@diagrammo/dgmo 0.47.0 → 0.48.0

package/src/visualizations/parse.ts

@@ -1,1560 +0,0 @@
-// ============================================================
-// Visualization parser — Story 109.2 (arch-review).
-//
-// The unified parse state machine, relocated verbatim out of d3.ts. Moving it
-// here breaks the d3 -> dgmo-router -> chart-type-registry -> d3 cycle that
-// Story 109.1 left behind: dgmo-router + chart-type-registry now import
-// parseVisualization from this module, not from d3.ts. Behavior is identical.
-// ============================================================
-
-import type { PaletteColors } from '../palettes';
-import { resolveColorWithDiagnostic } from '../colors';
-import { formatDgmoError, makeDgmoError, suggest } from '../diagnostics';
-import {
-  collectIndentedValues,
-  extractColor,
-  measureIndent,
-  normalizeNumericToken,
-  parseFirstLine,
-  peelTrailingColorName,
-  splitNameAndMeta,
-} from '../utils/parsing';
-import {
-  TIMELINE_REGISTRY,
-  withTagAliases,
-} from '../utils/reserved-key-registry';
-import {
-  matchTagBlockHeading,
-  stripDefaultModifier,
-  validateTagGroupNames,
-  validateTagValues,
-  tagAttrKey,
-} from '../utils/tag-groups';
-import type { TagGroup } from '../utils/tag-groups';
-import type { Writable } from '../utils/brand';
-import { parseTimelineEventLine } from '../timeline/parser';
-import {
-  DEFAULT_CLOUD_OPTIONS,
-  type ParsedVisualization,
-  type QuadrantLabel,
-  type VennOverlap,
-  type WordCloudWord,
-  type ParsedVizFull,
-} from './types';
-
-/**
- * Parses D3 chart text format into structured data. Returns the discriminated
- * {@link ParsedVisualization} union; internally the single state machine fills a
- * fat {@link ParsedVizFull} accumulator, which is a structural superset of every
- * variant, so the narrowing is sound and runtime-identical.
- */
-export function parseVisualization(
-  content: string,
-  palette?: PaletteColors
-): ParsedVisualization {
-  return parseVisualizationFull(content, palette) as ParsedVisualization;
-}
-
-function parseVisualizationFull(
-  content: string,
-  palette?: PaletteColors
-): ParsedVizFull {
-  const result: ParsedVizFull = {
-    type: null,
-    title: null,
-    titleLineNumber: null,
-    orientation: 'horizontal',
-    periods: [],
-    data: [],
-    words: [],
-    cloudOptions: { ...DEFAULT_CLOUD_OPTIONS },
-    links: [],
-    arcOrder: 'appearance',
-    arcNodeGroups: [],
-    timelineEvents: [],
-    timelineGroups: [],
-    timelineEras: [],
-    timelineMarkers: [],
-    timelineTagGroups: [],
-    timelineSort: null,
-    timelineScale: true,
-    timelineSwimlanes: false,
-    vennSets: [],
-    vennOverlaps: [],
-    quadrantLabels: {
-      topRight: null,
-      topLeft: null,
-      bottomLeft: null,
-      bottomRight: null,
-    },
-    quadrantPoints: [],
-    quadrantXAxis: null,
-    quadrantXAxisLineNumber: null,
-    quadrantYAxis: null,
-    quadrantYAxisLineNumber: null,
-    quadrantTitleLineNumber: null,
-    diagnostics: [],
-    error: null,
-  };
-
-  const fail = (line: number, message: string): ParsedVizFull => {
-    const diag = makeDgmoError(line, message);
-    result.diagnostics.push(diag);
-    result.error = formatDgmoError(diag);
-    return result;
-  };
-
-  const warn = (line: number, message: string): void => {
-    result.diagnostics.push(makeDgmoError(line, message, 'warning'));
-  };
-
-  if (!content?.trim()) {
-    return fail(0, 'Empty content');
-  }
-
-  const lines = content.split('\n');
-  const freeformLines: string[] = [];
-  let currentArcGroup: string | null = null;
-  let currentTimelineGroup: string | null = null;
-  let currentTimelineGroupMeta: Record<string, string> = {};
-  let currentTimelineTagGroup: Writable<TagGroup> | null = null;
-  let inTimelineEraBlock = false;
-  let timelineEraBlockIndent = 0;
-  let inTimelineMarkerBlock = false;
-  let timelineMarkerBlockIndent = 0;
-  let inSlopePeriodBlock = false;
-  const timelineAliasMap = new Map<string, string>();
-  const VALID_D3_TYPES = new Set([
-    'slope',
-    'wordcloud',
-    'arc',
-    'timeline',
-    'venn',
-    'quadrant',
-    'sequence',
-  ]);
-  let firstLineParsed = false;
-
-  for (let i = 0; i < lines.length; i++) {
-    // In-bounds by loop guard.
-    const rawLine = lines[i]!;
-    const line = rawLine.trim();
-    const indent = measureIndent(rawLine);
-    const lineNumber = i + 1;
-
-    // Skip empty lines
-    if (!line) continue;
-
-    // Skip comments
-    if (line.startsWith('//')) continue;
-
-    // First non-empty, non-comment line: chart type + optional title
-    if (!firstLineParsed) {
-      firstLineParsed = true;
-      const firstLineResult = parseFirstLine(line);
-      if (firstLineResult && VALID_D3_TYPES.has(firstLineResult.chartType)) {
-        result.type = firstLineResult.chartType as ParsedVisualization['type'];
-        if (firstLineResult.title) {
-          result.title = firstLineResult.title;
-          result.titleLineNumber = lineNumber;
-        }
-        continue;
-      }
-      // Not a bare chart type — fall through to normal parsing
-    }
-
-    // Timeline tag group heading: `tag Name [alias X]`
-    if (result.type === 'timeline' && indent === 0) {
-      const tagBlockMatch = matchTagBlockHeading(line);
-      if (tagBlockMatch) {
-        const newGroup: Writable<TagGroup> = {
-          name: tagBlockMatch.name,
-          ...(tagBlockMatch.alias !== undefined && {
-            alias: tagBlockMatch.alias,
-          }),
-          entries: [],
-          lineNumber,
-        };
-        currentTimelineTagGroup = newGroup;
-        if (tagBlockMatch.alias) {
-          timelineAliasMap.set(
-            tagBlockMatch.alias.toLowerCase(),
-            tagBlockMatch.name.toLowerCase()
-          );
-        } else {
-          timelineAliasMap.set(
-            tagBlockMatch.name.toLowerCase(),
-            tagBlockMatch.name.toLowerCase()
-          );
-        }
-        result.timelineTagGroups.push(newGroup);
-        continue;
-      }
-    }
-
-    // Timeline tag group entries (indented under tag heading)
-    if (currentTimelineTagGroup && indent > 0) {
-      const { text: entryText, isDefault } = stripDefaultModifier(line);
-      const { label, color } = extractColor(
-        entryText,
-        palette,
-        result.diagnostics,
-        lineNumber
-      );
-      if (color) {
-        if (isDefault) {
-          currentTimelineTagGroup.defaultValue = label;
-        } else if (currentTimelineTagGroup.entries.length === 0) {
-          currentTimelineTagGroup.defaultValue = label;
-        }
-        currentTimelineTagGroup.entries.push({
-          value: label,
-          color,
-          lineNumber,
-        });
-        continue;
-      }
-    }
-
-    // End tag group on non-indented line
-    if (currentTimelineTagGroup && indent === 0) {
-      currentTimelineTagGroup = null;
-    }
-
-    // [Group] container headers for arc / timeline (§1.5 trailing-token):
-    //   `[Group]`           — no color
-    //   `[Group] color`     — trailing-token color (recognized palette word)
-    //   `[Group] key: value`  — timeline group metadata
-    if (result.type === 'timeline') {
-      const tlGroupMatch = line.match(/^\[(.+?)\]\s*(.*?)\s*$/);
-      if (tlGroupMatch) {
-        const name = tlGroupMatch[1]!.trim();
-        const trailing = tlGroupMatch[2] ?? '';
-        let color: string | null = null;
-        let groupMeta: Record<string, string> = {};
-
-        if (trailing) {
-          const { label: peeled, colorName } = peelTrailingColorName(
-            ' ' + trailing
-          );
-          if (colorName && peeled.trim() === '') {
-            color =
-              resolveColorWithDiagnostic(
-                colorName,
-                lineNumber,
-                result.diagnostics,
-                palette
-              ) ?? null;
-          } else {
-            const tlRegistry = withTagAliases(
-              TIMELINE_REGISTRY,
-              new Set([
-                ...timelineAliasMap.keys(),
-                ...timelineAliasMap.values(),
-              ])
-            );
-            const split = splitNameAndMeta(
-              trailing,
-              tlRegistry,
-              timelineAliasMap
-            );
-            if (split.color) {
-              color =
-                resolveColorWithDiagnostic(
-                  split.color,
-                  lineNumber,
-                  result.diagnostics,
-                  palette
-                ) ?? null;
-            }
-            groupMeta = { ...split.meta };
-            if (groupMeta['color']) {
-              color =
-                resolveColorWithDiagnostic(
-                  groupMeta['color'],
-                  lineNumber,
-                  result.diagnostics,
-                  palette
-                ) ?? null;
-              delete groupMeta['color'];
-            }
-          }
-        }
-
-        result.timelineGroups.push({
-          name,
-          color,
-          metadata: groupMeta,
-          lineNumber,
-        });
-        currentTimelineGroup = name;
-        currentTimelineGroupMeta = groupMeta;
-        continue;
-      }
-    }
-
-    // [Group] for arc (original narrow regex)
-    const groupMatch = line.match(
-      /^\[(.+?)\](?:\s+(red|orange|yellow|green|blue|purple|teal|cyan|gray|black|white))?\s*$/
-    );
-    if (groupMatch) {
-      if (result.type === 'arc') {
-        const name = groupMatch[1]!.trim();
-        const color = groupMatch[2]
-          ? (resolveColorWithDiagnostic(
-              groupMatch[2].trim(),
-              lineNumber,
-              result.diagnostics,
-              palette
-            ) ?? null)
-          : null;
-        result.arcNodeGroups.push({ name, nodes: [], color, lineNumber });
-        currentArcGroup = name;
-      }
-      continue;
-    }
-
-    // Reject legacy ## group syntax
-    if (
-      /^#{2,}\s+/.test(line) &&
-      (result.type === 'arc' || result.type === 'timeline')
-    ) {
-      const name = line
-        .replace(/^#{2,}\s+/, '')
-        .replace(/\s*\([^)]*\)\s*$/, '')
-        .trim();
-      result.diagnostics.push(
-        makeDgmoError(
-          lineNumber,
-          `'## ${name}' is no longer supported. Use '[${name}]' instead`,
-          'warning'
-        )
-      );
-      continue;
-    }
-
-    // Clear group context on un-indented lines (except [Group] already handled above)
-    if (indent === 0) {
-      currentArcGroup = null;
-      currentTimelineGroup = null;
-      currentTimelineGroupMeta = {};
-    }
-
-    // Arc link line (§1.5 trailing-token):
-    //   `source -> target [color] [weight]` — color before weight
-    if (result.type === 'arc') {
-      const linkMatch = line.match(
-        /^(.+?)\s*->\s*(.+?)(?:\s+(red|orange|yellow|green|blue|purple|teal|cyan|gray|black|white))?(?:\s+(-?[\d,_]+(?:\.[\d]+)?))?$/
-      );
-      if (linkMatch) {
-        // Capture groups 1 and 2 are guaranteed by the regex match.
-        const source = linkMatch[1]!.trim();
-        const target = linkMatch[2]!.trim();
-        if (source.endsWith(':') || target.endsWith(':')) {
-          result.diagnostics.push(
-            makeDgmoError(
-              lineNumber,
-              `Trailing colon is not valid in arc edges — write '${source.replace(/:$/, '')} -> ${target.replace(/:$/, '')}' instead`
-            )
-          );
-          continue;
-        }
-        const linkColor = linkMatch[3]
-          ? (resolveColorWithDiagnostic(
-              linkMatch[3].trim(),
-              lineNumber,
-              result.diagnostics,
-              palette
-            ) ?? null)
-          : null;
-        const arcValue = linkMatch[4]
-          ? parseFloat(normalizeNumericToken(linkMatch[4]) ?? linkMatch[4])
-          : 1;
-        result.links.push({
-          source,
-          target,
-          value: arcValue,
-          color: linkColor,
-          lineNumber,
-        });
-        // Assign nodes to current group (first-appearance wins)
-        if (currentArcGroup !== null) {
-          const group = result.arcNodeGroups.find(
-            (g) => g.name === currentArcGroup
-          );
-          if (group) {
-            const allGrouped = new Set(
-              result.arcNodeGroups.flatMap((g) => g.nodes)
-            );
-            if (!allGrouped.has(source)) group.nodes.push(source);
-            if (!allGrouped.has(target)) group.nodes.push(target);
-          }
-        }
-        continue;
-      }
-    }
-
-    // Timeline era block entries (indented under bare `era`)
-    if (result.type === 'timeline' && inTimelineEraBlock) {
-      if (indent <= timelineEraBlockIndent) {
-        inTimelineEraBlock = false;
-        // fall through to process this line normally
-      } else {
-        if (line.startsWith('//')) continue;
-        // Timeline era block entry (\u00a71.5 trailing-token):
-        //   `<start> -> <end> Label`           (no color)
-        //   `<start> -> <end> Label color`     (trailing color word)
-        // Color (group 4) must be a recognized lowercase palette word.
-        const eraEntryMatch = line.match(
-          /^(\d{4}(?:-\d{2})?(?:-\d{2}(?: \d{2}:\d{2})?)?)\s*(?:->|\u2013>)\s*(\d{4}(?:-\d{2})?(?:-\d{2}(?: \d{2}:\d{2})?)?)\s+(.+?)(?:\s+(red|orange|yellow|green|blue|purple|teal|cyan|gray|black|white))?\s*$/
-        );
-        if (eraEntryMatch) {
-          const colorAnnotation = eraEntryMatch[4]?.trim() || null;
-          result.timelineEras.push({
-            // Capture groups 1-3 guaranteed by the regex match.
-            startDate: eraEntryMatch[1]!,
-            endDate: eraEntryMatch[2]!,
-            label: eraEntryMatch[3]!.trim(),
-            color: colorAnnotation
-              ? (resolveColorWithDiagnostic(
-                  colorAnnotation,
-                  lineNumber,
-                  result.diagnostics,
-                  palette
-                ) ?? null)
-              : null,
-            lineNumber,
-          });
-        } else {
-          warn(
-            lineNumber,
-            `Unrecognized era entry: "${line}" — use '<start> -> <end> Label [color]' (e.g. '1716 -> 1717 Reign').`
-          );
-        }
-        continue;
-      }
-    }
-
-    // Timeline marker block entries (indented under bare `marker`)
-    if (result.type === 'timeline' && inTimelineMarkerBlock) {
-      if (indent <= timelineMarkerBlockIndent) {
-        inTimelineMarkerBlock = false;
-        // fall through to process this line normally
-      } else {
-        if (line.startsWith('//')) continue;
-        // Timeline marker block entry (§1.5 trailing-token).
-        const markerEntryMatch = line.match(
-          /^(\d{4}(?:-\d{2})?(?:-\d{2}(?: \d{2}:\d{2})?)?)\s+(.+?)(?:\s+(red|orange|yellow|green|blue|purple|teal|cyan|gray|black|white))?\s*$/
-        );
-        if (markerEntryMatch) {
-          const colorAnnotation = markerEntryMatch[3]?.trim() || null;
-          result.timelineMarkers.push({
-            // Capture groups 1-2 guaranteed by the regex match.
-            date: markerEntryMatch[1]!,
-            label: markerEntryMatch[2]!.trim(),
-            color: colorAnnotation
-              ? (resolveColorWithDiagnostic(
-                  colorAnnotation,
-                  lineNumber,
-                  result.diagnostics,
-                  palette
-                ) ?? null)
-              : null,
-            lineNumber,
-          });
-        } else {
-          warn(
-            lineNumber,
-            `Unrecognized marker entry: "${line}" — use '<date> Label [color]' (e.g. '1776 Independence').`
-          );
-        }
-        continue;
-      }
-    }
-
-    // Timeline era/marker block starters and inline forms
-    if (result.type === 'timeline') {
-      // Bare `era` keyword starts a block
-      if (line.toLowerCase() === 'era') {
-        inTimelineEraBlock = true;
-        timelineEraBlockIndent = indent;
-        continue;
-      }
-
-      // Bare `marker` keyword starts a block
-      if (line.toLowerCase() === 'marker') {
-        inTimelineMarkerBlock = true;
-        timelineMarkerBlockIndent = indent;
-        continue;
-      }
-
-      // Timeline era lines, inline (\u00a71.5 trailing-token):
-      //   `era YYYY->YYYY Label`        \u2014 no color
-      //   `era YYYY->YYYY Label color`  \u2014 trailing-token color (recognized palette word)
-      const eraMatch = line.match(
-        /^era\s+(\d{4}(?:-\d{2})?(?:-\d{2}(?: \d{2}:\d{2})?)?)\s*(?:->|\u2013>)\s*(\d{4}(?:-\d{2})?(?:-\d{2}(?: \d{2}:\d{2})?)?)\s+(.+?)(?:\s+(red|orange|yellow|green|blue|purple|teal|cyan|gray|black|white))?\s*$/
-      );
-      if (eraMatch) {
-        const colorAnnotation = eraMatch[4]?.trim() || null;
-        result.timelineEras.push({
-          // Capture groups 1-3 guaranteed by the regex match.
-          startDate: eraMatch[1]!,
-          endDate: eraMatch[2]!,
-          label: eraMatch[3]!.trim(),
-          color: colorAnnotation
-            ? (resolveColorWithDiagnostic(
-                colorAnnotation,
-                lineNumber,
-                result.diagnostics,
-                palette
-              ) ?? null)
-            : null,
-          lineNumber,
-        });
-        continue;
-      }
-
-      // Timeline marker lines, inline (§1.5 trailing-token):
-      //   `marker YYYY Label`        — no color
-      //   `marker YYYY Label color`  — trailing-token color
-      const markerMatch = line.match(
-        /^marker\s+(\d{4}(?:-\d{2})?(?:-\d{2}(?: \d{2}:\d{2})?)?)\s+(.+?)(?:\s+(red|orange|yellow|green|blue|purple|teal|cyan|gray|black|white))?\s*$/
-      );
-      if (markerMatch) {
-        const colorAnnotation = markerMatch[3]?.trim() || null;
-        result.timelineMarkers.push({
-          // Capture groups 1-2 guaranteed by the regex match.
-          date: markerMatch[1]!,
-          label: markerMatch[2]!.trim(),
-          color: colorAnnotation
-            ? (resolveColorWithDiagnostic(
-                colorAnnotation,
-                lineNumber,
-                result.diagnostics,
-                palette
-              ) ?? null)
-            : null,
-          lineNumber,
-        });
-        continue;
-      }
-    }
-
-    // Timeline sort directive: `sort time | group | tag:X`
-    if (result.type === 'timeline') {
-      const sortMatch = line.match(/^sort\s+(time|group|tag:(.+))$/i);
-      if (sortMatch) {
-        const kind = sortMatch[1]!.toLowerCase();
-        if (kind === 'time') result.timelineSort = 'time';
-        else if (kind === 'group') result.timelineSort = 'group';
-        else {
-          result.timelineSort = 'tag';
-          result.timelineDefaultSwimlaneTG = sortMatch[2]!.trim();
-        }
-        continue;
-      }
-    }
-
-    // Timeline display directives (§15.6, space-separated, no colon):
-    //   no-scale         — drop the date axis/scale
-    //   swimlanes        — force lane backgrounds
-    //   active-tag <X>   — authored active tag group / `none` / value-ramp label
-    if (result.type === 'timeline') {
-      const lower = line.toLowerCase();
-      if (lower === 'no-scale') {
-        result.timelineScale = false;
-        continue;
-      }
-      if (lower === 'swimlanes') {
-        result.timelineSwimlanes = true;
-        continue;
-      }
-      // `lane-by <group>`: the canonical swimlane-axis directive (view-state).
-      // Equivalent to `sort tag:<group>`. `swimlane` is a chart type, hence the
-      // `lane-by` keyword.
-      const laneByMatch = line.match(/^lane-by\s+(.+)$/i);
-      if (laneByMatch) {
-        result.timelineSort = 'tag';
-        result.timelineDefaultSwimlaneTG = laneByMatch[1]!.trim();
-        continue;
-      }
-      const activeTagMatch = line.match(/^active-tag\s+(.+)$/i);
-      if (activeTagMatch) {
-        result.timelineActiveTag = activeTagMatch[1]!.trim();
-        continue;
-      }
-    }
-
-    // Timeline event lines: date-first syntax
-    if (result.type === 'timeline') {
-      const tlRegistry = withTagAliases(
-        TIMELINE_REGISTRY,
-        new Set([...timelineAliasMap.keys(), ...timelineAliasMap.values()])
-      );
-
-      const eventResult = parseTimelineEventLine(
-        line,
-        lineNumber,
-        currentTimelineGroup,
-        currentTimelineGroupMeta,
-        timelineAliasMap,
-        tlRegistry
-      );
-
-      if (eventResult) {
-        result.diagnostics.push(...eventResult.diagnostics);
-        if (eventResult.event) {
-          result.timelineEvents.push(eventResult.event);
-        }
-        continue;
-      }
-
-      // Bare name with no date prefix (not a keyword)
-      if (
-        !/^(era|marker|tag|sort|active-tag|swimlanes|no-scale|lane-by)\b/i.test(
-          line
-        ) &&
-        !line.startsWith('[')
-      ) {
-        warn(lineNumber, `Expected a date at the start — e.g., '1718 ${line}'`);
-        continue;
-      }
-    }
-
-    // Venn diagram DSL
-    if (result.type === 'venn') {
-      // Skip cross-chart bare-keyword options so they don't get parsed as
-      // a 4th set name (the bare-keyword block at line ~1132 runs AFTER
-      // type-specific parsing).
-      if (/^(solid-fill|no-name|no-value|no-percent|no-title)$/i.test(line)) {
-        // Fall through to the bare-keyword block below.
-      } else if (/\+/.test(line)) {
-        // Build lookup of known set names and aliases for label extraction
-        const knownSetRefs = new Set<string>();
-        for (const s of result.vennSets) {
-          knownSetRefs.add(s.name.toLowerCase());
-          if (s.alias) knownSetRefs.add(s.alias.toLowerCase());
-        }
-
-        const segments = line
-          .split('+')
-          .map((s) => s.trim())
-          .filter(Boolean);
-        if (segments.length >= 2) {
-          // All segments except the last are pure set references
-          const rawSets = segments.slice(0, -1);
-          // In-bounds by length check (segments.length >= 2).
-          const lastSeg = segments[segments.length - 1]!;
-
-          // For the last segment, extract set reference and optional label.
-          // Find where the set reference ends and label begins.
-          // Try progressively shorter prefixes against known set names/aliases.
-          const words = lastSeg.split(/\s+/);
-          let matchLen = 0;
-          for (let w = words.length; w >= 1; w--) {
-            const candidate = words.slice(0, w).join(' ');
-            if (knownSetRefs.has(candidate.toLowerCase())) {
-              matchLen = w;
-              break;
-            }
-          }
-          let lastSetRef: string;
-          let label: string | null;
-          if (matchLen > 0) {
-            lastSetRef = words.slice(0, matchLen).join(' ');
-            label =
-              words.length > matchLen ? words.slice(matchLen).join(' ') : null;
-          } else {
-            // No known set matched — assume first word is the set ref, rest is label
-            // words is non-empty (split always returns at least one).
-            lastSetRef = words[0]!;
-            label = words.length > 1 ? words.slice(1).join(' ') : null;
-          }
-          rawSets.push(lastSetRef);
-          result.vennOverlaps.push({ sets: rawSets, label, lineNumber });
-          continue;
-        }
-      }
-
-      // Set declaration (§1.5 universal trailing-token, §2A.2 modifier order):
-      //   `Name` / `Name color` / `Name as <alias>` / `Name as <alias> color`
-      // Color is the line-trailing token, peeled first; alias follows the name
-      // via the `as` keyword. Only attempt set parsing if the line wasn't a
-      // bare-keyword option (handled above).
-      if (!/^(solid-fill|no-name|no-value|no-percent|no-title)$/i.test(line)) {
-        // Peel a trailing color word from the whole line first so the
-        // remaining text is `Name [alias <alias>]` / `Name [as <alias>]`.
-        const { label: lineWithoutColor, colorName } =
-          peelTrailingColorName(line);
-        let color: string | null = null;
-        if (colorName) {
-          color =
-            resolveColorWithDiagnostic(
-              colorName,
-              lineNumber,
-              result.diagnostics,
-              palette
-            ) ?? null;
-        }
-
-        const setDeclMatch = lineWithoutColor.match(
-          /^(.+?)(?:\s+as\s+([A-Za-z][A-Za-z0-9_]{0,11}))?\s*$/i
-        );
-        if (setDeclMatch) {
-          // Capture group 1 guaranteed by the regex match.
-          const name = setDeclMatch[1]!.trim();
-          const alias = setDeclMatch[2]?.trim() ?? null;
-          result.vennSets.push({ name, alias, color, lineNumber });
-          continue;
-        }
-      }
-    }
-
-    // Quadrant-specific parsing
-    if (result.type === 'quadrant') {
-      // x-label Low, High  — or indented multi-line
-      const xAxisMatch = line.match(/^x-label\s+(.*)/i);
-      if (xAxisMatch) {
-        // Capture group 1 guaranteed by the regex match.
-        const val = xAxisMatch[1]!.trim();
-        let parts: string[];
-        if (val) {
-          parts = val.split(',').map((s) => s.trim());
-        } else {
-          const collected = collectIndentedValues(lines, i);
-          i = collected.newIndex;
-          parts = collected.values;
-        }
-        if (parts.length >= 2) {
-          // In-bounds by length check above.
-          result.quadrantXAxis = [parts[0]!, parts[1]!];
-          result.quadrantXAxisLineNumber = lineNumber;
-        }
-        continue;
-      }
-
-      // y-label Low, High  — or indented multi-line
-      const yAxisMatch = line.match(/^y-label\s+(.*)/i);
-      if (yAxisMatch) {
-        // Capture group 1 guaranteed by the regex match.
-        const val = yAxisMatch[1]!.trim();
-        let parts: string[];
-        if (val) {
-          parts = val.split(',').map((s) => s.trim());
-        } else {
-          const collected = collectIndentedValues(lines, i);
-          i = collected.newIndex;
-          parts = collected.values;
-        }
-        if (parts.length >= 2) {
-          // In-bounds by length check above.
-          result.quadrantYAxis = [parts[0]!, parts[1]!];
-          result.quadrantYAxisLineNumber = lineNumber;
-        }
-        continue;
-      }
-
-      // Quadrant position labels (§1.5 trailing-token):
-      //   `top-right Label`        — no color
-      //   `top-right Label color`  — trailing-token color (recognized palette word)
-      const quadrantLabelRe =
-        /^(top-right|top-left|bottom-left|bottom-right)\s+(.+)/i;
-      const quadrantMatch = line.match(quadrantLabelRe);
-      if (quadrantMatch) {
-        // Capture groups 1-2 guaranteed by the regex match.
-        const position = quadrantMatch[1]!.toLowerCase();
-        const labelPart = quadrantMatch[2]!.trim();
-        // Peel trailing recognized color word from the label.
-        const { label: text, colorName } = peelTrailingColorName(labelPart);
-        const color = colorName
-          ? (resolveColorWithDiagnostic(
-              colorName,
-              lineNumber,
-              result.diagnostics,
-              palette
-            ) ?? null)
-          : null;
-        const label: QuadrantLabel = { text, color, lineNumber };
-
-        if (position === 'top-right') result.quadrantLabels.topRight = label;
-        else if (position === 'top-left') result.quadrantLabels.topLeft = label;
-        else if (position === 'bottom-left')
-          result.quadrantLabels.bottomLeft = label;
-        else if (position === 'bottom-right')
-          result.quadrantLabels.bottomRight = label;
-        continue;
-      }
-
-      // Data points: Label x, y  OR  Label x y
-      const pointMatch = line.match(
-        /^(.+?)\s+(-?[0-9][0-9,_]*(?:\.[0-9]+)?)\s*([,\s])\s*(-?[0-9][0-9,_]*(?:\.[0-9]+)?)\s*$/
-      );
-      if (pointMatch) {
-        // Capture groups 1-4 guaranteed by the regex match.
-        const label = pointMatch[1]!.trim();
-        // Skip if it looks like a quadrant position keyword
-        const lowerLabel = label.toLowerCase();
-        if (
-          lowerLabel !== 'top-right' &&
-          lowerLabel !== 'top-left' &&
-          lowerLabel !== 'bottom-left' &&
-          lowerLabel !== 'bottom-right'
-        ) {
-          const x = parseFloat(
-            normalizeNumericToken(pointMatch[2]!) ?? pointMatch[2]!
-          );
-          const y = parseFloat(
-            normalizeNumericToken(pointMatch[4]!) ?? pointMatch[4]!
-          );
-          result.quadrantPoints.push({
-            label,
-            x,
-            y,
-            lineNumber,
-          });
-        }
-        continue;
-      }
-    }
-
-    // ── Space-separated options (no colon) ──────────────────
-    const spaceIdx = line.indexOf(' ');
-    if (spaceIdx >= 0) {
-      const firstToken = line.substring(0, spaceIdx).toLowerCase();
-      const restValue = line.substring(spaceIdx + 1).trim();
-
-      if (
-        firstToken === 'chart' &&
-        VALID_D3_TYPES.has(restValue.toLowerCase())
-      ) {
-        result.type = restValue.toLowerCase() as ParsedVisualization['type'];
-        continue;
-      }
-
-      if (firstToken === 'title') {
-        result.title = restValue;
-        result.titleLineNumber = lineNumber;
-        if (result.type === 'quadrant') {
-          result.quadrantTitleLineNumber = lineNumber;
-        }
-        continue;
-      }
-
-      if (firstToken === 'order') {
-        const v = restValue.toLowerCase();
-        if (v === 'name' || v === 'group' || v === 'degree') {
-          result.arcOrder = v;
-        }
-        continue;
-      }
-
-      // `layout arc|chord` (#26): render the same pairwise edges as the other
-      // preset. `chord` → circular chord; default `arc` → linear.
-      if (firstToken === 'layout') {
-        const v = restValue.toLowerCase();
-        if (v === 'arc' || v === 'chord') result.layout = v;
-        continue;
-      }
-
-      if (firstToken === 'rotate') {
-        const v = restValue.toLowerCase();
-        if (v === 'none' || v === 'mixed' || v === 'angled') {
-          result.cloudOptions.rotate = v;
-        }
-        continue;
-      }
-
-      if (firstToken === 'max') {
-        const v = parseInt(restValue, 10);
-        if (!isNaN(v) && v > 0) {
-          result.cloudOptions.max = v;
-        }
-        continue;
-      }
-
-      if (firstToken === 'size') {
-        const parts = restValue.split(',').map((s) => parseInt(s.trim(), 10));
-        if (
-          parts.length === 2 &&
-          parts.every((n) => !isNaN(n) && n > 0) &&
-          // In-bounds by length === 2 check.
-          parts[0]! < parts[1]!
-        ) {
-          // In-bounds by length === 2 check.
-          result.cloudOptions.minSize = parts[0]!;
-          result.cloudOptions.maxSize = parts[1]!;
-        }
-        continue;
-      }
-    }
-
-    // ── Bare-keyword no-* flags (show-everything default) ──────
-    {
-      const bareToken = line.toLowerCase();
-      if (bareToken === 'no-name') {
-        result.noName = true;
-        continue;
-      }
-      if (bareToken === 'no-value') {
-        result.noValue = true;
-        continue;
-      }
-      if (bareToken === 'no-percent') {
-        result.noPercent = true;
-        continue;
-      }
-      if (bareToken === 'solid-fill') {
-        result.solidFill = true;
-        continue;
-      }
-      if (bareToken === 'no-title') {
-        result.noTitle = true;
-        continue;
-      }
-      // Silent-ignore unrecognized no-* flags (typos, future flags).
-      if (bareToken.startsWith('no-')) {
-        continue;
-      }
-    }
-
-    // ── Slope chart: period directive + right-scan data rows ──
-    if (result.type === 'slope') {
-      // Period block: indented lines inside `period` block
-      // (blank lines are pre-filtered at loop top, so only non-indented lines close the block)
-      if (inSlopePeriodBlock) {
-        if (indent > 0) {
-          result.periods.push(line);
-          continue;
-        }
-        // Non-indented line → close block, fall through to process normally
-        inSlopePeriodBlock = false;
-      }
-
-      // Period directive: `period Label1 Label2` or bare `period` (block open)
-      // Only accept before data rows start (F4: prevent keyword shadowing labels)
-      if (result.data.length === 0) {
-        const periodMatch = line.match(/^period\b(.*)$/i);
-        if (periodMatch) {
-          if (result.periods.length > 0 && !inSlopePeriodBlock) {
-            // F5: warn on duplicate period directives
-            warn(
-              lineNumber,
-              `Duplicate 'period' directive — periods are already defined`
-            );
-          }
-          // Capture group 1 guaranteed by the regex match.
-          const rest = periodMatch[1]!.trim();
-          if (rest) {
-            // One-line: `period 1715 1725`
-            const periodLabels = rest.split(/\s+/);
-            result.periods.push(...periodLabels);
-          } else {
-            // Block open: bare `period`
-            inSlopePeriodBlock = true;
-          }
-          continue;
-        }
-      }
-
-      // Migration error: bare period line (old syntax — comma-separated, no keyword)
-      // F1: Only fire when ALL comma-separated tokens are short (≤20 chars) and non-empty
-      if (
-        result.periods.length === 0 &&
-        line.includes(',') &&
-        !line.includes(':')
-      ) {
-        const tokens = line
-          .split(',')
-          .map((t) => t.trim())
-          .filter(Boolean);
-        const looksLikePeriods =
-          tokens.length >= 2 && tokens.every((t) => t.length <= 20);
-        if (looksLikePeriods) {
-          return fail(
-            lineNumber,
-            `Period lines require the 'period' keyword — use 'period ${tokens.join(' ')}'`
-          );
-        }
-      }
-
-      // Migration error: old colon syntax in data rows
-      // F2: Only fire when content after colon is predominantly numeric (old "Label: val1, val2" pattern)
-      if (line.includes(':')) {
-        const colonPos = line.indexOf(':');
-        const afterColon = line.substring(colonPos + 1).trim();
-        const numericTokens = afterColon
-          .split(/[,\s]+/)
-          .filter((v) => /^-?\d/.test(v));
-        // Only trigger if most tokens after the colon are numeric (old data pattern)
-        if (numericTokens.length >= 1) {
-          const allTokens = afterColon.split(/[,\s]+/).filter(Boolean);
-          if (numericTokens.length >= allTokens.length * 0.5) {
-            const label = line.substring(0, colonPos).trim();
-            return fail(
-              lineNumber,
-              `Colons are no longer used in slope data rows — use '${label} ${numericTokens.join(' ')}'`
-            );
-          }
-        }
-      }
-
-      // Right-scan data row parsing (requires periods to be known)
-      if (result.periods.length >= 2) {
-        const P = result.periods.length;
-        const tokens = line.split(/\s+/);
-        const values: number[] = [];
-
-        // Scan from right, capped at P values
-        let rightIdx = tokens.length - 1;
-        while (rightIdx >= 0 && values.length < P) {
-          // In-bounds by while condition (rightIdx >= 0 && < tokens.length).
-          const tok = tokens[rightIdx]!;
-          const raw = normalizeNumericToken(tok) ?? tok;
-          const num = parseFloat(raw);
-          if (!isNaN(num) && /^-?\d/.test(raw)) {
-            values.unshift(num);
-            rightIdx--;
-          } else {
-            break;
-          }
-        }
-
-        if (values.length < P) {
-          warn(
-            lineNumber,
-            `Data row has ${values.length} numeric value(s) but ${P} period(s) are defined — expected ${P} values`
-          );
-          continue;
-        }
-
-        // Remaining left tokens = label
-        const labelTokens = tokens.slice(0, rightIdx + 1);
-        const joinedLabel = labelTokens.join(' ');
-
-        if (!joinedLabel) {
-          warn(
-            lineNumber,
-            `Data row has no label — add a label before the numeric values`
-          );
-          continue;
-        }
-
-        // Color annotation (§1.5 trailing-token): `Label color` → split.
-        const { label: labelPart, colorName: colorWord } =
-          peelTrailingColorName(joinedLabel);
-        const colorPart = colorWord
-          ? (resolveColorWithDiagnostic(
-              colorWord,
-              lineNumber,
-              result.diagnostics,
-              palette
-            ) ?? null)
-          : null;
-
-        if (!labelPart) {
-          warn(
-            lineNumber,
-            `Data row has no label — add a label before the numeric values`
-          );
-          continue;
-        }
-
-        // F3: Warn on purely numeric labels — likely a mistake
-        if (/^\d[\d,.]*$/.test(labelPart)) {
-          warn(
-            lineNumber,
-            `Label '${labelPart}' looks numeric — this may indicate too many values or a missing label`
-          );
-        }
-
-        result.data.push({
-          label: labelPart,
-          values,
-          color: colorPart,
-          lineNumber,
-        });
-        continue;
-      }
-
-      // If we get here in a slope chart, it's an unrecognized line
-      if (firstLineParsed) {
-        warn(lineNumber, `Unexpected line: '${line}'.`);
-      }
-      continue;
-    }
-
-    // ── Colon-separated metadata / options (legacy + data lines) ──
-    const colonIndex = line.indexOf(':');
-
-    if (colonIndex !== -1) {
-      const rawKey = line.substring(0, colonIndex).trim();
-      const key = rawKey.toLowerCase();
-
-      // Check for color annotation in raw key: "Label(color)"
-      const colorMatch = rawKey.match(/^(.+?)\(([^)]+)\)\s*$/);
-
-      if (key === 'title') {
-        result.title = line.substring(colonIndex + 1).trim();
-        result.titleLineNumber = lineNumber;
-        if (result.type === 'quadrant') {
-          result.quadrantTitleLineNumber = lineNumber;
-        }
-        continue;
-      }
-
-      if (key === 'order') {
-        const v = line
-          .substring(colonIndex + 1)
-          .trim()
-          .toLowerCase();
-        if (v === 'name' || v === 'group' || v === 'degree') {
-          result.arcOrder = v;
-        }
-        continue;
-      }
-
-      if (key === 'rotate') {
-        const v = line
-          .substring(colonIndex + 1)
-          .trim()
-          .toLowerCase();
-        if (v === 'none' || v === 'mixed' || v === 'angled') {
-          result.cloudOptions.rotate = v;
-        }
-        continue;
-      }
-
-      if (key === 'max') {
-        const v = parseInt(line.substring(colonIndex + 1).trim(), 10);
-        if (!isNaN(v) && v > 0) {
-          result.cloudOptions.max = v;
-        }
-        continue;
-      }
-
-      if (key === 'size') {
-        const v = line.substring(colonIndex + 1).trim();
-        const parts = v.split(',').map((s) => parseInt(s.trim(), 10));
-        if (
-          parts.length === 2 &&
-          parts.every((n) => !isNaN(n) && n > 0) &&
-          // In-bounds by length === 2 check.
-          parts[0]! < parts[1]!
-        ) {
-          // In-bounds by length === 2 check.
-          result.cloudOptions.minSize = parts[0]!;
-          result.cloudOptions.maxSize = parts[1]!;
-        }
-        continue;
-      }
-
-      // Data line: "Label: value1, value2" or "Label(color): value1, value2"
-      // Capture groups 1-2 guaranteed by the regex match.
-      const labelPart = colorMatch ? colorMatch[1]!.trim() : rawKey;
-      const colorPart = colorMatch
-        ? (resolveColorWithDiagnostic(
-            colorMatch[2]!.trim(),
-            lineNumber,
-            result.diagnostics,
-            palette
-          ) ?? null)
-        : null;
-      const valuePart = line.substring(colonIndex + 1).trim();
-      const values = valuePart.split(',').map((v) => v.trim());
-
-      // Check if this looks like a data line (values should be numeric)
-      const numericValues: number[] = [];
-      let allNumeric = true;
-      for (const v of values) {
-        const num = parseFloat(v);
-        if (isNaN(num)) {
-          allNumeric = false;
-          break;
-        }
-        numericValues.push(num);
-      }
-
-      if (allNumeric && numericValues.length > 0) {
-        // Wordcloud does not use colon data format — skip to freeform handling
-        if (result.type !== 'wordcloud') {
-          result.data.push({
-            label: labelPart,
-            values: numericValues,
-            color: colorPart,
-            lineNumber,
-          });
-          continue;
-        }
-      }
-    }
-
-    // For wordcloud: collect non-metadata lines for freeform fallback
-    if (result.type === 'wordcloud') {
-      if (colonIndex === -1 && !line.includes(' ')) {
-        // Single bare word — structured mode
-        result.words.push({ text: line, weight: 10, lineNumber });
-      } else if (colonIndex === -1) {
-        // Try "word weight" or "multi-word-label weight" space-separated format
-        const lastSpace = line.lastIndexOf(' ');
-        const rawWeight = lastSpace >= 0 ? line.substring(lastSpace + 1) : '';
-        const maybeWeight =
-          lastSpace >= 0
-            ? parseFloat(normalizeNumericToken(rawWeight) ?? rawWeight)
-            : NaN;
-        if (lastSpace >= 0 && !isNaN(maybeWeight) && maybeWeight > 0) {
-          const wordText = line.substring(0, lastSpace).trim();
-          result.words.push({
-            text: wordText,
-            weight: maybeWeight,
-            lineNumber,
-          });
-        } else {
-          freeformLines.push(line);
-        }
-      } else {
-        // Non-numeric colon line — freeform text
-        freeformLines.push(line);
-      }
-      continue;
-    }
-
-    // Catch-all: nothing matched this line
-    // Skip on first line — chart type suggestion is handled post-loop
-    if (firstLineParsed) {
-      warn(lineNumber, `Unexpected line: '${line}'.`);
-    }
-  }
-
-  // Validation
-  if (!result.type) {
-    const validD3Types = [...VALID_D3_TYPES];
-    const firstNonEmpty =
-      lines.find((l) => l.trim() && !l.trim().startsWith('//'))?.trim() ?? '';
-    // split always returns at least one element.
-    const hint = suggest(
-      firstNonEmpty.split(/\s/)[0]!.toLowerCase(),
-      validD3Types
-    );
-    let msg = `Unsupported chart type: "${firstNonEmpty.split(/\s/)[0]}". Supported types: ${validD3Types.join(', ')}`;
-    if (hint) msg += `. ${hint}`;
-    return fail(1, msg);
-  }
-
-  // Sequence diagrams are parsed by their own dedicated parser
-  if (result.type === 'sequence') {
-    return result;
-  }
-
-  if (result.type === 'wordcloud') {
-    // If no structured words were found, parse freeform text as word frequencies
-    if (result.words.length === 0 && freeformLines.length > 0) {
-      result.words = tokenizeFreeformText(freeformLines.join(' '));
-    }
-    if (result.words.length === 0) {
-      warn(
-        1,
-        'No words found. Add words as "word weight" (space-separated), one per line, or paste freeform text'
-      );
-    }
-    // Apply max word limit (words are already sorted by weight desc for freeform)
-    if (
-      result.cloudOptions.max > 0 &&
-      result.words.length > result.cloudOptions.max
-    ) {
-      result.words = result.words
-        .slice()
-        .sort((a, b) => b.weight - a.weight)
-        .slice(0, result.cloudOptions.max);
-    }
-    return result;
-  }
-
-  if (result.type === 'arc') {
-    if (result.links.length === 0) {
-      warn(
-        1,
-        'No links found. Add links as "Source -> Target weight" (e.g., "Alice -> Bob 5")'
-      );
-    }
-    // Validate arc ordering vs groups
-    if (result.arcNodeGroups.length > 0) {
-      if (result.arcOrder === 'name' || result.arcOrder === 'degree') {
-        warn(
-          1,
-          `Cannot use "order ${result.arcOrder}" with [Group] headers. Use "order group" or remove group headers.`
-        );
-        result.arcOrder = 'group';
-      }
-      if (result.arcOrder === 'appearance') {
-        result.arcOrder = 'group';
-      }
-    }
-    return result;
-  }
-
-  if (result.type === 'timeline') {
-    if (result.timelineEvents.length === 0) {
-      warn(
-        1,
-        'No events found. Add events like "1718 Event Name" or "1716 -> 1717 Event Name"'
-      );
-    }
-    // Validate tag values and inject defaults
-    if (result.timelineTagGroups.length > 0) {
-      validateTagValues(
-        result.timelineEvents,
-        result.timelineTagGroups,
-        (line, msg) =>
-          result.diagnostics.push(makeDgmoError(line, msg, 'warning')),
-        suggest
-      );
-      validateTagGroupNames(
-        result.timelineTagGroups,
-        (line, msg) =>
-          result.diagnostics.push(makeDgmoError(line, msg, 'warning')),
-        (line, msg) => {
-          const diag = makeDgmoError(line, msg);
-          result.diagnostics.push(diag);
-          if (!result.error) result.error = formatDgmoError(diag);
-        }
-      );
-      for (const group of result.timelineTagGroups) {
-        if (!group.defaultValue) continue;
-        const key = tagAttrKey(group.name);
-        for (const event of result.timelineEvents) {
-          if (!event.metadata[key]) {
-            event.metadata[key] = group.defaultValue;
-          }
-        }
-      }
-    }
-
-    return result;
-  }
-
-  if (result.type === 'venn') {
-    if (result.vennSets.length < 2) {
-      return fail(
-        1,
-        'At least 2 sets are required. Add set names (e.g., "Apples", "Oranges")'
-      );
-    }
-    if (result.vennSets.length > 3) {
-      return fail(1, 'Venn diagrams support 2–3 sets');
-    }
-    // Build lookup: full name (lowercase) and alias → canonical name
-    const setNameLower = new Map<string, string>(
-      result.vennSets.map((s) => [s.name.toLowerCase(), s.name])
-    );
-    const aliasLower = new Map<string, string>();
-    for (const s of result.vennSets) {
-      if (s.alias) aliasLower.set(s.alias.toLowerCase(), s.name);
-    }
-    const resolveSetRef = (ref: string): string | null =>
-      setNameLower.get(ref.toLowerCase()) ??
-      aliasLower.get(ref.toLowerCase()) ??
-      null;
-
-    // Resolve intersection set references; drop invalid ones with a diagnostic
-    const validOverlaps: VennOverlap[] = [];
-    for (const ov of result.vennOverlaps) {
-      const resolvedSets: string[] = [];
-      let valid = true;
-      for (const ref of ov.sets) {
-        const resolved = resolveSetRef(ref);
-        if (!resolved) {
-          result.diagnostics.push(
-            makeDgmoError(
-              ov.lineNumber,
-              `Intersection references unknown set or alias "${ref}" — declare the set first on its own line ('Name [as <alias>] [color]', e.g. 'Navigation as nav blue'), then reference it by that name or alias. (§16.4)`
-            )
-          );
-          if (!result.error)
-            // diagnostics non-empty: we just pushed to it.
-            result.error = formatDgmoError(
-              result.diagnostics[result.diagnostics.length - 1]!
-            );
-          valid = false;
-          break;
-        }
-        resolvedSets.push(resolved);
-      }
-      if (valid) validOverlaps.push({ ...ov, sets: resolvedSets.sort() });
-    }
-    result.vennOverlaps = validOverlaps;
-    return result;
-  }
-
-  if (result.type === 'quadrant') {
-    if (result.quadrantPoints.length === 0) {
-      warn(
-        1,
-        'No data points found. Add points as "Label x, y" (e.g., "Item A 0.5, 0.7")'
-      );
-    }
-    return result;
-  }
-
-  // Slope chart validation
-  if (result.periods.length < 2) {
-    return fail(
-      1,
-      "Missing 'period' directive. Add 'period 2020 2024' before data rows (minimum 2 periods required)"
-    );
-  }
-
-  if (result.data.length === 0) {
-    warn(
-      1,
-      "No data lines found. Add data as 'Label value1 value2' (e.g., 'Blackbeard 40 4')"
-    );
-  }
-
-  // Validate value counts match period count — warn and skip mismatched items
-  for (const item of result.data) {
-    if (item.values.length !== result.periods.length) {
-      warn(
-        item.lineNumber,
-        `Data item "${item.label}" has ${item.values.length} value(s) but ${result.periods.length} period(s) are defined`
-      );
-    }
-  }
-  result.data = result.data.filter(
-    (item) => item.values.length === result.periods.length
-  );
-
-  return result;
-}
-
-const STOP_WORDS = new Set([
-  'a',
-  'an',
-  'the',
-  'and',
-  'or',
-  'but',
-  'in',
-  'on',
-  'at',
-  'to',
-  'for',
-  'of',
-  'with',
-  'by',
-  'is',
-  'am',
-  'are',
-  'was',
-  'were',
-  'be',
-  'been',
-  'being',
-  'have',
-  'has',
-  'had',
-  'do',
-  'does',
-  'did',
-  'will',
-  'would',
-  'could',
-  'should',
-  'may',
-  'might',
-  'shall',
-  'can',
-  'it',
-  'its',
-  'this',
-  'that',
-  'these',
-  'those',
-  'i',
-  'me',
-  'my',
-  'we',
-  'us',
-  'our',
-  'you',
-  'your',
-  'he',
-  'him',
-  'his',
-  'she',
-  'her',
-  'they',
-  'them',
-  'their',
-  'what',
-  'which',
-  'who',
-  'whom',
-  'how',
-  'when',
-  'where',
-  'why',
-  'not',
-  'no',
-  'nor',
-  'so',
-  'if',
-  'then',
-  'than',
-  'too',
-  'very',
-  'just',
-  'about',
-  'up',
-  'out',
-  'from',
-  'into',
-  'over',
-  'after',
-  'before',
-  'between',
-  'under',
-  'again',
-  'there',
-  'here',
-  'all',
-  'each',
-  'every',
-  'both',
-  'few',
-  'more',
-  'most',
-  'other',
-  'some',
-  'such',
-  'only',
-  'own',
-  'same',
-  'also',
-  'as',
-  'because',
-  'until',
-  'while',
-  'during',
-  'through',
-]);
-
-function tokenizeFreeformText(text: string): WordCloudWord[] {
-  const counts = new Map<string, number>();
-
-  // Split on non-letter/non-apostrophe chars, lowercase everything
-  const tokens = text
-    .toLowerCase()
-    .split(/[^a-zA-Z']+/)
-    .filter(Boolean);
-
-  for (const raw of tokens) {
-    // Strip leading/trailing apostrophes
-    const word = raw.replace(/^'+|'+$/g, '');
-    if (word.length < 2 || STOP_WORDS.has(word)) continue;
-    counts.set(word, (counts.get(word) ?? 0) + 1);
-  }
-
-  return Array.from(counts.entries())
-    .map(([text, count]) => ({ text, weight: count, lineNumber: 0 }))
-    .sort((a, b) => b.weight - a.weight);
-}