@pas7/llm-seo 0.1.6

package/dist/adapters/index.js

@@ -0,0 +1,673 @@
+import { existsSync } from 'fs';
+import { readFile, readdir } from 'fs/promises';
+import { join } from 'path';
+
+// src/adapters/next/manifest.ts
+function fromNextContentManifest(manifest, options = {}) {
+  const { slugPrefix = "", defaultLocale } = options;
+  return manifest.items.map((item) => {
+    const result = {
+      slug: slugPrefix + item.slug
+    };
+    if (item.locales !== void 0) {
+      result.locales = item.locales;
+    } else if (defaultLocale !== void 0) {
+      result.locales = [defaultLocale];
+    }
+    if (item.publishedAt !== void 0) {
+      result.publishedAt = item.publishedAt;
+    }
+    if (item.updatedAt !== void 0) {
+      result.updatedAt = item.updatedAt;
+    }
+    if (item.title !== void 0) {
+      result.title = item.title;
+    }
+    if (item.description !== void 0) {
+      result.description = item.description;
+    }
+    if (item.priority !== void 0) {
+      result.priority = item.priority;
+    }
+    if (item.canonicalOverride !== void 0) {
+      result.canonicalOverride = item.canonicalOverride;
+    }
+    return result;
+  });
+}
+function parseFrontmatter(content) {
+  const frontmatterRegex = /^---\s*\n([\s\S]*?)\n---\s*\n([\s\S]*)$/;
+  const match = content.match(frontmatterRegex);
+  if (!match) {
+    return { frontmatter: {}, body: content };
+  }
+  const frontmatterText = match[1] ?? "";
+  const body = match[2] ?? content;
+  const frontmatter = {};
+  const lines = frontmatterText.split("\n");
+  for (const line of lines) {
+    const colonIndex = line.indexOf(":");
+    if (colonIndex === -1) continue;
+    const key = line.slice(0, colonIndex).trim();
+    let value = line.slice(colonIndex + 1).trim();
+    if (value.startsWith("[") && value.endsWith("]")) {
+      value = value.slice(1, -1).split(",").map((item) => item.trim().replace(/^['"]|['"]$/g, "")).filter((item) => item.length > 0);
+    } else {
+      value = value.replace(/^['"]|['"]$/g, "");
+    }
+    if (key === "title") frontmatter.title = value;
+    else if (key === "description") frontmatter.description = value;
+    else if (key === "date" || key === "publishedAt") frontmatter.date = value;
+    else if (key === "updated" || key === "updatedAt") frontmatter.updated = value;
+    else if (key === "locale") frontmatter.locale = value;
+    else if (key === "locales") frontmatter.locales = value;
+    else if (key === "priority") frontmatter.priority = parseInt(value, 10);
+    else frontmatter[key] = value;
+  }
+  return { frontmatter, body };
+}
+function filePathToSlug(filePath, extensions) {
+  let slug = filePath;
+  for (const ext of extensions) {
+    if (slug.endsWith(ext)) {
+      slug = slug.slice(0, -ext.length);
+      break;
+    }
+  }
+  if (slug.endsWith("/index")) {
+    slug = slug.slice(0, -6);
+  }
+  if (slug === "index") {
+    slug = "";
+  }
+  if (!slug.startsWith("/")) {
+    slug = "/" + slug;
+  }
+  return slug || "/";
+}
+async function scanDirectory(dir, extensions, basePath = "") {
+  const files = [];
+  if (!existsSync(dir)) {
+    return files;
+  }
+  const entries = await readdir(dir, { withFileTypes: true });
+  for (const entry of entries) {
+    const fullPath = join(dir, entry.name);
+    const relativePath = basePath ? join(basePath, entry.name) : entry.name;
+    if (entry.isDirectory()) {
+      const subFiles = await scanDirectory(fullPath, extensions, relativePath);
+      files.push(...subFiles);
+    } else if (entry.isFile()) {
+      const hasValidExtension = extensions.some((ext) => entry.name.endsWith(ext));
+      if (hasValidExtension) {
+        files.push(relativePath);
+      }
+    }
+  }
+  return files;
+}
+async function createManifestFromPagesDir(options) {
+  const {
+    pagesDir,
+    routePrefix = "",
+    defaultLocale,
+    extensions = [".mdx", ".md"]
+  } = options;
+  const files = await scanDirectory(pagesDir, extensions);
+  const items = [];
+  for (const file of files) {
+    const fullPath = join(pagesDir, file);
+    const content = await readFile(fullPath, "utf-8");
+    const { frontmatter } = parseFrontmatter(content);
+    const slug = routePrefix + filePathToSlug(file, extensions);
+    const item = {
+      slug
+    };
+    if (frontmatter.locales !== void 0) {
+      item.locales = frontmatter.locales;
+    } else if (frontmatter.locale !== void 0) {
+      item.locales = [frontmatter.locale];
+    } else if (defaultLocale !== void 0) {
+      item.locales = [defaultLocale];
+    }
+    if (frontmatter.date !== void 0) {
+      item.publishedAt = frontmatter.date;
+    }
+    if (frontmatter.updated !== void 0) {
+      item.updatedAt = frontmatter.updated;
+    }
+    if (frontmatter.title !== void 0) {
+      item.title = frontmatter.title;
+    }
+    if (frontmatter.description !== void 0) {
+      item.description = frontmatter.description;
+    }
+    if (frontmatter.priority !== void 0) {
+      item.priority = frontmatter.priority;
+    }
+    items.push(item);
+  }
+  return items.sort((a, b) => a.slug.localeCompare(b.slug));
+}
+function createManifestFromData(items, options = {}) {
+  const { slugPrefix = "", defaultLocale } = options;
+  return items.map((item) => {
+    const slugPath = Array.isArray(item.params.slug) ? "/" + item.params.slug.join("/") : item.params.slug.startsWith("/") ? item.params.slug : "/" + item.params.slug;
+    const result = {
+      slug: slugPrefix + slugPath
+    };
+    if (item.locale !== void 0) {
+      result.locales = [item.locale];
+    } else if (defaultLocale !== void 0) {
+      result.locales = [defaultLocale];
+    }
+    if (item.publishedAt !== void 0) {
+      result.publishedAt = item.publishedAt;
+    }
+    if (item.updatedAt !== void 0) {
+      result.updatedAt = item.updatedAt;
+    }
+    if (item.title !== void 0) {
+      result.title = item.title;
+    }
+    if (item.description !== void 0) {
+      result.description = item.description;
+    }
+    if (item.priority !== void 0) {
+      result.priority = item.priority;
+    }
+    return result;
+  });
+}
+function extractPagePaths(buildManifest) {
+  const pages = [];
+  for (const pagePath of Object.keys(buildManifest.pages)) {
+    if (pagePath.startsWith("/_")) {
+      continue;
+    }
+    const normalizedPath = normalizeNextPath(pagePath);
+    pages.push(normalizedPath);
+  }
+  return pages.sort((a, b) => a.localeCompare(b));
+}
+function normalizeNextPath(path) {
+  let normalized = path.replace(/\.(html|json)$/, "");
+  if (normalized === "/index" || normalized === "") {
+    normalized = "/";
+  }
+  if (!normalized.startsWith("/")) {
+    normalized = `/${normalized}`;
+  }
+  return normalized;
+}
+function generateNextManifest(options, buildManifest) {
+  const pages = [];
+  if (buildManifest) {
+    const pagePaths = extractPagePaths(buildManifest);
+    for (const path of pagePaths) {
+      pages.push({
+        path,
+        title: void 0,
+        // Will be filled by crawler or manually
+        description: void 0,
+        optional: false
+      });
+    }
+  }
+  return {
+    baseUrl: options.baseUrl,
+    title: options.title,
+    description: options.description,
+    pages
+  };
+}
+
+// src/core/normalize/sort.ts
+function compareStrings(a, b) {
+  return a.localeCompare(b, "en", { sensitivity: "case", numeric: true });
+}
+function sortStrings(items) {
+  return [...items].sort(compareStrings);
+}
+function sortBy(items, keyFn) {
+  return [...items].sort((a, b) => compareStrings(keyFn(a), keyFn(b)));
+}
+
+// src/core/normalize/text.ts
+function normalizeLineEndings(text, lineEndings) {
+  const normalized = text.replace(/\r\n/g, "\n").replace(/\r/g, "\n");
+  return lineEndings === "crlf" ? normalized.replace(/\n/g, "\r\n") : normalized;
+}
+function normalizeLineWhitespace(text) {
+  const lines = text.split(/\r?\n/);
+  return lines.map((line) => line.trimEnd().replace(/[ \t]+/g, (match) => " ".repeat(match.length === 0 ? 0 : match.length))).map((line) => line.replace(/  +/g, " ")).join("\n");
+}
+
+// src/core/generate/llms-txt.ts
+function createLlmsTxt(options) {
+  const { config, canonicalUrls } = options;
+  const lineEndings = config.format?.lineEndings ?? "lf";
+  const lines = [];
+  lines.push(`# ${config.brand.name}`);
+  lines.push("");
+  if (config.brand.tagline) {
+    lines.push(`> ${config.brand.tagline}`);
+    lines.push("");
+  }
+  if (config.brand.description) {
+    lines.push(config.brand.description);
+    lines.push("");
+  }
+  const hubs = config.sections?.hubs ?? [];
+  if (hubs.length > 0) {
+    lines.push("## Sections");
+    lines.push("");
+    const sortedHubs = sortStrings(hubs);
+    for (const hub of sortedHubs) {
+      const hubLabel = getHubLabel(hub);
+      lines.push(`- [${hub}](${hub}) - ${hubLabel}`);
+    }
+    lines.push("");
+  }
+  if (canonicalUrls.length > 0) {
+    lines.push("## URLs");
+    lines.push("");
+    const sortedUrls = sortStrings(canonicalUrls);
+    for (const url of sortedUrls) {
+      lines.push(`- ${url}`);
+    }
+    lines.push("");
+  }
+  const hasPolicies = config.policy?.geoPolicy || config.policy?.citationRules || config.policy?.restrictedClaims;
+  if (hasPolicies) {
+    lines.push("## Policies");
+    lines.push("");
+    if (config.policy?.geoPolicy) {
+      lines.push(`- GEO: ${config.policy.geoPolicy}`);
+    }
+    if (config.policy?.citationRules) {
+      lines.push(`- Citations: ${config.policy.citationRules}`);
+    }
+    if (config.policy?.restrictedClaims) {
+      const status = config.policy.restrictedClaims.enable ? "Enabled" : "Disabled";
+      lines.push(`- Restricted Claims: ${status}`);
+    }
+    lines.push("");
+  }
+  const hasContact = config.contact?.email || config.contact?.social || config.contact?.phone;
+  const hasBooking = config.booking?.url;
+  if (hasContact || hasBooking) {
+    lines.push("## Contact");
+    lines.push("");
+    if (config.contact?.email) {
+      lines.push(`- Email: ${config.contact.email}`);
+    }
+    if (config.contact?.phone) {
+      lines.push(`- Phone: ${config.contact.phone}`);
+    }
+    if (config.contact?.social) {
+      if (config.contact.social.twitter) {
+        lines.push(`- Twitter: ${config.contact.social.twitter}`);
+      }
+      if (config.contact.social.linkedin) {
+        lines.push(`- LinkedIn: ${config.contact.social.linkedin}`);
+      }
+      if (config.contact.social.github) {
+        lines.push(`- GitHub: ${config.contact.social.github}`);
+      }
+    }
+    if (config.booking?.url) {
+      const label = config.booking.label ?? "Book consultation";
+      lines.push(`- Booking: ${config.booking.url} (${label})`);
+    }
+    lines.push("");
+  }
+  const hasMachineHints = config.machineHints?.robots || config.machineHints?.sitemap || config.machineHints?.llmsTxt || config.machineHints?.llmsFullTxt;
+  if (hasMachineHints) {
+    lines.push("## Machine Hints");
+    lines.push("");
+    if (config.machineHints?.robots) {
+      lines.push(`- robots.txt: ${config.machineHints.robots}`);
+    }
+    if (config.machineHints?.sitemap) {
+      lines.push(`- sitemap.xml: ${config.machineHints.sitemap}`);
+    }
+    if (config.machineHints?.llmsTxt) {
+      lines.push(`- llms.txt: ${config.machineHints.llmsTxt}`);
+    }
+    if (config.machineHints?.llmsFullTxt) {
+      lines.push(`- llms-full.txt: ${config.machineHints.llmsFullTxt}`);
+    }
+    lines.push("");
+  }
+  let content = lines.join("\n");
+  content = normalizeLineWhitespace(content);
+  content = normalizeLineEndings(content, lineEndings);
+  const finalLines = content.split(lineEndings === "crlf" ? "\r\n" : "\n");
+  return {
+    content,
+    byteSize: Buffer.byteLength(content, "utf-8"),
+    lineCount: finalLines.length
+  };
+}
+function getHubLabel(hub) {
+  const labels = {
+    "/services": "Services overview",
+    "/blog": "Blog posts",
+    "/projects": "Our projects",
+    "/cases": "Case studies",
+    "/contact": "Contact us",
+    "/about": "About us",
+    "/products": "Products",
+    "/docs": "Documentation",
+    "/faq": "Frequently asked questions",
+    "/pricing": "Pricing information",
+    "/team": "Our team",
+    "/careers": "Career opportunities",
+    "/news": "News and updates",
+    "/resources": "Resources",
+    "/support": "Support center"
+  };
+  return labels[hub] ?? formatHubLabel(hub);
+}
+function formatHubLabel(hub) {
+  const clean = hub.replace(/^\//, "");
+  return clean.replace(/[-_]/g, " ").replace(/\b\w/g, (char) => char.toUpperCase());
+}
+function generateLlmsTxt(manifest, _options) {
+  const canonicalUrls = manifest.pages.map((page) => `${manifest.baseUrl}${page.path}`);
+  const config = {
+    site: { baseUrl: manifest.baseUrl },
+    brand: {
+      name: manifest.title,
+      locales: ["en"],
+      ...manifest.description && { description: manifest.description }
+    },
+    manifests: {},
+    output: {
+      paths: {
+        llmsTxt: "public/llms.txt",
+        llmsFullTxt: "public/llms-full.txt"
+      }
+    }
+  };
+  return createLlmsTxt({ config, canonicalUrls }).content;
+}
+
+// src/core/generate/llms-full-txt.ts
+function createLlmsFullTxt(options) {
+  const { config, canonicalUrls, manifestItems } = options;
+  const lineEndings = config.format?.lineEndings ?? "lf";
+  const lines = [];
+  lines.push(`# ${config.brand.name} - Full LLM Context`);
+  lines.push("");
+  if (config.brand.tagline) {
+    lines.push(`> ${config.brand.tagline}`);
+    lines.push("");
+  }
+  if (config.brand.description) {
+    lines.push(config.brand.description);
+    lines.push("");
+  }
+  if (config.brand.org) {
+    lines.push(`Organization: ${config.brand.org}`);
+  }
+  lines.push(`Locales: ${config.brand.locales.join(", ")}`);
+  lines.push("");
+  if (canonicalUrls.length > 0) {
+    lines.push("## All Canonical URLs");
+    lines.push("");
+    const sortedUrls = sortStrings(canonicalUrls);
+    for (const url of sortedUrls) {
+      lines.push(`- ${url}`);
+    }
+    lines.push("");
+  }
+  const hasPolicies = config.policy?.geoPolicy || config.policy?.citationRules || config.policy?.restrictedClaims;
+  if (hasPolicies) {
+    lines.push("## Policies");
+    lines.push("");
+    if (config.policy?.geoPolicy) {
+      lines.push("### GEO Policy");
+      lines.push(config.policy.geoPolicy);
+      lines.push("");
+    }
+    if (config.policy?.citationRules) {
+      lines.push("### Citation Rules");
+      lines.push(config.policy.citationRules);
+      lines.push("");
+    }
+    if (config.policy?.restrictedClaims) {
+      lines.push("### Restricted Claims");
+      const status = config.policy.restrictedClaims.enable ? "Enabled" : "Disabled";
+      lines.push(`Status: ${status}`);
+      if (config.policy.restrictedClaims.forbidden && config.policy.restrictedClaims.forbidden.length > 0) {
+        lines.push(`Forbidden terms: ${config.policy.restrictedClaims.forbidden.join(", ")}`);
+      }
+      if (config.policy.restrictedClaims.whitelist && config.policy.restrictedClaims.whitelist.length > 0) {
+        lines.push(`Exceptions: ${config.policy.restrictedClaims.whitelist.join(", ")}`);
+      }
+      lines.push("");
+    }
+  }
+  const hasSocial = config.contact?.social?.twitter || config.contact?.social?.linkedin || config.contact?.social?.github;
+  const hasBooking = config.booking?.url;
+  if (hasSocial || hasBooking) {
+    lines.push("## Social & Booking");
+    lines.push("");
+    if (config.contact?.social?.twitter) {
+      lines.push(`- Twitter: ${config.contact.social.twitter}`);
+    }
+    if (config.contact?.social?.linkedin) {
+      lines.push(`- LinkedIn: ${config.contact.social.linkedin}`);
+    }
+    if (config.contact?.social?.github) {
+      lines.push(`- GitHub: ${config.contact.social.github}`);
+    }
+    if (config.booking?.url) {
+      const label = config.booking.label ?? "Book consultation";
+      lines.push(`- Booking: ${config.booking.url} (${label})`);
+    }
+    lines.push("");
+  }
+  const hasMachineHints = config.machineHints?.robots || config.machineHints?.sitemap || config.machineHints?.llmsTxt || config.machineHints?.llmsFullTxt;
+  if (hasMachineHints) {
+    lines.push("## Machine Hints");
+    lines.push("");
+    if (config.machineHints?.robots) {
+      lines.push(`- robots.txt: ${config.machineHints.robots}`);
+    }
+    if (config.machineHints?.sitemap) {
+      lines.push(`- sitemap.xml: ${config.machineHints.sitemap}`);
+    }
+    if (config.machineHints?.llmsTxt) {
+      lines.push(`- llms.txt: ${config.machineHints.llmsTxt}`);
+    }
+    if (config.machineHints?.llmsFullTxt) {
+      lines.push(`- llms-full.txt: ${config.machineHints.llmsFullTxt}`);
+    }
+    lines.push("");
+  }
+  const hubs = config.sections?.hubs ?? [];
+  if (hubs.length > 0 || manifestItems.length > 0) {
+    lines.push("## Sitemap");
+    lines.push("");
+    if (hubs.length > 0) {
+      const sortedHubs = sortStrings(hubs);
+      for (const hub of sortedHubs) {
+        lines.push(`- [${hub}](${hub}) - ${getHubLabel2(hub)}`);
+      }
+    }
+    if (manifestItems.length > 0) {
+      const sortedItems = sortBy(manifestItems, (item) => item.slug);
+      for (const item of sortedItems) {
+        const url = item.canonicalOverride ?? `${config.site.baseUrl}${item.slug}`;
+        const title = item.title ?? item.slug;
+        const locales = item.locales?.join(", ") ?? config.brand.locales[0] ?? "en";
+        lines.push(`- [${title}](${url}) (${locales})`);
+      }
+    }
+    lines.push("");
+  }
+  let content = lines.join("\n");
+  content = normalizeLineWhitespace(content);
+  content = normalizeLineEndings(content, lineEndings);
+  const finalLines = content.split(lineEndings === "crlf" ? "\r\n" : "\n");
+  return {
+    content,
+    byteSize: Buffer.byteLength(content, "utf-8"),
+    lineCount: finalLines.length
+  };
+}
+function getHubLabel2(hub) {
+  const labels = {
+    "/services": "Services overview",
+    "/blog": "Blog posts",
+    "/projects": "Our projects",
+    "/cases": "Case studies",
+    "/contact": "Contact us",
+    "/about": "About us",
+    "/products": "Products",
+    "/docs": "Documentation",
+    "/faq": "Frequently asked questions",
+    "/pricing": "Pricing information",
+    "/team": "Our team",
+    "/careers": "Career opportunities",
+    "/news": "News and updates",
+    "/resources": "Resources",
+    "/support": "Support center"
+  };
+  return labels[hub] ?? formatHubLabel2(hub);
+}
+function formatHubLabel2(hub) {
+  const clean = hub.replace(/^\//, "");
+  return clean.replace(/[-_]/g, " ").replace(/\b\w/g, (char) => char.toUpperCase());
+}
+function generateLlmsFullTxt(manifest, _options) {
+  const canonicalUrls = manifest.pages.map((page) => `${manifest.baseUrl}${page.path}`);
+  const manifestItems = manifest.pages.map((page) => ({
+    slug: page.path,
+    title: page.title,
+    description: page.description
+  }));
+  const config = {
+    site: { baseUrl: manifest.baseUrl },
+    brand: {
+      name: manifest.title,
+      locales: ["en"],
+      ...manifest.description && { description: manifest.description }
+    },
+    manifests: {},
+    output: {
+      paths: {
+        llmsTxt: "public/llms.txt",
+        llmsFullTxt: "public/llms-full.txt"
+      }
+    }
+  };
+  return createLlmsFullTxt({ config, canonicalUrls, manifestItems }).content;
+}
+
+// src/adapters/next/build-hooks.ts
+function generateBuildScripts(options = {}) {
+  const {
+    configPath = "llm-seo.config.ts",
+    emitCitations = false,
+    packageManager = "pnpm"
+  } = options;
+  const configFlag = `--config ${configPath}`;
+  const citationsFlag = emitCitations ? " --emit-citations" : "";
+  const runCommand = packageManager === "npm" ? "npm run" : packageManager;
+  return {
+    "build:seo": `llm-seo generate ${configFlag}${citationsFlag}`,
+    postbuild: `${runCommand} build:seo && llm-seo check --fail-on error`,
+    "check:seo": "llm-seo check --fail-on warn"
+  };
+}
+function createRobotsLlmsPolicySnippet(options) {
+  const {
+    baseUrl,
+    allowLlmsTxt = true,
+    allowLlmsFullTxt = true,
+    allowSitemap = true,
+    additionalPaths = [],
+    userAgent = "*"
+  } = options;
+  const lines = ["# LLM SEO"];
+  try {
+    const url = new URL(baseUrl);
+    lines.push(`Host: ${url.host}`);
+  } catch {
+  }
+  lines.push(`User-agent: ${userAgent}`);
+  if (allowLlmsTxt) {
+    lines.push("Allow: /llms.txt");
+  }
+  if (allowLlmsFullTxt) {
+    lines.push("Allow: /llms-full.txt");
+  }
+  if (allowSitemap) {
+    lines.push("Allow: /sitemap.xml");
+  }
+  for (const path of additionalPaths) {
+    lines.push(`Allow: ${path}`);
+  }
+  return lines.join("\n");
+}
+function createNextConfig(options) {
+  const {
+    locales = [],
+    defaultLocale = "en",
+    trailingSlash = false,
+    unoptimizedImages = true
+  } = options;
+  const config = {
+    output: "export",
+    trailingSlash,
+    images: {
+      unoptimized: unoptimizedImages
+    }
+  };
+  if (locales.length > 0) {
+    config.i18n = {
+      locales,
+      defaultLocale
+    };
+  }
+  return config;
+}
+async function postBuildHook(options) {
+  const { outputDir, manifest, generateFull = false } = options;
+  const files = [];
+  try {
+    generateLlmsTxt(manifest);
+    const llmsTxtPath = `${outputDir}/llms.txt`;
+    files.push(llmsTxtPath);
+    if (generateFull) {
+      generateLlmsFullTxt(manifest);
+      const llmsFullTxtPath = `${outputDir}/llms-full.txt`;
+      files.push(llmsFullTxtPath);
+    }
+    return {
+      files,
+      success: true
+    };
+  } catch (error) {
+    const message = error instanceof Error ? error.message : "Unknown error";
+    return {
+      files,
+      success: false,
+      error: message
+    };
+  }
+}
+function createNextPlugin() {
+  return {
+    name: "llm-seo",
+    postBuild: postBuildHook
+  };
+}
+
+export { createManifestFromData, createManifestFromPagesDir, createNextConfig, createNextPlugin, createRobotsLlmsPolicySnippet, extractPagePaths, fromNextContentManifest, generateBuildScripts, generateNextManifest, postBuildHook };
+//# sourceMappingURL=index.js.map
+//# sourceMappingURL=index.js.map