obs-map-plugin / src / parser.ts

/**
 * parser.ts — Note Parser (Pure)
 *
 * Parse markdown content into an array of Place objects.
 * This is a pure function with no side effects.
 */

export interface Place {
  name: string;
  url?: string;
  fields: Record<string, string>;
  notes: string[];
  lat?: number;
  lng?: number;
  startLine: number;
  endLine: number;
}

/** Regex for top-level bullet: `* ` or `- ` at column 0 */
const TOP_BULLET_RE = /^[*-] /;

/**
 * Regex for sub-bullet: any leading whitespace (tab/spaces, 1+ chars for tab,
 * 2+ chars for spaces) followed by `* ` or `- `. Uses a flat character class
 * instead of nested quantifiers to avoid catastrophic backtracking (ReDoS).
 */
const SUB_BULLET_RE = /^[\t ]{2,}[*-] |^\t[*-] /;

/** Regex for structured field: key (word chars + hyphens), colon, space, then value */
const FIELD_RE = /^([\w-]+): (.*)$/;

/** Regex for markdown link: [text](url) or [text](url "title") */
const MD_LINK_RE = /^\[([^\]]*)\]\(([^)"]*?)(?:\s+"[^"]*")?\)$/;

/** Regex for wiki-link: [[Page]] or [[Target|Display]] */
const WIKI_LINK_RE = /^\[\[([^\]]*)\]\]$/;

/**
 * Regex matching a `geo:` sub-bullet line in raw note content.
 * Shared between parser and mapView write-back logic.
 */
export const GEO_LINE_RE = /^[\t ]+[*-] geo: .*/;

/**
 * Regex for valid geo coordinates.
 * Requires digits (not just a dot), optional decimal part with digits after dot.
 * Format: lat,lng with optional space after comma.
 */
const GEO_RE = /^(-?\d+(?:\.\d+)?),\s*(-?\d+(?:\.\d+)?)$/;

/**
 * Parse the name portion of a top-level bullet, handling markdown links,
 * wiki-links, and plain text.
 */
function parseName(raw: string): { name: string; url?: string } {
  // Try markdown link
  const mdMatch = raw.match(MD_LINK_RE);
  if (mdMatch) {
    const text = mdMatch[1];
    const href = mdMatch[2];
    return {
      name: text,
      url: href || undefined,
    };
  }

  // Try wiki-link
  const wikiMatch = raw.match(WIKI_LINK_RE);
  if (wikiMatch) {
    const inner = wikiMatch[1];
    const pipeIdx = inner.indexOf("|");
    if (pipeIdx !== -1) {
      const display = inner.substring(pipeIdx + 1);
      return { name: display };
    }
    return { name: inner };
  }

  // Plain text
  return { name: raw };
}

/**
 * Parse geo field value into lat/lng if valid.
 */
function parseGeo(value: string): { lat?: number; lng?: number } {
  const match = value.match(GEO_RE);
  if (!match) return {};

  const lat = parseFloat(match[1]);
  const lng = parseFloat(match[2]);

  if (lat < -90 || lat > 90 || lng < -180 || lng > 180) return {};

  return { lat, lng };
}

/**
 * Extract the text content from a sub-bullet line, stripping indentation
 * and bullet prefix.
 */
function extractSubBulletText(line: string): string {
  return line.replace(SUB_BULLET_RE, "").trim();
}

export function parsePlaces(content: string): Place[] {
  if (!content) return [];

  // Normalize Windows line endings
  const normalized = content.replace(/\r\n/g, "\n").replace(/\r/g, "\n");
  const lines = normalized.split("\n");

  const places: Place[] = [];
  let current: {
    name: string;
    url?: string;
    fields: Record<string, string>;
    notes: string[];
    startLine: number;
    endLine: number;
  } | null = null;

  for (let i = 0; i < lines.length; i++) {
    const line = lines[i];

    if (TOP_BULLET_RE.test(line)) {
      // Finalize previous place
      if (current) {
        finalizePlace(current, places);
      }

      // Extract raw name after bullet prefix
      const raw = line.replace(/^[*-] /, "").trim();
      const { name, url } = parseName(raw);

      current = {
        name,
        url,
        fields: Object.create(null) as Record<string, string>,
        notes: [],
        startLine: i,
        endLine: i,
      };
    } else if (SUB_BULLET_RE.test(line) && current) {
      // Sub-bullet belongs to current place
      current.endLine = i;
      const text = extractSubBulletText(line);

      // Try to parse as field
      const fieldMatch = text.match(FIELD_RE);
      if (fieldMatch) {
        const key = fieldMatch[1].toLowerCase();
        const value = fieldMatch[2].trim();
        current.fields[key] = value;
      } else if (text) {
        current.notes.push(text);
      }
    }
    // Non-bullet lines are ignored (dead zones)
  }

  // Finalize last place
  if (current) {
    finalizePlace(current, places);
  }

  return places;
}

/**
 * Finalize a place block: parse geo, exclude empty names, push to results.
 */
function finalizePlace(
  block: {
    name: string;
    url?: string;
    fields: Record<string, string>;
    notes: string[];
    startLine: number;
    endLine: number;
  },
  places: Place[]
): void {
  // Exclude empty/whitespace-only names
  if (!block.name.trim()) return;

  const place: Place = {
    name: block.name,
    url: block.url,
    fields: block.fields,
    notes: block.notes,
    startLine: block.startLine,
    endLine: block.endLine,
  };

  // Parse geo if present
  if (block.fields.geo) {
    const { lat, lng } = parseGeo(block.fields.geo);
    if (lat !== undefined && lng !== undefined) {
      place.lat = lat;
      place.lng = lng;
    }
  }

  places.push(place);
}