| 1 | #!/usr/bin/env python3 |
| 2 | """ |
| 3 | PPT Master - Project Specification Helpers |
| 4 | |
| 5 | Scaffold and validate the Markdown planning artifacts used by project_manager.py. |
| 6 | The module keeps schema parsing and deterministic scaffold rendering independent |
| 7 | from the broader project-management command surface. |
| 8 | |
| 9 | Usage: |
| 10 | Import validate_project_artifacts() or scaffold_project_artifact(). |
| 11 | |
| 12 | Examples: |
| 13 | from project_management.project_specs import validate_markdown_schema |
| 14 | |
| 15 | Dependencies: |
| 16 | None (only uses the standard library and local project modules) |
| 17 | """ |
| 18 | |
| 19 | from __future__ import annotations |
| 20 | |
| 21 | import json |
| 22 | import math |
| 23 | import re |
| 24 | from pathlib import Path |
| 25 | from typing import Mapping |
| 26 | |
| 27 | from .paths import ( |
| 28 | SCAFFOLD_DIR, |
| 29 | SCHEMA_DIR, |
| 30 | SCRIPTS_DIR, |
| 31 | SKILL_DIR, |
| 32 | ) |
| 33 | |
| 34 | try: |
| 35 | from project_utils import ( |
| 36 | CANVAS_FORMATS, |
| 37 | get_project_info as get_project_info_common, |
| 38 | validate_communication_trace, |
| 39 | ) |
| 40 | except ImportError: |
| 41 | import sys |
| 42 | |
| 43 | tools_dir = SCRIPTS_DIR |
| 44 | if str(tools_dir) not in sys.path: |
| 45 | sys.path.insert(0, str(tools_dir)) |
| 46 | from project_utils import ( # type: ignore |
| 47 | CANVAS_FORMATS, |
| 48 | get_project_info as get_project_info_common, |
| 49 | validate_communication_trace, |
| 50 | ) |
| 51 | |
| 52 | try: |
| 53 | from visualization_catalog import ( |
| 54 | LEGACY_STRUCTURE_INTENT_KIND, |
| 55 | VISUALIZATION_SVG_KIND, |
| 56 | VisualizationCatalogError, |
| 57 | resolve_visualization_reference, |
| 58 | ) |
| 59 | except ImportError: |
| 60 | import sys |
| 61 | |
| 62 | if str(SCRIPTS_DIR) not in sys.path: |
| 63 | sys.path.insert(0, str(SCRIPTS_DIR)) |
| 64 | from visualization_catalog import ( # type: ignore |
| 65 | LEGACY_STRUCTURE_INTENT_KIND, |
| 66 | VISUALIZATION_SVG_KIND, |
| 67 | VisualizationCatalogError, |
| 68 | resolve_visualization_reference, |
| 69 | ) |
| 70 | |
| 71 | |
| 72 | TOOLS_DIR = SCRIPTS_DIR |
| 73 | |
| 74 | _CUSTOM_REFERENCE_CATALOGS = ( |
| 75 | ("mode", "mode", "mode_references", SKILL_DIR / "references" / "modes"), |
| 76 | ( |
| 77 | "visual_style", |
| 78 | "visual_style", |
| 79 | "visual_style_references", |
| 80 | SKILL_DIR / "references" / "visual-styles", |
| 81 | ), |
| 82 | ( |
| 83 | "colors", |
| 84 | "image_rendering", |
| 85 | "image_rendering_references", |
| 86 | SKILL_DIR / "references" / "image-renderings", |
| 87 | ), |
| 88 | ) |
| 89 | |
| 90 | _MARKDOWN_H2_RE = re.compile(r"^##[ \t]+(.+?)[ \t]*$", re.MULTILINE) |
| 91 | _MARKDOWN_SUBHEADING_RE = re.compile(r"^#{3,6}[ \t]+(.+?)[ \t]*$", re.MULTILINE) |
| 92 | _MARKDOWN_DATA_LINE_RE = re.compile( |
| 93 | r"^[ \t]*-[ \t]+(?:\*\*)?([^:\n*]+?)(?:\*\*)?[ \t]*:[ \t]*(.*)$", |
| 94 | re.MULTILINE, |
| 95 | ) |
| 96 | _IMAGE_PATH_SUFFIXES = frozenset( |
| 97 | { |
| 98 | ".bmp", |
| 99 | ".emf", |
| 100 | ".gif", |
| 101 | ".jpeg", |
| 102 | ".jpg", |
| 103 | ".png", |
| 104 | ".svg", |
| 105 | ".tif", |
| 106 | ".tiff", |
| 107 | ".webp", |
| 108 | ".wmf", |
| 109 | } |
| 110 | ) |
| 111 | _IMAGE_ACQUISITION_SOURCES = frozenset( |
| 112 | {"ai", "web", "user", "formula", "placeholder", "slice"} |
| 113 | ) |
| 114 | _IMAGE_CROP_POLICIES = frozenset({"adaptive", "no-crop"}) |
| 115 | _LEGACY_IMAGE_METADATA_KEYS = frozenset( |
| 116 | { |
| 117 | "image_rendering", |
| 118 | "image_rendering_behavior", |
| 119 | "image_rendering_references", |
| 120 | } |
| 121 | ) |
| 122 | _LEGACY_SPEC_LOCK_FORBIDDEN = frozenset({"Mixing icon libraries"}) |
| 123 | _SCAFFOLD_TOKEN_RE = re.compile(r"\{\{[A-Z_]+\}\}") |
| 124 | _SCHEMA_MARKER_RE = re.compile( |
| 125 | r"^<!--[ \t]+ppt-master-schema:[ \t]*([a-z0-9-]+/v[1-9][0-9]*)[ \t]+-->$", |
| 126 | re.IGNORECASE, |
| 127 | ) |
| 128 | |
| 129 | def _normalize_schema_value(value: str) -> str: |
| 130 | """Normalize a Markdown scalar before enum, pattern, and catalog checks.""" |
| 131 | normalized = value.strip() |
| 132 | if ( |
| 133 | len(normalized) >= 2 |
| 134 | and normalized[0] == normalized[-1] |
| 135 | and normalized[0] in "'\"`" |
| 136 | ): |
| 137 | return normalized[1:-1].strip() |
| 138 | return normalized |
| 139 | |
| 140 | |
| 141 | def _extract_schema_marker(text: str) -> tuple[str | None, str | None]: |
| 142 | """Read the optional version marker from the first non-empty line.""" |
| 143 | first_line = next((line.strip() for line in text.splitlines() if line.strip()), "") |
| 144 | if not first_line.startswith("<!--") or "ppt-master-schema:" not in first_line: |
| 145 | return None, None |
| 146 | match = _SCHEMA_MARKER_RE.fullmatch(first_line) |
| 147 | if match is None: |
| 148 | return None, "has a malformed ppt-master-schema marker" |
| 149 | return match.group(1).casefold(), None |
| 150 | |
| 151 | |
| 152 | def _parse_markdown_sections( |
| 153 | text: str, |
| 154 | *, |
| 155 | report_duplicate_fields: bool, |
| 156 | ) -> tuple[list[dict[str, object]], list[str]]: |
| 157 | """Parse H2 sections, data lines, and nested headings from Markdown.""" |
| 158 | headings = list(_MARKDOWN_H2_RE.finditer(text)) |
| 159 | sections: list[dict[str, object]] = [] |
| 160 | errors: list[str] = [] |
| 161 | |
| 162 | for index, heading_match in enumerate(headings): |
| 163 | body_start = heading_match.end() |
| 164 | body_end = headings[index + 1].start() if index + 1 < len(headings) else len(text) |
| 165 | body = text[body_start:body_end] |
| 166 | fields: dict[str, str] = {} |
| 167 | field_names: dict[str, str] = {} |
| 168 | |
| 169 | for field_match in _MARKDOWN_DATA_LINE_RE.finditer(body): |
| 170 | field_name = field_match.group(1).strip() |
| 171 | field_key = field_name |
| 172 | if field_key in fields and report_duplicate_fields: |
| 173 | errors.append( |
| 174 | f"section '{heading_match.group(1).strip()}' repeats data key " |
| 175 | f"'{field_name}'" |
| 176 | ) |
| 177 | continue |
| 178 | fields[field_key] = field_match.group(2).strip() |
| 179 | field_names[field_key] = field_name |
| 180 | |
| 181 | sections.append( |
| 182 | { |
| 183 | "heading": heading_match.group(1).strip(), |
| 184 | "offset": heading_match.start(), |
| 185 | "body": body, |
| 186 | "fields": fields, |
| 187 | "field_names": field_names, |
| 188 | "subheadings": [ |
| 189 | match.group(1).strip() |
| 190 | for match in _MARKDOWN_SUBHEADING_RE.finditer(body) |
| 191 | ], |
| 192 | } |
| 193 | ) |
| 194 | |
| 195 | return sections, errors |
| 196 | |
| 197 | |
| 198 | def parse_markdown_artifact( |
| 199 | markdown_path: Path, |
| 200 | *, |
| 201 | report_duplicate_fields: bool = False, |
| 202 | ) -> list[dict[str, object]]: |
| 203 | """Parse one Markdown planning artifact without changing it. |
| 204 | |
| 205 | This is the public read-only entry point for consumers that need the same |
| 206 | heading/data-line grammar as schema validation. Keeping the parser here |
| 207 | prevents runtime projections from drifting into their own lock grammar. |
| 208 | """ |
| 209 | text = markdown_path.read_text(encoding="utf-8") |
| 210 | sections, errors = _parse_markdown_sections( |
| 211 | text, |
| 212 | report_duplicate_fields=report_duplicate_fields, |
| 213 | ) |
| 214 | if errors: |
| 215 | raise ValueError("; ".join(errors)) |
| 216 | return sections |
| 217 | |
| 218 | |
| 219 | def _looks_like_image_path(raw: str) -> bool: |
| 220 | """Return whether one lock token looks like a project image path.""" |
| 221 | token = raw.strip().strip("`'\"").replace("\\", "/") |
| 222 | return bool(token) and Path(token).suffix.casefold() in _IMAGE_PATH_SUFFIXES |
| 223 | |
| 224 | |
| 225 | def parse_spec_lock_image_value(key: str, value: str) -> dict[str, str]: |
| 226 | """Parse one image-lock row while preserving supported legacy rows. |
| 227 | |
| 228 | Current rows use ``<path> | source=... | pattern=... | crop=...``. Legacy |
| 229 | rows remain readable, but any row that starts using named metadata must |
| 230 | provide the complete current contract. |
| 231 | """ |
| 232 | normalized_key = str(key).strip() |
| 233 | normalized_value = str(value).strip() |
| 234 | parts = [part.strip() for part in normalized_value.split("|")] |
| 235 | path_part = parts[0] if parts else "" |
| 236 | |
| 237 | if _looks_like_image_path(normalized_key) and not _looks_like_image_path(path_part): |
| 238 | parts.insert(0, normalized_key) |
| 239 | path_part = normalized_key |
| 240 | elif ( |
| 241 | len(parts) >= 2 |
| 242 | and parts[0].casefold() in _IMAGE_ACQUISITION_SOURCES |
| 243 | and _looks_like_image_path(parts[1]) |
| 244 | ): |
| 245 | path_part = parts[1] |
| 246 | |
| 247 | metadata_parts = [part for part in parts[1:] if "=" in part] |
| 248 | if not metadata_parts: |
| 249 | legacy_crop = ( |
| 250 | "no-crop" |
| 251 | if any( |
| 252 | re.search(r"(?<![a-z])no-crop(?![a-z])", part, re.IGNORECASE) |
| 253 | for part in parts[1:] |
| 254 | ) |
| 255 | else "" |
| 256 | ) |
| 257 | return { |
| 258 | "path": path_part, |
| 259 | "source": "", |
| 260 | "pattern": "", |
| 261 | "crop": legacy_crop, |
| 262 | "legacy": "true", |
| 263 | } |
| 264 | |
| 265 | metadata: dict[str, str] = {} |
| 266 | unsupported_parts: list[str] = [] |
| 267 | for part in parts[1:]: |
| 268 | field, separator, raw = part.partition("=") |
| 269 | if not separator: |
| 270 | unsupported_parts.append(part) |
| 271 | continue |
| 272 | field = field.strip().casefold() |
| 273 | if field in metadata: |
| 274 | raise ValueError(f"repeats metadata field {field!r}") |
| 275 | metadata[field] = raw.strip() |
| 276 | |
| 277 | expected_fields = {"source", "pattern", "crop"} |
| 278 | unknown_fields = sorted(set(metadata) - expected_fields) |
| 279 | missing_fields = sorted(expected_fields - set(metadata)) |
| 280 | if unsupported_parts: |
| 281 | shown = ", ".join(repr(part) for part in unsupported_parts) |
| 282 | raise ValueError(f"has unsupported metadata token(s) {shown}") |
| 283 | if unknown_fields: |
| 284 | raise ValueError( |
| 285 | f"has unknown metadata field(s) {', '.join(unknown_fields)}" |
| 286 | ) |
| 287 | if missing_fields: |
| 288 | raise ValueError( |
| 289 | f"misses metadata field(s) {', '.join(missing_fields)}" |
| 290 | ) |
| 291 | if not _looks_like_image_path(path_part): |
| 292 | raise ValueError(f"has invalid image path {path_part!r}") |
| 293 | |
| 294 | normalized_path = path_part.replace("\\", "/") |
| 295 | path = Path(normalized_path) |
| 296 | if ( |
| 297 | path.is_absolute() |
| 298 | or len(path.parts) != 2 |
| 299 | or path.parts[0] != "images" |
| 300 | or path.name in {"", ".", ".."} |
| 301 | or ":" in path.name |
| 302 | or normalized_path != f"images/{path.name}" |
| 303 | ): |
| 304 | raise ValueError( |
| 305 | f"must use canonical project path images/<filename>, got {path_part!r}" |
| 306 | ) |
| 307 | |
| 308 | source = metadata["source"].casefold() |
| 309 | if source not in _IMAGE_ACQUISITION_SOURCES: |
| 310 | allowed = ", ".join(sorted(_IMAGE_ACQUISITION_SOURCES)) |
| 311 | raise ValueError(f"source must be one of {allowed}, got {metadata['source']!r}") |
| 312 | if not metadata["pattern"]: |
| 313 | raise ValueError("pattern must be non-empty") |
| 314 | crop = metadata["crop"].casefold() |
| 315 | if crop not in _IMAGE_CROP_POLICIES: |
| 316 | allowed = ", ".join(sorted(_IMAGE_CROP_POLICIES)) |
| 317 | raise ValueError(f"crop must be one of {allowed}, got {metadata['crop']!r}") |
| 318 | |
| 319 | return { |
| 320 | "path": normalized_path, |
| 321 | "source": source, |
| 322 | "pattern": metadata["pattern"], |
| 323 | "crop": crop, |
| 324 | "legacy": "false", |
| 325 | } |
| 326 | |
| 327 | |
| 328 | def parse_spec_lock_artifact( |
| 329 | lock_path: Path, |
| 330 | *, |
| 331 | report_duplicate_fields: bool = False, |
| 332 | compatibility_warnings: list[str] | None = None, |
| 333 | ) -> list[dict[str, object]]: |
| 334 | """Parse one execution lock and normalize supported legacy image rows. |
| 335 | |
| 336 | New locks use ``- <key>: <path> | source=... | pattern=... | crop=...``. |
| 337 | Some versioned projects instead placed the image path before the colon. |
| 338 | Preserve those projects by projecting the key path back into the value so |
| 339 | every consumer sees the same path-first image value. |
| 340 | """ |
| 341 | sections = parse_markdown_artifact( |
| 342 | lock_path, |
| 343 | report_duplicate_fields=report_duplicate_fields, |
| 344 | ) |
| 345 | normalized_sections: list[dict[str, object]] = [] |
| 346 | compatibility_keys: list[str] = [] |
| 347 | |
| 348 | for section in sections: |
| 349 | if str(section.get("heading", "")).strip().casefold() != "images": |
| 350 | normalized_sections.append(section) |
| 351 | continue |
| 352 | raw_fields = section.get("fields") |
| 353 | if not isinstance(raw_fields, dict): |
| 354 | normalized_sections.append(section) |
| 355 | continue |
| 356 | |
| 357 | fields: dict[str, str] = {} |
| 358 | for raw_key, raw_value in raw_fields.items(): |
| 359 | key = str(raw_key) |
| 360 | value = str(raw_value).strip() |
| 361 | value_path = value.split("|", 1)[0].strip() |
| 362 | if _looks_like_image_path(key) and not _looks_like_image_path(value_path): |
| 363 | value = f"{key} | {value}" if value else key |
| 364 | compatibility_keys.append(key) |
| 365 | fields[key] = value |
| 366 | |
| 367 | normalized_section = dict(section) |
| 368 | normalized_section["fields"] = fields |
| 369 | normalized_sections.append(normalized_section) |
| 370 | |
| 371 | if compatibility_warnings is not None and compatibility_keys: |
| 372 | sample = ", ".join(compatibility_keys[:3]) |
| 373 | suffix = "" if len(compatibility_keys) <= 3 else ", ..." |
| 374 | compatibility_warnings.append( |
| 375 | f"{lock_path.name} images: normalized {len(compatibility_keys)} legacy " |
| 376 | "path-as-key row(s); new locks should use '- <key>: <path> | " |
| 377 | "source=... | pattern=... | crop=...' " |
| 378 | f"(found: {sample}{suffix})" |
| 379 | ) |
| 380 | return normalized_sections |
| 381 | |
| 382 | |
| 383 | def parse_spec_lock( |
| 384 | lock_path: Path, |
| 385 | *, |
| 386 | report_duplicate_fields: bool = False, |
| 387 | compatibility_warnings: list[str] | None = None, |
| 388 | ) -> dict[str, dict[str, str]]: |
| 389 | """Return one execution lock as ``{section: {key: value}}`.""" |
| 390 | sections = parse_spec_lock_artifact( |
| 391 | lock_path, |
| 392 | report_duplicate_fields=report_duplicate_fields, |
| 393 | compatibility_warnings=compatibility_warnings, |
| 394 | ) |
| 395 | parsed: dict[str, dict[str, str]] = {} |
| 396 | for section in sections: |
| 397 | raw_fields = section.get("fields") |
| 398 | if not isinstance(raw_fields, dict): |
| 399 | continue |
| 400 | parsed[str(section.get("heading", "")).strip()] = { |
| 401 | str(key): str(value) for key, value in raw_fields.items() |
| 402 | } |
| 403 | return parsed |
| 404 | |
| 405 | |
| 406 | def default_spec_lock_forbidden() -> frozenset[str]: |
| 407 | """Return the versioned scaffold's universal forbidden-item defaults.""" |
| 408 | sections = parse_markdown_artifact(SCAFFOLD_DIR / "spec_lock.md") |
| 409 | section = next( |
| 410 | ( |
| 411 | item |
| 412 | for item in sections |
| 413 | if str(item.get("heading", "")).strip().casefold() == "forbidden" |
| 414 | ), |
| 415 | None, |
| 416 | ) |
| 417 | if section is None: |
| 418 | raise ValueError("spec-lock scaffold has no forbidden section") |
| 419 | current = frozenset( |
| 420 | re.sub(r"^-[ \t]+", "", line.strip()) |
| 421 | for line in str(section.get("body", "")).splitlines() |
| 422 | if line.strip() |
| 423 | ) |
| 424 | return current | _LEGACY_SPEC_LOCK_FORBIDDEN |
| 425 | |
| 426 | |
| 427 | def _load_markdown_schema(schema_path: Path) -> dict[str, object]: |
| 428 | """Load and sanity-check one versioned Markdown schema.""" |
| 429 | with schema_path.open("r", encoding="utf-8") as stream: |
| 430 | schema = json.load(stream) |
| 431 | contract = schema.get("x-markdown") |
| 432 | if not isinstance(contract, dict): |
| 433 | raise ValueError(f"Schema is missing x-markdown: {schema_path}") |
| 434 | if contract.get("version") != 1: |
| 435 | raise ValueError(f"Unsupported Markdown schema version: {schema_path}") |
| 436 | return schema |
| 437 | |
| 438 | |
| 439 | def _catalog_values( |
| 440 | schema_path: Path, |
| 441 | value_catalog: Mapping[str, object], |
| 442 | ) -> tuple[Path, dict[str, object]]: |
| 443 | """Resolve a schema-declared JSON catalog and object pointer.""" |
| 444 | relative_path = value_catalog.get("path") |
| 445 | pointer = value_catalog.get("pointer", []) |
| 446 | if not isinstance(relative_path, str) or not isinstance(pointer, list): |
| 447 | raise ValueError("value_catalog requires string path and list pointer") |
| 448 | |
| 449 | catalog_path = (schema_path.parent / relative_path).resolve() |
| 450 | with catalog_path.open("r", encoding="utf-8") as stream: |
| 451 | node: object = json.load(stream) |
| 452 | for pointer_part in pointer: |
| 453 | if not isinstance(node, dict): |
| 454 | raise ValueError(f"pointer enters a non-object at '{pointer_part}'") |
| 455 | node = node[str(pointer_part)] |
| 456 | if not isinstance(node, dict): |
| 457 | raise ValueError("catalog pointer does not resolve to an object") |
| 458 | return catalog_path, node |
| 459 | |
| 460 | |
| 461 | def _validate_catalog_values( |
| 462 | *, |
| 463 | markdown_name: str, |
| 464 | section_id: str, |
| 465 | fields: Mapping[str, object], |
| 466 | schema_path: Path, |
| 467 | value_catalog: Mapping[str, object], |
| 468 | ) -> list[str]: |
| 469 | """Validate catalog membership and any schema-declared asset path.""" |
| 470 | errors: list[str] = [] |
| 471 | relative_path = value_catalog.get("path") |
| 472 | try: |
| 473 | catalog_path, catalog = _catalog_values(schema_path, value_catalog) |
| 474 | except (OSError, KeyError, UnicodeError, json.JSONDecodeError, ValueError) as exc: |
| 475 | return [ |
| 476 | f"{markdown_name} schema: cannot read value catalog " |
| 477 | f"'{relative_path}': {exc}" |
| 478 | ] |
| 479 | |
| 480 | asset_pattern = value_catalog.get("asset_path_pattern") |
| 481 | for field_value in fields.values(): |
| 482 | value = _normalize_schema_value(str(field_value)) |
| 483 | if value not in catalog: |
| 484 | errors.append( |
| 485 | f"{markdown_name} schema: section '{section_id}' value '{value}' " |
| 486 | f"is absent from {catalog_path.name}" |
| 487 | ) |
| 488 | continue |
| 489 | if not isinstance(asset_pattern, str): |
| 490 | continue |
| 491 | try: |
| 492 | relative_asset_path = asset_pattern.format(key=value, value=value) |
| 493 | except (KeyError, ValueError) as exc: |
| 494 | errors.append( |
| 495 | f"{markdown_name} schema: invalid asset_path_pattern " |
| 496 | f"'{asset_pattern}': {exc}" |
| 497 | ) |
| 498 | break |
| 499 | asset_path = (schema_path.parent / relative_asset_path).resolve() |
| 500 | if asset_path.suffix.casefold() != ".svg" or not asset_path.is_file(): |
| 501 | errors.append( |
| 502 | f"{markdown_name} schema: section '{section_id}' value '{value}' " |
| 503 | f"does not resolve to an SVG asset at {asset_path}" |
| 504 | ) |
| 505 | return errors |
| 506 | |
| 507 | |
| 508 | def _validate_section( |
| 509 | *, |
| 510 | markdown_name: str, |
| 511 | section_id: str, |
| 512 | section: Mapping[str, object], |
| 513 | definition: Mapping[str, object], |
| 514 | schema_path: Path, |
| 515 | ) -> list[str]: |
| 516 | """Apply one section definition to one matched Markdown section.""" |
| 517 | errors: list[str] = [] |
| 518 | fields = section["fields"] |
| 519 | assert isinstance(fields, dict) |
| 520 | |
| 521 | required_fields = definition.get("required_fields", []) |
| 522 | allow_empty = { |
| 523 | str(field_name) |
| 524 | for field_name in definition.get("allow_empty_fields", []) |
| 525 | } |
| 526 | if isinstance(required_fields, list): |
| 527 | for field_name in required_fields: |
| 528 | field_key = str(field_name) |
| 529 | if field_key not in fields: |
| 530 | errors.append( |
| 531 | f"{markdown_name} schema: section '{section_id}' is missing field " |
| 532 | f"'{field_name}'" |
| 533 | ) |
| 534 | elif ( |
| 535 | field_key not in allow_empty |
| 536 | and not _normalize_schema_value(str(fields[field_key])) |
| 537 | ): |
| 538 | errors.append( |
| 539 | f"{markdown_name} schema: section '{section_id}' field " |
| 540 | f"'{field_name}' must not be empty" |
| 541 | ) |
| 542 | |
| 543 | allowed_fields = definition.get("allowed_fields") |
| 544 | if isinstance(allowed_fields, list): |
| 545 | allowed = {str(field_name) for field_name in allowed_fields} |
| 546 | for field_name in fields: |
| 547 | if field_name not in allowed: |
| 548 | errors.append( |
| 549 | f"{markdown_name} schema: section '{section_id}' has unknown " |
| 550 | f"field '{field_name}'" |
| 551 | ) |
| 552 | |
| 553 | field_enums = definition.get("field_enums", {}) |
| 554 | if isinstance(field_enums, dict): |
| 555 | for field_name, allowed in field_enums.items(): |
| 556 | field_key = str(field_name) |
| 557 | if field_key not in fields or not isinstance(allowed, list): |
| 558 | continue |
| 559 | value = _normalize_schema_value(str(fields[field_key])) |
| 560 | if value not in [str(item) for item in allowed]: |
| 561 | errors.append( |
| 562 | f"{markdown_name} schema: section '{section_id}' field " |
| 563 | f"'{field_name}' has illegal value '{value}'" |
| 564 | ) |
| 565 | |
| 566 | field_patterns = definition.get("field_patterns", {}) |
| 567 | if isinstance(field_patterns, dict): |
| 568 | for field_name, pattern in field_patterns.items(): |
| 569 | field_key = str(field_name) |
| 570 | if field_key not in fields: |
| 571 | continue |
| 572 | value = _normalize_schema_value(str(fields[field_key])) |
| 573 | if re.fullmatch(str(pattern), value) is None: |
| 574 | errors.append( |
| 575 | f"{markdown_name} schema: section '{section_id}' field " |
| 576 | f"'{field_name}' does not match '{pattern}'" |
| 577 | ) |
| 578 | |
| 579 | field_value_rules = definition.get("field_value_rules", []) |
| 580 | if isinstance(field_value_rules, list): |
| 581 | for rule in field_value_rules: |
| 582 | if not isinstance(rule, dict): |
| 583 | continue |
| 584 | key_pattern = rule.get("key_pattern") |
| 585 | value_pattern = rule.get("value_pattern") |
| 586 | if not isinstance(key_pattern, str) or not isinstance( |
| 587 | value_pattern, str |
| 588 | ): |
| 589 | continue |
| 590 | requirement = str(rule.get("requirement", "match its value grammar")) |
| 591 | for field_name, raw_value in fields.items(): |
| 592 | if re.fullmatch(key_pattern, str(field_name)) is None: |
| 593 | continue |
| 594 | value = str(raw_value).strip() |
| 595 | if bool(rule.get("normalize", True)): |
| 596 | value = _normalize_schema_value(value) |
| 597 | value_matches = re.fullmatch(value_pattern, value) is not None |
| 598 | if value_matches and rule.get("numeric") == "positive_finite": |
| 599 | try: |
| 600 | number = float(value) |
| 601 | except ValueError: |
| 602 | value_matches = False |
| 603 | else: |
| 604 | value_matches = math.isfinite(number) and number > 0 |
| 605 | if not value_matches: |
| 606 | errors.append( |
| 607 | f"{markdown_name} schema: section '{section_id}' field " |
| 608 | f"'{field_name}' must {requirement}; found '{value}'" |
| 609 | ) |
| 610 | |
| 611 | minimum = definition.get("min_entries") |
| 612 | if isinstance(minimum, int) and len(fields) < minimum: |
| 613 | errors.append( |
| 614 | f"{markdown_name} schema: section '{section_id}' needs at least " |
| 615 | f"{minimum} data line(s)" |
| 616 | ) |
| 617 | |
| 618 | min_body_chars = definition.get("min_body_chars") |
| 619 | if isinstance(min_body_chars, int) and len(str(section["body"]).strip()) < min_body_chars: |
| 620 | errors.append( |
| 621 | f"{markdown_name} schema: section '{section_id}' must contain content" |
| 622 | ) |
| 623 | |
| 624 | entry_key_pattern = definition.get("entry_key_pattern") |
| 625 | if isinstance(entry_key_pattern, str): |
| 626 | field_names = section["field_names"] |
| 627 | assert isinstance(field_names, dict) |
| 628 | for field_name in field_names.values(): |
| 629 | if re.fullmatch(entry_key_pattern, str(field_name)) is None: |
| 630 | errors.append( |
| 631 | f"{markdown_name} schema: section '{section_id}' has malformed " |
| 632 | f"entry key '{field_name}'" |
| 633 | ) |
| 634 | |
| 635 | value_enum = definition.get("value_enum") |
| 636 | if isinstance(value_enum, list): |
| 637 | allowed_values = [str(item) for item in value_enum] |
| 638 | for field_value in fields.values(): |
| 639 | value = _normalize_schema_value(str(field_value)) |
| 640 | if value not in allowed_values: |
| 641 | errors.append( |
| 642 | f"{markdown_name} schema: section '{section_id}' has illegal " |
| 643 | f"value '{value}'" |
| 644 | ) |
| 645 | |
| 646 | value_pattern = definition.get("value_pattern") |
| 647 | if isinstance(value_pattern, str): |
| 648 | for field_value in fields.values(): |
| 649 | value = _normalize_schema_value(str(field_value)) |
| 650 | if re.fullmatch(value_pattern, value) is None: |
| 651 | errors.append( |
| 652 | f"{markdown_name} schema: section '{section_id}' has malformed " |
| 653 | f"value '{value}'" |
| 654 | ) |
| 655 | |
| 656 | value_catalog = definition.get("value_catalog") |
| 657 | if isinstance(value_catalog, dict): |
| 658 | errors.extend( |
| 659 | _validate_catalog_values( |
| 660 | markdown_name=markdown_name, |
| 661 | section_id=section_id, |
| 662 | fields=fields, |
| 663 | schema_path=schema_path, |
| 664 | value_catalog=value_catalog, |
| 665 | ) |
| 666 | ) |
| 667 | return errors |
| 668 | |
| 669 | |
| 670 | def _condition_applies( |
| 671 | when: Mapping[str, object], |
| 672 | matched: Mapping[str, dict[str, object] | None], |
| 673 | ) -> bool: |
| 674 | """Return whether a schema condition applies to the matched document.""" |
| 675 | section = matched.get(str(when.get("section", ""))) |
| 676 | if section is None: |
| 677 | return False |
| 678 | applies = True |
| 679 | field_name = when.get("field") |
| 680 | if field_name is not None: |
| 681 | fields = section["fields"] |
| 682 | assert isinstance(fields, dict) |
| 683 | field_value = fields.get(str(field_name)) |
| 684 | applies = field_value is not None |
| 685 | if applies and "equals" in when: |
| 686 | applies = _normalize_schema_value(str(field_value)) == str(when["equals"]) |
| 687 | body_regex = when.get("body_regex") |
| 688 | if applies and isinstance(body_regex, str): |
| 689 | applies = re.search(body_regex, str(section["body"])) is not None |
| 690 | return applies |
| 691 | |
| 692 | |
| 693 | def _validate_condition( |
| 694 | *, |
| 695 | markdown_name: str, |
| 696 | condition_id: str, |
| 697 | then: Mapping[str, object], |
| 698 | matched: Mapping[str, dict[str, object] | None], |
| 699 | ) -> list[str]: |
| 700 | """Apply one active cross-section condition.""" |
| 701 | errors: list[str] = [] |
| 702 | |
| 703 | required_sections = then.get("required_sections", []) |
| 704 | if isinstance(required_sections, list): |
| 705 | for section_id in required_sections: |
| 706 | if matched.get(str(section_id)) is None: |
| 707 | errors.append( |
| 708 | f"{markdown_name} schema: condition '{condition_id}' requires " |
| 709 | f"section '{section_id}'" |
| 710 | ) |
| 711 | |
| 712 | forbidden_sections = then.get("forbidden_sections", []) |
| 713 | if isinstance(forbidden_sections, list): |
| 714 | for section_id in forbidden_sections: |
| 715 | if matched.get(str(section_id)) is not None: |
| 716 | errors.append( |
| 717 | f"{markdown_name} schema: condition '{condition_id}' forbids " |
| 718 | f"section '{section_id}'" |
| 719 | ) |
| 720 | |
| 721 | field_groups = then.get("required_fields", []) |
| 722 | if isinstance(field_groups, list): |
| 723 | for group in field_groups: |
| 724 | if not isinstance(group, dict): |
| 725 | continue |
| 726 | target_id = str(group.get("section", "")) |
| 727 | target = matched.get(target_id) |
| 728 | if target is None: |
| 729 | continue |
| 730 | target_fields = target["fields"] |
| 731 | assert isinstance(target_fields, dict) |
| 732 | for field_name in group.get("fields", []): |
| 733 | field_key = str(field_name) |
| 734 | if field_key not in target_fields: |
| 735 | errors.append( |
| 736 | f"{markdown_name} schema: condition '{condition_id}' requires " |
| 737 | f"field '{field_name}' in section '{target_id}'" |
| 738 | ) |
| 739 | elif not _normalize_schema_value(str(target_fields[field_key])): |
| 740 | errors.append( |
| 741 | f"{markdown_name} schema: condition '{condition_id}' requires " |
| 742 | f"non-empty field '{field_name}' in section '{target_id}'" |
| 743 | ) |
| 744 | |
| 745 | field_values = then.get("field_values", []) |
| 746 | if isinstance(field_values, list): |
| 747 | for value_rule in field_values: |
| 748 | if not isinstance(value_rule, dict): |
| 749 | continue |
| 750 | target_id = str(value_rule.get("section", "")) |
| 751 | target = matched.get(target_id) |
| 752 | if target is None: |
| 753 | continue |
| 754 | target_fields = target["fields"] |
| 755 | assert isinstance(target_fields, dict) |
| 756 | target_field = str(value_rule.get("field", "")) |
| 757 | value = target_fields.get(target_field) |
| 758 | allowed = value_rule.get("enum", []) |
| 759 | if value is None or not isinstance(allowed, list): |
| 760 | continue |
| 761 | normalized = _normalize_schema_value(str(value)) |
| 762 | if normalized not in [str(item) for item in allowed]: |
| 763 | errors.append( |
| 764 | f"{markdown_name} schema: condition '{condition_id}' requires " |
| 765 | f"'{target_id}.{target_field}' to be one of {allowed}" |
| 766 | ) |
| 767 | |
| 768 | subheading_rules = then.get("required_subheadings", []) |
| 769 | if isinstance(subheading_rules, list): |
| 770 | for rule in subheading_rules: |
| 771 | if not isinstance(rule, dict): |
| 772 | continue |
| 773 | target_id = str(rule.get("section", "")) |
| 774 | target = matched.get(target_id) |
| 775 | if target is None: |
| 776 | continue |
| 777 | heading = str(rule.get("heading", "")) |
| 778 | subheadings = target["subheadings"] |
| 779 | assert isinstance(subheadings, list) |
| 780 | if not any( |
| 781 | str(item).startswith(heading) |
| 782 | for item in subheadings |
| 783 | ): |
| 784 | errors.append( |
| 785 | f"{markdown_name} schema: condition '{condition_id}' requires " |
| 786 | f"subheading '{heading}' in section '{target_id}'" |
| 787 | ) |
| 788 | return errors |
| 789 | |
| 790 | |
| 791 | def _validate_slides( |
| 792 | *, |
| 793 | markdown_name: str, |
| 794 | slide_contract: Mapping[str, object], |
| 795 | matched: Mapping[str, dict[str, object] | None], |
| 796 | ) -> list[str]: |
| 797 | """Validate repeated slide blocks inside the configured outline section.""" |
| 798 | outline = matched.get(str(slide_contract.get("section", ""))) |
| 799 | heading_pattern = str(slide_contract.get("heading_pattern", "")) |
| 800 | if outline is None or not heading_pattern: |
| 801 | return [] |
| 802 | |
| 803 | body = str(outline["body"]) |
| 804 | heading_matches = [ |
| 805 | match |
| 806 | for match in _MARKDOWN_SUBHEADING_RE.finditer(body) |
| 807 | if re.match(heading_pattern, match.group(1)) |
| 808 | ] |
| 809 | if not heading_matches: |
| 810 | return [f"{markdown_name} schema: content outline has no Slide blocks"] |
| 811 | |
| 812 | errors: list[str] = [] |
| 813 | required_fields = slide_contract.get("required_fields", []) |
| 814 | if not isinstance(required_fields, list): |
| 815 | return errors |
| 816 | for index, slide_match in enumerate(heading_matches): |
| 817 | block_end = ( |
| 818 | heading_matches[index + 1].start() |
| 819 | if index + 1 < len(heading_matches) |
| 820 | else len(body) |
| 821 | ) |
| 822 | block = body[slide_match.end():block_end] |
| 823 | for field_name in required_fields: |
| 824 | pattern = ( |
| 825 | rf"^[ \t]*-[ \t]+(?:\*\*)?{re.escape(str(field_name))}" |
| 826 | rf"(?:\*\*)?[ \t]*:" |
| 827 | ) |
| 828 | if re.search(pattern, block, flags=re.MULTILINE) is None: |
| 829 | errors.append( |
| 830 | f"{markdown_name} schema: '{slide_match.group(1)}' is missing " |
| 831 | f"field '{field_name}'" |
| 832 | ) |
| 833 | return errors |
| 834 | |
| 835 | |
| 836 | def _validate_references( |
| 837 | *, |
| 838 | markdown_path: Path, |
| 839 | markdown_name: str, |
| 840 | rules: object, |
| 841 | matched: Mapping[str, dict[str, object] | None], |
| 842 | ) -> list[str]: |
| 843 | """Validate schema-declared cross-section keys and project assets.""" |
| 844 | if not isinstance(rules, list): |
| 845 | return [] |
| 846 | errors: list[str] = [] |
| 847 | project_root = markdown_path.parent.resolve() |
| 848 | |
| 849 | for rule in rules: |
| 850 | if not isinstance(rule, dict): |
| 851 | continue |
| 852 | rule_id = str(rule.get("id", "reference")) |
| 853 | source_id = str(rule.get("from_section", "")) |
| 854 | source = matched.get(source_id) |
| 855 | if source is None: |
| 856 | continue |
| 857 | source_fields = source["fields"] |
| 858 | assert isinstance(source_fields, dict) |
| 859 | |
| 860 | target_id = rule.get("target_section") |
| 861 | target_fields: Mapping[str, object] | None = None |
| 862 | if isinstance(target_id, str): |
| 863 | target = matched.get(target_id) |
| 864 | if target is None: |
| 865 | continue |
| 866 | raw_target_fields = target["fields"] |
| 867 | assert isinstance(raw_target_fields, dict) |
| 868 | target_fields = raw_target_fields |
| 869 | |
| 870 | component = rule.get("value_component") |
| 871 | asset_pattern = rule.get("asset_path_pattern") |
| 872 | for source_key, raw_value in source_fields.items(): |
| 873 | value = _normalize_schema_value(str(raw_value)) |
| 874 | reference_value = value |
| 875 | if isinstance(component, dict): |
| 876 | separator = str(component.get("separator", "|")) |
| 877 | index = component.get("index", 0) |
| 878 | parts = [part.strip() for part in value.split(separator)] |
| 879 | if not isinstance(index, int) or index >= len(parts): |
| 880 | errors.append( |
| 881 | f"{markdown_name} schema: reference '{rule_id}' cannot " |
| 882 | f"parse value '{value}' from section '{source_id}'" |
| 883 | ) |
| 884 | continue |
| 885 | reference_value = _normalize_schema_value(parts[index]) |
| 886 | |
| 887 | if target_fields is not None: |
| 888 | if reference_value not in target_fields: |
| 889 | errors.append( |
| 890 | f"{markdown_name} schema: reference '{rule_id}' value " |
| 891 | f"'{reference_value}' from '{source_id}.{source_key}' is not " |
| 892 | f"declared in section '{target_id}'" |
| 893 | ) |
| 894 | |
| 895 | if isinstance(asset_pattern, str): |
| 896 | asset_value = reference_value |
| 897 | suffix_match = re.search(r"\{value\}(\.[A-Za-z0-9]+)$", asset_pattern) |
| 898 | if ( |
| 899 | suffix_match is not None |
| 900 | and asset_value.casefold().endswith( |
| 901 | suffix_match.group(1).casefold() |
| 902 | ) |
| 903 | ): |
| 904 | asset_value = asset_value[: -len(suffix_match.group(1))] |
| 905 | try: |
| 906 | relative_asset = asset_pattern.format(value=asset_value) |
| 907 | except (KeyError, ValueError) as exc: |
| 908 | errors.append( |
| 909 | f"{markdown_name} schema: reference '{rule_id}' has invalid " |
| 910 | f"asset_path_pattern '{asset_pattern}': {exc}" |
| 911 | ) |
| 912 | continue |
| 913 | asset_path = (project_root / relative_asset).resolve() |
| 914 | try: |
| 915 | asset_path.relative_to(project_root) |
| 916 | except ValueError: |
| 917 | errors.append( |
| 918 | f"{markdown_name} schema: reference '{rule_id}' escapes the " |
| 919 | f"project root for value '{reference_value}'" |
| 920 | ) |
| 921 | continue |
| 922 | if asset_path.suffix.casefold() != ".svg" or not asset_path.is_file(): |
| 923 | errors.append( |
| 924 | f"{markdown_name} schema: reference '{rule_id}' value " |
| 925 | f"'{reference_value}' does not resolve to {asset_path}" |
| 926 | ) |
| 927 | return errors |
| 928 | |
| 929 | |
| 930 | def _validate_strict_data_surface( |
| 931 | markdown_name: str, |
| 932 | text: str, |
| 933 | sections: list[dict[str, object]], |
| 934 | matched: Mapping[str, dict[str, object] | None], |
| 935 | ) -> list[str]: |
| 936 | """Reject unknown lock sections and prose outside the data-line grammar.""" |
| 937 | errors: list[str] = [] |
| 938 | section_ids = { |
| 939 | int(section["offset"]): section_id |
| 940 | for section_id, section in matched.items() |
| 941 | if section is not None |
| 942 | } |
| 943 | first_offset = min((int(section["offset"]) for section in sections), default=len(text)) |
| 944 | for line in text[:first_offset].splitlines(): |
| 945 | stripped = line.strip() |
| 946 | if not stripped or stripped.startswith("<!--") or stripped.startswith("# "): |
| 947 | continue |
| 948 | errors.append(f"{markdown_name} schema: unsupported preamble line '{stripped}'") |
| 949 | |
| 950 | for section in sections: |
| 951 | heading = str(section["heading"]) |
| 952 | section_id = section_ids.get(int(section["offset"])) |
| 953 | if section_id is None: |
| 954 | errors.append(f"{markdown_name} schema: unknown section '{heading}'") |
| 955 | continue |
| 956 | for line in str(section["body"]).splitlines(): |
| 957 | stripped = line.strip() |
| 958 | if not stripped: |
| 959 | continue |
| 960 | if section_id == "forbidden" and stripped.startswith("- "): |
| 961 | continue |
| 962 | if _MARKDOWN_DATA_LINE_RE.fullmatch(line) is not None: |
| 963 | continue |
| 964 | errors.append( |
| 965 | f"{markdown_name} schema: section '{section_id}' has unsupported " |
| 966 | f"line '{stripped}'" |
| 967 | ) |
| 968 | return errors |
| 969 | |
| 970 | |
| 971 | def _validate_spec_lock_relations( |
| 972 | markdown_path: Path, |
| 973 | matched: Mapping[str, dict[str, object] | None], |
| 974 | ) -> list[str]: |
| 975 | """Validate cross-section references that JSON field rules cannot express.""" |
| 976 | markdown_name = markdown_path.name |
| 977 | errors: list[str] = [] |
| 978 | |
| 979 | def fields(section_id: str) -> dict[str, str]: |
| 980 | section = matched.get(section_id) |
| 981 | if section is None: |
| 982 | return {} |
| 983 | raw_fields = section["fields"] |
| 984 | assert isinstance(raw_fields, dict) |
| 985 | return {str(key): str(value) for key, value in raw_fields.items()} |
| 986 | |
| 987 | for section_id, selector_field, references_field, catalog_dir in ( |
| 988 | _CUSTOM_REFERENCE_CATALOGS |
| 989 | ): |
| 990 | section_fields = fields(section_id) |
| 991 | is_custom = ( |
| 992 | _normalize_schema_value(section_fields.get(selector_field, "")) == "custom" |
| 993 | ) |
| 994 | raw_references = _normalize_schema_value( |
| 995 | section_fields.get(references_field, "") |
| 996 | ) |
| 997 | if not is_custom: |
| 998 | if raw_references: |
| 999 | errors.append( |
| 1000 | f"{markdown_name} schema: field '{references_field}' is valid " |
| 1001 | f"only when '{selector_field}' is custom" |
| 1002 | ) |
| 1003 | continue |
| 1004 | if not raw_references: |
| 1005 | continue |
| 1006 | references = [item.strip() for item in raw_references.split(",")] |
| 1007 | duplicates = sorted( |
| 1008 | reference |
| 1009 | for reference in set(references) |
| 1010 | if references.count(reference) > 1 |
| 1011 | ) |
| 1012 | if duplicates: |
| 1013 | errors.append( |
| 1014 | f"{markdown_name} schema: field '{references_field}' repeats " |
| 1015 | f"catalog id(s) {', '.join(duplicates)}" |
| 1016 | ) |
| 1017 | for reference in references: |
| 1018 | catalog_file = catalog_dir / f"{reference}.md" |
| 1019 | if reference == "custom" or not catalog_file.is_file(): |
| 1020 | errors.append( |
| 1021 | f"{markdown_name} schema: field '{references_field}' references " |
| 1022 | f"unknown catalog id '{reference}'" |
| 1023 | ) |
| 1024 | |
| 1025 | for key, value in fields("images").items(): |
| 1026 | if key.strip().casefold() in _LEGACY_IMAGE_METADATA_KEYS: |
| 1027 | continue |
| 1028 | try: |
| 1029 | parse_spec_lock_image_value(key, value) |
| 1030 | except ValueError as exc: |
| 1031 | errors.append( |
| 1032 | f"{markdown_name} schema: images row {key!r} {exc}" |
| 1033 | ) |
| 1034 | |
| 1035 | rhythm = fields("page_rhythm") |
| 1036 | layouts = fields("pptx_layouts") |
| 1037 | page_pptx_layouts = fields("page_pptx_layouts") |
| 1038 | page_layouts = fields("page_layouts") |
| 1039 | page_visualizations = fields("page_visualizations") |
| 1040 | legacy_page_charts = fields("page_charts") |
| 1041 | structure = fields("pptx_structure") |
| 1042 | |
| 1043 | for layout_key, raw_value in layouts.items(): |
| 1044 | parts = [part.strip() for part in raw_value.split("|")] |
| 1045 | if len(parts) != 3: |
| 1046 | continue |
| 1047 | _, _, source = parts |
| 1048 | if source.startswith("template:"): |
| 1049 | basename = source.removeprefix("template:").strip() |
| 1050 | if basename.casefold().endswith(".svg"): |
| 1051 | basename = basename[:-4] |
| 1052 | template_path = markdown_path.parent / "templates" / f"{basename}.svg" |
| 1053 | if not template_path.is_file(): |
| 1054 | errors.append( |
| 1055 | f"{markdown_name} schema: layout '{layout_key}' references " |
| 1056 | f"missing template SVG '{basename}.svg'" |
| 1057 | ) |
| 1058 | elif source not in rhythm: |
| 1059 | errors.append( |
| 1060 | f"{markdown_name} schema: layout '{layout_key}' has unknown " |
| 1061 | f"prototype source '{source}'" |
| 1062 | ) |
| 1063 | |
| 1064 | if structure.get("mode") == "structured": |
| 1065 | expected_pages = set(rhythm) |
| 1066 | for section_id, mapping in ( |
| 1067 | ("page_pptx_layouts", page_pptx_layouts), |
| 1068 | ("page_layouts", page_layouts), |
| 1069 | ): |
| 1070 | missing = sorted(expected_pages - set(mapping)) |
| 1071 | extra = sorted(set(mapping) - expected_pages) |
| 1072 | if missing: |
| 1073 | errors.append( |
| 1074 | f"{markdown_name} schema: section '{section_id}' misses pages " |
| 1075 | f"{', '.join(missing)}" |
| 1076 | ) |
| 1077 | if extra: |
| 1078 | errors.append( |
| 1079 | f"{markdown_name} schema: section '{section_id}' has unknown " |
| 1080 | f"pages {', '.join(extra)}" |
| 1081 | ) |
| 1082 | |
| 1083 | overlapping_visualization_pages = sorted( |
| 1084 | set(page_visualizations) & set(legacy_page_charts) |
| 1085 | ) |
| 1086 | if overlapping_visualization_pages: |
| 1087 | errors.append( |
| 1088 | f"{markdown_name} schema: pages " |
| 1089 | f"{', '.join(overlapping_visualization_pages)} are declared in both " |
| 1090 | "page_visualizations and legacy page_charts; keep only " |
| 1091 | "page_visualizations" |
| 1092 | ) |
| 1093 | |
| 1094 | for section_id, mapping, allow_legacy_bare in ( |
| 1095 | ("page_visualizations", page_visualizations, False), |
| 1096 | ("page_charts", legacy_page_charts, True), |
| 1097 | ): |
| 1098 | unknown_pages = sorted(set(mapping) - set(rhythm)) |
| 1099 | if unknown_pages: |
| 1100 | errors.append( |
| 1101 | f"{markdown_name} schema: {section_id} has unknown pages " |
| 1102 | f"{', '.join(unknown_pages)}" |
| 1103 | ) |
| 1104 | for page_key, raw_reference in mapping.items(): |
| 1105 | reference = _normalize_schema_value(raw_reference) |
| 1106 | try: |
| 1107 | entry = resolve_visualization_reference( |
| 1108 | reference, |
| 1109 | allow_legacy_bare=allow_legacy_bare, |
| 1110 | ) |
| 1111 | except VisualizationCatalogError as exc: |
| 1112 | errors.append( |
| 1113 | f"{markdown_name} schema: {section_id}.{page_key} " |
| 1114 | f"cannot resolve visualization {reference!r}: {exc}" |
| 1115 | ) |
| 1116 | continue |
| 1117 | if entry.kind == LEGACY_STRUCTURE_INTENT_KIND: |
| 1118 | if section_id != "page_charts" or not allow_legacy_bare: |
| 1119 | errors.append( |
| 1120 | f"{markdown_name} schema: {section_id}.{page_key} resolves " |
| 1121 | "to a legacy Structure intent outside page_charts" |
| 1122 | ) |
| 1123 | if entry.path is not None: |
| 1124 | errors.append( |
| 1125 | f"{markdown_name} schema: {section_id}.{page_key} legacy " |
| 1126 | "Structure intent unexpectedly has an asset path" |
| 1127 | ) |
| 1128 | continue |
| 1129 | if entry.kind != VISUALIZATION_SVG_KIND: |
| 1130 | errors.append( |
| 1131 | f"{markdown_name} schema: {section_id}.{page_key} resolves " |
| 1132 | f"to unsupported kind {entry.kind!r}" |
| 1133 | ) |
| 1134 | continue |
| 1135 | if entry.path is None: |
| 1136 | errors.append( |
| 1137 | f"{markdown_name} schema: {section_id}.{page_key} does not " |
| 1138 | "resolve to an SVG asset path" |
| 1139 | ) |
| 1140 | continue |
| 1141 | asset_path = Path(entry.path) |
| 1142 | if asset_path.suffix.casefold() != ".svg" or not asset_path.is_file(): |
| 1143 | errors.append( |
| 1144 | f"{markdown_name} schema: {section_id}.{page_key} does not " |
| 1145 | f"resolve to an SVG asset at {asset_path}" |
| 1146 | ) |
| 1147 | |
| 1148 | info = get_project_info_common(str(markdown_path.parent)) |
| 1149 | format_key = str(info.get("format", "unknown")) |
| 1150 | canvas = CANVAS_FORMATS.get(format_key) |
| 1151 | canvas_fields = fields("canvas") |
| 1152 | if canvas is not None: |
| 1153 | expected_format = str(canvas["name"]) |
| 1154 | expected_viewbox = str(canvas["viewbox"]) |
| 1155 | if ( |
| 1156 | "format" in canvas_fields |
| 1157 | and _normalize_schema_value(canvas_fields["format"]) != expected_format |
| 1158 | ): |
| 1159 | errors.append( |
| 1160 | f"{markdown_name} schema: canvas.format must be '{expected_format}'" |
| 1161 | ) |
| 1162 | if ( |
| 1163 | "viewBox" in canvas_fields |
| 1164 | and _normalize_schema_value(canvas_fields["viewBox"]) != expected_viewbox |
| 1165 | ): |
| 1166 | errors.append( |
| 1167 | f"{markdown_name} schema: canvas.viewBox must be '{expected_viewbox}'" |
| 1168 | ) |
| 1169 | return errors |
| 1170 | |
| 1171 | |
| 1172 | def validate_markdown_schema(markdown_path: Path, schema_path: Path) -> list[str]: |
| 1173 | """Validate one existing Markdown artifact against a versioned schema.""" |
| 1174 | try: |
| 1175 | text = markdown_path.read_text(encoding="utf-8-sig") |
| 1176 | schema = _load_markdown_schema(schema_path) |
| 1177 | except (OSError, UnicodeError, json.JSONDecodeError, ValueError) as exc: |
| 1178 | return [f"Schema validation could not read {markdown_path.name}: {exc}"] |
| 1179 | |
| 1180 | contract = schema["x-markdown"] |
| 1181 | assert isinstance(contract, dict) |
| 1182 | marker, marker_error = _extract_schema_marker(text) |
| 1183 | if marker_error is not None: |
| 1184 | return [f"{markdown_path.name} schema: {marker_error}"] |
| 1185 | expected_marker = contract.get("marker") |
| 1186 | if isinstance(expected_marker, str): |
| 1187 | if marker is None: |
| 1188 | return [ |
| 1189 | f"{markdown_path.name} schema: missing ppt-master-schema marker " |
| 1190 | f"'{expected_marker}'" |
| 1191 | ] |
| 1192 | if marker != expected_marker.casefold(): |
| 1193 | return [ |
| 1194 | f"{markdown_path.name} schema: marker '{marker}' does not match " |
| 1195 | f"'{expected_marker}'" |
| 1196 | ] |
| 1197 | sections, parse_errors = _parse_markdown_sections( |
| 1198 | text, |
| 1199 | report_duplicate_fields=contract.get("parser") == "heading-data-lines-v1", |
| 1200 | ) |
| 1201 | definitions = contract.get("sections", []) |
| 1202 | if not isinstance(definitions, list): |
| 1203 | return [f"Schema validation could not read {schema_path.name}: sections must be a list"] |
| 1204 | |
| 1205 | markdown_name = markdown_path.name |
| 1206 | errors = [f"{markdown_name} schema: {message}" for message in parse_errors] |
| 1207 | unresolved_patterns = contract.get("unresolved_patterns", []) |
| 1208 | if isinstance(unresolved_patterns, list): |
| 1209 | for pattern in unresolved_patterns: |
| 1210 | if not isinstance(pattern, str): |
| 1211 | continue |
| 1212 | matches = list(re.finditer(pattern, text)) |
| 1213 | if matches: |
| 1214 | errors.append( |
| 1215 | f"{markdown_name} schema: contains {len(matches)} unresolved " |
| 1216 | f"placeholder(s) matching '{pattern}'" |
| 1217 | ) |
| 1218 | |
| 1219 | matched: dict[str, dict[str, object] | None] = {} |
| 1220 | for definition in definitions: |
| 1221 | if not isinstance(definition, dict): |
| 1222 | continue |
| 1223 | section_id = str(definition.get("id", "")) |
| 1224 | pattern = str(definition.get("pattern", "")) |
| 1225 | if not section_id or not pattern: |
| 1226 | continue |
| 1227 | candidates = [ |
| 1228 | section |
| 1229 | for section in sections |
| 1230 | if re.fullmatch(pattern, str(section["heading"])) |
| 1231 | ] |
| 1232 | if len(candidates) > 1: |
| 1233 | errors.append( |
| 1234 | f"{markdown_name} schema: section '{section_id}' appears more than once" |
| 1235 | ) |
| 1236 | section = candidates[0] if candidates else None |
| 1237 | matched[section_id] = section |
| 1238 | if definition.get("required") is True and section is None: |
| 1239 | errors.append(f"{markdown_name} schema: missing section '{section_id}'") |
| 1240 | continue |
| 1241 | if section is not None: |
| 1242 | errors.extend( |
| 1243 | _validate_section( |
| 1244 | markdown_name=markdown_name, |
| 1245 | section_id=section_id, |
| 1246 | section=section, |
| 1247 | definition=definition, |
| 1248 | schema_path=schema_path, |
| 1249 | ) |
| 1250 | ) |
| 1251 | |
| 1252 | section_order = contract.get("section_order", []) |
| 1253 | if isinstance(section_order, list): |
| 1254 | ordered_sections = [ |
| 1255 | (str(section_id), matched.get(str(section_id))) |
| 1256 | for section_id in section_order |
| 1257 | if matched.get(str(section_id)) is not None |
| 1258 | ] |
| 1259 | offsets = [int(section["offset"]) for _, section in ordered_sections if section] |
| 1260 | if offsets != sorted(offsets): |
| 1261 | expected = " -> ".join(str(section_id) for section_id in section_order) |
| 1262 | errors.append( |
| 1263 | f"{markdown_name} schema: sections are out of order; expected {expected}" |
| 1264 | ) |
| 1265 | |
| 1266 | conditions = contract.get("conditions", []) |
| 1267 | if isinstance(conditions, list): |
| 1268 | for condition in conditions: |
| 1269 | if not isinstance(condition, dict): |
| 1270 | continue |
| 1271 | when = condition.get("when", {}) |
| 1272 | then = condition.get("then", {}) |
| 1273 | if not isinstance(when, dict) or not isinstance(then, dict): |
| 1274 | continue |
| 1275 | if _condition_applies(when, matched): |
| 1276 | errors.extend( |
| 1277 | _validate_condition( |
| 1278 | markdown_name=markdown_name, |
| 1279 | condition_id=str(condition.get("id", "conditional rule")), |
| 1280 | then=then, |
| 1281 | matched=matched, |
| 1282 | ) |
| 1283 | ) |
| 1284 | |
| 1285 | errors.extend( |
| 1286 | _validate_references( |
| 1287 | markdown_path=markdown_path, |
| 1288 | markdown_name=markdown_name, |
| 1289 | rules=contract.get("references"), |
| 1290 | matched=matched, |
| 1291 | ) |
| 1292 | ) |
| 1293 | |
| 1294 | slide_contract = contract.get("slides") |
| 1295 | if isinstance(slide_contract, dict): |
| 1296 | errors.extend( |
| 1297 | _validate_slides( |
| 1298 | markdown_name=markdown_name, |
| 1299 | slide_contract=slide_contract, |
| 1300 | matched=matched, |
| 1301 | ) |
| 1302 | ) |
| 1303 | if contract.get("strict_lines") is True: |
| 1304 | errors.extend( |
| 1305 | _validate_strict_data_surface(markdown_name, text, sections, matched) |
| 1306 | ) |
| 1307 | if schema.get("$id") == "ppt-master://schemas/spec-lock/v1": |
| 1308 | errors.extend(_validate_spec_lock_relations(markdown_path, matched)) |
| 1309 | return errors |
| 1310 | |
| 1311 | |
| 1312 | def validate_project_artifacts( |
| 1313 | project_path: Path, |
| 1314 | project_info: Mapping[str, object] | None = None, |
| 1315 | *, |
| 1316 | include_design: bool = True, |
| 1317 | ) -> tuple[list[str], list[str]]: |
| 1318 | """Validate the lock and, when requested, the human-facing design brief.""" |
| 1319 | info = project_info or get_project_info_common(str(project_path)) |
| 1320 | errors: list[str] = [] |
| 1321 | warnings: list[str] = [] |
| 1322 | artifacts: list[tuple[Path, Path, str]] = [] |
| 1323 | spec_name = info.get("spec_file") |
| 1324 | if include_design and isinstance(spec_name, str): |
| 1325 | artifacts.append( |
| 1326 | ( |
| 1327 | project_path / spec_name, |
| 1328 | SCHEMA_DIR / "design_spec.schema.json", |
| 1329 | "design", |
| 1330 | ) |
| 1331 | ) |
| 1332 | lock_path = project_path / "spec_lock.md" |
| 1333 | if lock_path.is_file(): |
| 1334 | artifacts.append((lock_path, SCHEMA_DIR / "spec_lock.schema.json", "lock")) |
| 1335 | elif isinstance(spec_name, str): |
| 1336 | errors.append( |
| 1337 | "Communication trace: missing spec_lock.md with a " |
| 1338 | "## communication section." |
| 1339 | ) |
| 1340 | |
| 1341 | legacy_design = False |
| 1342 | legacy_lock = False |
| 1343 | versioned_lock_valid = False |
| 1344 | for artifact_path, schema_path, artifact_kind in artifacts: |
| 1345 | try: |
| 1346 | text = artifact_path.read_text(encoding="utf-8-sig") |
| 1347 | except (OSError, UnicodeError) as exc: |
| 1348 | errors.append(f"Schema validation could not read {artifact_path.name}: {exc}") |
| 1349 | continue |
| 1350 | marker, marker_error = _extract_schema_marker(text) |
| 1351 | if marker_error is not None: |
| 1352 | errors.append(f"{artifact_path.name} schema: {marker_error}") |
| 1353 | continue |
| 1354 | if marker is None: |
| 1355 | warnings.append( |
| 1356 | f"{artifact_path.name}: legacy artifact has no ppt-master-schema " |
| 1357 | "marker; skipped versioned schema validation" |
| 1358 | ) |
| 1359 | legacy_design = legacy_design or artifact_kind == "design" |
| 1360 | legacy_lock = legacy_lock or artifact_kind == "lock" |
| 1361 | continue |
| 1362 | artifact_errors = validate_markdown_schema(artifact_path, schema_path) |
| 1363 | errors.extend(artifact_errors) |
| 1364 | if artifact_kind == "lock" and not artifact_errors: |
| 1365 | try: |
| 1366 | parse_spec_lock_artifact( |
| 1367 | artifact_path, |
| 1368 | compatibility_warnings=warnings, |
| 1369 | ) |
| 1370 | except (OSError, UnicodeError, ValueError) as exc: |
| 1371 | errors.append(f"spec_lock.md compatibility parse failed: {exc}") |
| 1372 | continue |
| 1373 | versioned_lock_valid = True |
| 1374 | if versioned_lock_valid: |
| 1375 | try: |
| 1376 | from svg_to_pptx.pptx_package.template_structure import ( |
| 1377 | TemplateStructureError, |
| 1378 | load_pptx_structure_lock, |
| 1379 | template_prototype_lock_errors, |
| 1380 | ) |
| 1381 | |
| 1382 | structure_lock = load_pptx_structure_lock(project_path) |
| 1383 | if structure_lock is not None: |
| 1384 | errors.extend(template_prototype_lock_errors(structure_lock)) |
| 1385 | except (ImportError, TemplateStructureError) as exc: |
| 1386 | errors.append(f"spec_lock.md structure preflight failed: {exc}") |
| 1387 | if legacy_design or legacy_lock: |
| 1388 | errors.extend( |
| 1389 | validate_communication_trace( |
| 1390 | project_path, |
| 1391 | check_lock=legacy_lock, |
| 1392 | check_design=legacy_design, |
| 1393 | ) |
| 1394 | ) |
| 1395 | return errors, warnings |
| 1396 | |
| 1397 | |
| 1398 | def scaffold_project_artifact(project_path: Path, artifact: str) -> str: |
| 1399 | """Render one versioned Markdown scaffold without overwriting user work.""" |
| 1400 | assets = { |
| 1401 | "design_spec": (SCAFFOLD_DIR / "design_spec.md", "design_spec.md"), |
| 1402 | "spec_lock": (SCAFFOLD_DIR / "spec_lock.md", "spec_lock.md"), |
| 1403 | } |
| 1404 | if artifact not in assets: |
| 1405 | raise ValueError(f"Unsupported scaffold artifact: {artifact}") |
| 1406 | if not project_path.exists() or not project_path.is_dir(): |
| 1407 | raise FileNotFoundError(f"Project directory does not exist: {project_path}") |
| 1408 | |
| 1409 | info = get_project_info_common(str(project_path)) |
| 1410 | format_key = str(info.get("format", "unknown")) |
| 1411 | if format_key not in CANVAS_FORMATS: |
| 1412 | raise ValueError( |
| 1413 | "Cannot derive the canvas format from the project directory name. " |
| 1414 | "Use a standard <name>_<format>_<YYYYMMDD> project path." |
| 1415 | ) |
| 1416 | canvas = CANVAS_FORMATS[format_key] |
| 1417 | created_date = str(info.get("date_formatted", "Unknown date")) |
| 1418 | if created_date == "Unknown date": |
| 1419 | created_date = "[fill]" |
| 1420 | context = { |
| 1421 | "PROJECT_NAME": str(info.get("name", project_path.name)), |
| 1422 | "CANVAS_NAME": str(canvas["name"]), |
| 1423 | "CANVAS_DIMENSIONS": str(canvas["dimensions"]), |
| 1424 | "VIEWBOX": str(canvas["viewbox"]), |
| 1425 | "CREATED_DATE": created_date, |
| 1426 | } |
| 1427 | |
| 1428 | scaffold_path, target_name = assets[artifact] |
| 1429 | target_path = project_path / target_name |
| 1430 | existing_spec = info.get("spec_file") if artifact == "design_spec" else None |
| 1431 | if isinstance(existing_spec, str): |
| 1432 | existing_path = project_path / existing_spec |
| 1433 | raise FileExistsError( |
| 1434 | f"Refusing to shadow existing design spec: {existing_path}" |
| 1435 | ) |
| 1436 | if target_path.exists() or target_path.is_symlink(): |
| 1437 | raise FileExistsError(f"Refusing to overwrite existing artifact: {target_path}") |
| 1438 | rendered = scaffold_path.read_text(encoding="utf-8") |
| 1439 | for key, value in context.items(): |
| 1440 | rendered = rendered.replace(f"{{{{{key}}}}}", value) |
| 1441 | unresolved = sorted(set(_SCAFFOLD_TOKEN_RE.findall(rendered))) |
| 1442 | if unresolved: |
| 1443 | raise ValueError( |
| 1444 | f"Unresolved scaffold token(s) in {scaffold_path}: {', '.join(unresolved)}" |
| 1445 | ) |
| 1446 | with target_path.open("x", encoding="utf-8") as stream: |
| 1447 | stream.write(rendered) |
| 1448 | return str(target_path) |
| 1449 |