary skeleton.
  4. conference_guidelines.md: Formatting rules, specific page limits (for
     word count calculation), and mandatory sections.

Processing Directives

Global Instruction: Do not analyze inputs in isolation. You must synthesize
information across all provided documents for every step.

Directive 1: Plotting & Visualization Plan

Synthesize experimental_log.md and idea.md to identify the most compelling
evidence.

  - Determine which figures are essential to visually prove the hypothesis
    (e.g., convergence rates, qualitative visual comparisons).
  - The plot_type MUST be exactly "plot" or "diagram". If it is a plot,
    specify the specific chart type (e.g., Radar Chart) inside the objective.
  - The data_source MUST be exactly "idea.md", "experimental_log.md", or
    "both".
  - Determine the ideal aspect_ratio for each figure. The aspect_ratio MUST
    be exactly one of: "1:1", "1:4", "2:3", "3:2", "3:4", "4:1", "4:3",
    "4:5", "5:4", "9:16", "16:9", "21:9".
  - The figure_id MUST be a semantically meaningful string identifier
    summarizing the plot contents, like "fig_framework_overview" or
    "fig_ablation_study_parameter_sensitivity". It MUST NOT contain the word
    "Figure".
  - Output Focus: Create an array of objects for the plotting_plan key.

Directive 2: Research Graph & Investigation Strategy (Intro & Related Work)

Provide search instructions for a downstream literature review agent to build
a Research Graph. Do not write the actual paper content.

Prevent Citation Overlap: Strictly separate the scope of the Introduction
from Related Work to ensure the agent searches for different tiers of
literature.

  - Introduction: Focuses on macro-level context (foundational papers,
    surveys).
  - Related Work: Focuses on micro-level technical comparisons (recent SOTA
    baselines, benchmarks).

Introduction Strategy (Macro-Level Context, 10-20 papers):

  - Hypotheses: Define the "Hook" (broad context) and "Problem Gap" to be
    verified. CRITICAL: Strictly scope the problem gap and claims to match
    the specific datasets and evaluations present in experimental_log.md.
    Do not over-claim generalization.
  - Search Directions: Provide 3-5 specific queries to find:
    1. Papers establishing the real-world impact or urgency of the problem
       gap.
    2. Good survey or review papers on the topic.
    3. 3-5 Foundational papers that established the sub-field.

Related Work Strategy (Micro-Level Technical Baselines, 30-50 papers):

  - Divide the field into 2-4 distinct methodology clusters that directly
    compete with or precede our approach.
  - For each cluster, define:
    1. Methodology Cluster Name: The technical category.
    2. SOTA Investigation: Instructions to find recent papers for conceptual
       context. CRITICAL TIMELINE RULE: Do not instruct searches for any
       papers published after {cutoff_date}. Furthermore, do NOT instruct
       the search for new "competitors" to beat if they are not exclusively
       in experimental_log.md.
    3. Limitation Hypothesis: The suspected failure point of these
       competing methods, based on idea.md.
    4. Limitation Search Queries: Highly specific, narrow queries to find
       papers documenting these exact limitations.
    5. The Bridge: How our proposed method resolves this specific limitation.

Output Focus: Populate the intro_related_work_plan key.

Directive 3: Section Writing Plan & Sizing Constraints

Outline the remaining sections (Abstract, Methodology, Experiments,
Conclusion, Appendix) into a detailed structural plan.

  - Structural Hierarchy: If Subsection X.1 is created, X.2 is mandatory.
    Do not create orphaned subsections. Omit subsections entirely if a
    section does not require division.
  - Content Specificity: Explicitly reference source materials.
    - Avoid: "Describe the model."
    - Require: "Formalize the Temporal-Aware Attention mechanism using
      Eq. 3 from idea.md."
  - Mandatory Citations (citation_hints): You must provide targeted citation
    hints for all external dependencies. Every hint must point to a single,
    unambiguous canonical paper.
    - Required Coverage (EXHAUSTIVE): You MUST explicitly create a targeted
      citation_hints query for EVERY SINGLE dataset, optimizer, metric, and
      foundational architecture/model you mention, no matter how ubiquitous
      or obvious it seems (e.g., AdamW, ResNet, ImageNet, CLIP, Transformer,
      LLaMA, GPT, LLaVA). If it is in the experimental_log.md or idea.md,
      it MUST have a citation hint.
      1. All baseline methods compared against.
      2. All datasets evaluated on.
      3. All standard metrics utilized.
      4. All foundational algorithms (e.g., ResNet, Transformer, Diffusion
         models), foundational models (e.g., LLMs, VLMs), optimizers
         (e.g., AdamW), or frameworks built upon.
    - Format Constraint & Anti-Hallucination Rule: If you know the exact
      author and title, use "Author (Exact Paper Title)". DO NOT guess or
      hallucinate authors. If you do not know the exact author, use this
      format: "research paper or technical report introducing '[Exact
      Model/Dataset/Metric Name]'".
  - Output Focus: Populate the section_plan key.

Guidelines on Scientific Depth & Mathematical Rigor:

  - Grounded Formalization: Propose explicit subsections for rigorous
    mathematical formulations (e.g., loss functions, core algorithms,
    theoretical proofs). You must base these strictly on idea.md and
    experimental_log.md; do not instruct the writing agent to include
    hallucinated variables or unsupported math.

Strict Output Format (JSON)

You must output a single, valid JSON object with the following three
top-level keys: "plotting_plan", "intro_related_work_plan", and
"section_plan".
```

The full example output JSON from the paper (App. F.1, pp. 43–44) is at
`example-output.json`.

#!/usr/bin/env python3
"""
validate_outline.py — Validate workspace/outline.json against the JSON Schema.

Performs the structural checks defined in
`references/outline_schema.json` plus a few semantic checks the schema
cannot express:

  - figure_id snake_case + no "figure" infix
  - "no orphaned subsection" hierarchy rule
  - related_work_strategy.subsections has 2-4 entries (paper rule)
  - emits WARNINGS for missing-but-recommended things, not just ERRORS

Exit codes:
    0  valid (warnings allowed)
    1  schema or semantic errors

Usage:
    python validate_outline.py /path/to/outline.json
"""
import json
import os
import re
import sys

try:
    import jsonschema
    from jsonschema import Draft202012Validator
except ImportError:
    print("ERROR: jsonschema is required. Install with: pip install jsonschema",
          file=sys.stderr)
    sys.exit(2)

HERE = os.path.dirname(os.path.abspath(__file__))
SCHEMA_PATH = os.path.normpath(
    os.path.join(HERE, "..", "references", "outline_schema.json")
)

FIGID_RE = re.compile(r"^[a-z0-9_]+$")


def load_schema() -> dict:
    with open(SCHEMA_PATH) as f:
        return json.load(f)


def semantic_checks(outline: dict) -> tuple[list[str], list[str]]:
    """Return (errors, warnings)."""
    errors: list[str] = []
    warnings: list[str] = []

    # Figure ID checks
    for i, fig in enumerate(outline.get("plotting_plan", [])):
        fid = fig.get("figure_id", "")
        if not FIGID_RE.match(fid):
            errors.append(
                f"plotting_plan[{i}].figure_id={fid!r} is not snake_case "
                f"matching ^[a-z0-9_]+$"
            )
        # The word "figure" must not appear (allow the conventional fig_ prefix)
        body = re.sub(r"^fig_", "", fid)
        if "figure" in body.lower():
            errors.append(
                f"plotting_plan[{i}].figure_id={fid!r} contains 'figure' — "
                f"forbidden by Outline Agent prompt (App. F.1)"
            )
        # If plot_type is "plot", objective should mention a specific chart type
        if fig.get("plot_type") == "plot":
            obj = fig.get("objective", "").lower()
            chart_keywords = [
                "radar", "bar", "line", "scatter", "box", "violin",
                "heatmap", "histogram", "pie", "stacked", "grouped",
                "ridge", "density", "convergence", "training curve",
            ]
            if not any(k in obj for k in chart_keywords):
                warnings.append(
                    f"plotting_plan[{i}] is plot_type='plot' but objective "
                    f"does not name a specific chart type. The Outline Agent "
                    f"prompt requires this."
                )

    # Section hierarchy: no orphan subsections
    for i, section in enumerate(outline.get("section_plan", [])):
        subs = section.get("subsections") or []
        if len(subs) == 1:
            warnings.append(
                f"section_plan[{i}] {section.get('section_title')!r} has only "
                f"one subsection — this is an orphan. The Outline Agent prompt "
                f"requires that if X.1 exists, X.2 must too."
            )

    return errors, warnings


def main() -> int:
    if len(sys.argv) != 2:
        print(__doc__, file=sys.stderr)
        return 2

    outline_path = sys.argv[1]
    if not os.path.isfile(outline_path):
        print(f"ERROR: {outline_path} not found", file=sys.stderr)
        return 1

    try:
        with open(outline_path) as f:
            outline = json.load(f)
    except json.JSONDecodeError as e:
        print(f"ERROR: {outline_path} is not valid JSON: {e}", file=sys.stderr)
        return 1

    schema = load_schema()
    validator = Draft202012Validator(schema)
    schema_errors = sorted(validator.iter_errors(outline), key=lambda e: e.path)

    sem_errors, sem_warnings = semantic_checks(outline)

    if schema_errors:
        print("SCHEMA ERRORS:", file=sys.stderr)
        for e in schema_errors:
            path = "/".join(str(p) for p in e.absolute_path) or "(root)"
            print(f"  {path}: {e.message}", file=sys.stderr)

    if sem_errors:
        print("SEMANTIC ERRORS:", file=sys.stderr)
        for msg in sem_errors:
            print(f"  {msg}", file=sys.stderr)

    if sem_warnings:
        print("WARNINGS:")
        for msg in sem_warnings:
            print(f"  {msg}")

    if schema_errors or sem_errors:
        print(f"\nFAIL: {outline_path}", file=sys.stderr)
        return 1

    n_figs = len(outline.get("plotting_plan", []))
    n_relwork = len(
        outline.get("intro_related_work_plan", {})
        .get("related_work_strategy", {})
        .get("subsections", [])
    )
    n_sections = len(outline.get("section_plan", []))
    n_hints = sum(
        len(sub.get("citation_hints", []))
        for sec in outline.get("section_plan", [])
        for sub in (sec.get("subsections") or [])
    )
    print(
        f"OK: {outline_path} — {n_figs} figures, "
        f"{n_relwork} related-work clusters, {n_sections} sections, "
        f"{n_hints} citation hints"
    )
    return 0


if __name__ == "__main__":
    sys.exit(main())