explorbot 0.4.8 → 0.4.9

package/src/config.ts

@@ -5,8 +5,9 @@ import { pathToFileURL } from 'node:url';
 import { parseEnv } from 'node:util';
 import dedent from 'dedent';
 import matter from 'gray-matter';
-import { type SiteRecord, findGlobalConfig, globalEnvPath, isGlobalConfigPath, registerSite, resolveSiteTarget } from './global-config.js';
+import { EXPLORBOT_CONFIG_PATHS, type SiteRecord, findGlobalConfig, globalEnvPath, isGlobalConfigPath, loadSiteConfig, registerSite, resolveSiteTarget } from './global-config.js';
 import { getCliName } from './utils/cli-name.js';
+import { deepMerge } from './utils/merge.js';
 import { log, tag } from './utils/logger.js';
 
 export const PROVIDERS: Record<string, ProviderInfo> = {
@@ -268,7 +269,7 @@ const config: ExplorbotConfig = {
 
 type RuleEntry = string | Record<string, string>;
 
-export const EXPLORBOT_CONFIG_PATHS = ['explorbot.config.js', 'explorbot.config.mjs', 'explorbot.config.ts'];
+export { EXPLORBOT_CONFIG_PATHS };
 
 export const EXPLORBOT_ENV_VARS: EnvVar[] = [
   { name: 'EXPLORBOT_AI_PROVIDER', required: true, description: 'Provider name; fills every model role from its recommended models. Turns on config-free mode' },
@@ -323,6 +324,7 @@ export class ConfigParser {
   private runtimeTarget: string | null = null;
   private site: SiteRecord | null = null;
   private siteStartPath = '/';
+  private siteConfigPath: string | null = null;
 
   private constructor() {}
 
@@ -407,16 +409,18 @@ export class ConfigParser {
         log(`Configuration built from EXPLORBOT_* environment variables. Output: ${outputRoot}`);
       }
 
-      this.config = this.resolveConfig(loadedConfig as ExplorbotConfig, options);
-      await resolveConfigModels(this.config.ai);
-      this.runtimeTarget = target;
-      this.configPath = sourcePath;
+      let config = this.resolveConfig(loadedConfig as ExplorbotConfig, options);
+      await resolveConfigModels(config.ai);
       this.site = null;
+      this.siteConfigPath = null;
 
       if (resolvedPath && isGlobalConfigPath(resolvedPath)) {
-        this.enterGlobalMode(this.config, target);
+        config = await this.enterGlobalMode(config, target);
       }
 
+      this.config = config;
+      this.runtimeTarget = target;
+      this.configPath = sourcePath;
       this.applyEnvSpec(this.config);
 
       // Restore original directory after successful config load
@@ -473,6 +477,10 @@ export class ConfigParser {
     return this.site;
   }
 
+  public getSiteConfigPath(): string | null {
+    return this.siteConfigPath;
+  }
+
   public resolveTargetPath(target?: string): string {
     if (!this.site) {
       const configured = this.config?.playwright?.url || this.config?.web?.url;
@@ -512,6 +520,7 @@ export class ConfigParser {
       ConfigParser.instance.runtimeTarget = null;
       ConfigParser.instance.site = null;
       ConfigParser.instance.siteStartPath = '/';
+      ConfigParser.instance.siteConfigPath = null;
     }
   }
 
@@ -563,16 +572,24 @@ export class ConfigParser {
     }
   }
 
-  private enterGlobalMode(config: ExplorbotConfig, target: string | null): void {
+  private async enterGlobalMode(config: ExplorbotConfig, target: string | null): Promise<ExplorbotConfig> {
     const site = resolveSiteTarget(target || undefined, config.web?.url || config.playwright?.url);
     this.site = registerSite(site.baseUrl);
     this.siteStartPath = site.path;
 
-    config.dirs = { knowledge: 'knowledge', experience: 'experience', output: 'output' };
-    config.playwright = { ...config.playwright, browser: config.playwright?.browser || 'chromium', url: site.baseUrl };
+    const { path: sitePath, config: siteConfig } = await loadSiteConfig(this.site.dir, site.baseUrl);
+    this.siteConfigPath = sitePath;
+
+    const merged = deepMerge(config, siteConfig);
+    await resolveConfigModels(merged.ai);
+    resolveLangfuse(merged.ai);
+
+    merged.dirs = { knowledge: 'knowledge', experience: 'experience', output: 'output' };
+    merged.playwright = { ...merged.playwright, browser: merged.playwright?.browser || 'chromium', url: site.baseUrl };
     materializeKnowledge(this.site.dir);
 
     log(`Global mode: ${site.baseUrl} stored in ${this.site.dir}`);
+    return merged;
   }
 
   private applyEnvSpec(config: ExplorbotConfig): void {
@@ -720,21 +737,7 @@ export class ConfigParser {
       },
     };
 
-    return this.deepMerge(defaults, config);
-  }
-
-  private deepMerge(target: any, source: any): any {
-    const result = { ...target };
-
-    for (const key in source) {
-      if (source[key] && typeof source[key] === 'object' && !Array.isArray(source[key]) && source[key].constructor === Object) {
-        result[key] = this.deepMerge(result[key] || {}, source[key]);
-      } else {
-        result[key] = source[key];
-      }
-    }
-
-    return result;
+    return deepMerge(defaults, config);
   }
 
   public ensureDirectory(path: string): void {

package/src/global-config.ts

@@ -1,10 +1,14 @@
 import { existsSync, mkdirSync, readFileSync, readdirSync, writeFileSync } from 'node:fs';
 import os from 'node:os';
-import { join } from 'node:path';
+import { join, resolve } from 'node:path';
+import { pathToFileURL } from 'node:url';
+import dedent from 'dedent';
 
 const GLOBAL_CONFIG_NAMES = ['config.js', 'config.mjs', 'config.ts'];
 const SITE_DIRS = ['knowledge', 'experience', 'output'];
 
+export const EXPLORBOT_CONFIG_PATHS = ['explorbot.config.js', 'explorbot.config.mjs', 'explorbot.config.ts'];
+
 export function globalDir(): string {
   return join(os.homedir(), '.explorbot');
 }
@@ -18,11 +22,7 @@ export function globalConfigPath(): string {
 }
 
 export function findGlobalConfig(): string | null {
-  for (const name of GLOBAL_CONFIG_NAMES) {
-    const fullPath = join(globalDir(), name);
-    if (existsSync(fullPath)) return fullPath;
-  }
-  return null;
+  return firstExisting(globalDir(), GLOBAL_CONFIG_NAMES);
 }
 
 export function isGlobalConfigPath(configPath: string): boolean {
@@ -47,6 +47,52 @@ export function listSites(): SiteRecord[] {
     .sort((a, b) => b.lastRunAt.localeCompare(a.lastRunAt));
 }
 
+export function findSiteConfig(dir: string): string | null {
+  return firstExisting(dir, EXPLORBOT_CONFIG_PATHS);
+}
+
+function ensureSiteConfig(dir: string, baseUrl: string): string {
+  const existing = findSiteConfig(dir);
+  if (existing) return existing;
+
+  const path = join(dir, EXPLORBOT_CONFIG_PATHS[0]);
+  writeFileSync(path, siteConfigTemplate(baseUrl), 'utf8');
+  return path;
+}
+
+export async function loadSiteConfig(dir: string, baseUrl: string): Promise<{ path: string; config: any }> {
+  const path = ensureSiteConfig(dir, baseUrl);
+  const module = await import(pathToFileURL(resolve(path)).href);
+  const config = module.default || module;
+  validateSiteConfig(config, path, baseUrl);
+  return { path, config };
+}
+
+function validateSiteConfig(config: any, configPath: string, baseUrl: string): void {
+  const url = config?.web?.url;
+  if (!url) {
+    throw new Error(dedent`
+      Site config is missing web.url.
+        ${configPath}
+
+      Add it so the config states which site it configures:
+        web: { url: '${baseUrl}' },
+    `);
+  }
+
+  const declared = URL.parse(url)?.origin;
+  if (declared === baseUrl) return;
+
+  throw new Error(dedent`
+    Site config declares a different site.
+      ${configPath}
+      web.url:  ${url}
+      site:     ${baseUrl}
+
+    Fix web.url, or explore ${url} to register it as its own site.
+  `);
+}
+
 export function findSiteWith(subpath: string): SiteRecord | undefined {
   return listSites().find((site) => existsSync(join(site.dir, subpath)));
 }
@@ -107,6 +153,35 @@ export function resolveSiteTarget(target?: string, defaultBaseUrl?: string): Sit
   return { baseUrl: site.url, path };
 }
 
+function firstExisting(dir: string, names: string[]): string | null {
+  for (const name of names) {
+    const fullPath = join(dir, name);
+    if (existsSync(fullPath)) return fullPath;
+  }
+  return null;
+}
+
+function siteConfigTemplate(baseUrl: string): string {
+  return `// Config for ${baseUrl}
+// Extends ~/.explorbot/config.js — set only what differs.
+const config = {
+  web: {
+    url: '${baseUrl}',
+  },
+
+  // ai: {
+  //   model: 'openrouter/openai/gpt-oss-120b',
+  // },
+
+  // playwright: {
+  //   show: true,
+  // },
+};
+
+export default config;
+`;
+}
+
 function readSite(folder: string): SiteRecord | null {
   const dir = join(sitesDir(), folder);
   const metaPath = join(dir, 'site.json');

package/src/utils/html.ts

@@ -1420,6 +1420,12 @@ function cleanElement(element: parse5TreeAdapter.Element): void {
     'aria-labelledby',
     'aria-describedby',
     'aria-owns',
+    'aria-checked',
+    'aria-expanded',
+    'aria-selected',
+    'aria-pressed',
+    'aria-current',
+    'aria-disabled',
     'role',
     'title',
     'href',

package/src/utils/merge.ts

@@ -0,0 +1,13 @@
+export function deepMerge(target: any, source: any): any {
+  const result = { ...target };
+
+  for (const key in source) {
+    if (source[key] && typeof source[key] === 'object' && !Array.isArray(source[key]) && source[key].constructor === Object) {
+      result[key] = deepMerge(result[key] || {}, source[key]);
+      continue;
+    }
+    result[key] = source[key];
+  }
+
+  return result;
+}