feat: web search (#57)

backend/.gitignore

@@ -75,4 +75,5 @@ build/
 .DS_Store
 
 *.local
-.secret.*
+.secret.*
+licensed/

backend/functions/integrity-check.cjs

@@ -0,0 +1,11 @@
+#!/usr/bin/env node
+
+const fs = require('fs');
+const path = require('path');
+
+const file = path.resolve(__dirname, 'licensed/GeoLite2-City.mmdb');
+
+if (!fs.existsSync(file)) {
+    console.error(`Integrity check failed: ${file} does not exist.`);
+    process.exit(1);
+}

backend/functions/package-lock.json

@@ -24,6 +24,7 @@
         "htmlparser2": "^9.0.0",
         "jose": "^5.1.0",
         "langdetect": "^0.2.1",
+        "maxmind": "^4.3.18",
         "minio": "^7.1.3",
         "openai": "^4.20.0",
         "puppeteer": "^22.7.1",
@@ -8144,6 +8145,19 @@
         "tmpl": "1.0.5"
       }
     },
+    "node_modules/maxmind": {
+      "version": "4.3.18",
+      "resolved": "https://registry.npmjs.org/maxmind/-/maxmind-4.3.18.tgz",
+      "integrity": "sha512-5b9utU7ZxcGYTBaO7hCF0FXyfw3IpankLn+FnLW4RZS1zi97RBeSdfXJFJlk5UxNsMiFZlsdMT3lzvD+bD8MLQ==",
+      "dependencies": {
+        "mmdb-lib": "2.1.0",
+        "tiny-lru": "11.2.5"
+      },
+      "engines": {
+        "node": ">=12",
+        "npm": ">=6"
+      }
+    },
     "node_modules/media-typer": {
       "version": "0.3.0",
       "resolved": "https://registry.npmjs.org/media-typer/-/media-typer-0.3.0.tgz",
@@ -8375,6 +8389,15 @@
       "resolved": "https://registry.npmjs.org/mkdirp-classic/-/mkdirp-classic-0.5.3.tgz",
       "integrity": "sha512-gKLcREMhtuZRwRAfqP3RFW+TK4JqApVBtOIftVgjuABpAtpxhPGaDcfvbhNvD0B8iD1oUr/txX35NjcaY6Ns/A=="
     },
+    "node_modules/mmdb-lib": {
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/mmdb-lib/-/mmdb-lib-2.1.0.tgz",
+      "integrity": "sha512-tdDTZmnI5G4UoSctv2KxM/3VQt2XRj4CmR5R4VsAWsOUcS3LysHR34wtixWm/pXxXdkBDuN92auxkC0T2+qd1Q==",
+      "engines": {
+        "node": ">=10",
+        "npm": ">=6"
+      }
+    },
     "node_modules/mongodb": {
       "version": "5.9.2",
       "resolved": "https://registry.npmjs.org/mongodb/-/mongodb-5.9.2.tgz",
@@ -11059,6 +11082,14 @@
       "resolved": "https://registry.npmjs.org/tiktoken/-/tiktoken-1.0.13.tgz",
       "integrity": "sha512-JaL9ZnvTbGFMDIBeGdVkLt4qWTeCPw+n7Ock+wceAGRenuHA6nOOvMJFliNDyXsjg2osGKJWsXtO2xc74VxyDw=="
     },
+    "node_modules/tiny-lru": {
+      "version": "11.2.5",
+      "resolved": "https://registry.npmjs.org/tiny-lru/-/tiny-lru-11.2.5.tgz",
+      "integrity": "sha512-JpqM0K33lG6iQGKiigcwuURAKZlq6rHXfrgeL4/I8/REoyJTGU+tEMszvT/oTRVHG2OiylhGDjqPp1jWMlr3bw==",
+      "engines": {
+        "node": ">=12"
+      }
+    },
     "node_modules/tld-extract": {
       "version": "2.1.0",
       "resolved": "https://registry.npmjs.org/tld-extract/-/tld-extract-2.1.0.tgz",

backend/functions/package.json

@@ -2,7 +2,7 @@
   "name": "reader",
   "scripts": {
     "lint": "eslint --ext .js,.ts .",
-    "build": "tsc -p .",
+    "build": "node ./integrity-check.cjs && tsc -p .",
     "build:watch": "tsc --watch",
     "build:clean": "rm -rf ./build",
     "shell": "npm run build && firebase functions:shell",
@@ -44,6 +44,7 @@
     "htmlparser2": "^9.0.0",
     "jose": "^5.1.0",
     "langdetect": "^0.2.1",
+    "maxmind": "^4.3.18",
     "minio": "^7.1.3",
     "openai": "^4.20.0",
     "puppeteer": "^22.7.1",

backend/functions/src/cloud-functions/crawler.ts

@@ -2,7 +2,7 @@ import {
     assignTransferProtocolMeta, marshalErrorLike,
     RPCHost, RPCReflection,
     HashManager,
-    AssertionFailureError, ParamValidationError,
+    AssertionFailureError, ParamValidationError, Defer,
 } from 'civkit';
 import { singleton } from 'tsyringe';
 import { AsyncContext, CloudHTTPv2, Ctx, FirebaseStorageBucketControl, InsufficientBalanceError, Logger, OutputServerEventStream, RPCReflect } from '../shared';
@@ -34,6 +34,12 @@ export class CrawlerHost extends RPCHost {
     cacheValidMs = 1000 * 300;
     urlValidMs = 1000 * 3600 * 4;
 
+    indexText = `[Usage1] https://r.jina.ai/YOUR_URL
+[Usage2] https://s.jina.ai/YOUR_SEARCH_QUERY
+[Homepage] https://jina.ai/reader
+[Source code] https://github.com/jina-ai/reader
+`;
+
     constructor(
         protected globalLogger: Logger,
         protected puppeteerControl: PuppeteerControl,
@@ -357,10 +363,7 @@ ${this.content}
 [Balance left] ${latestUser.wallet.total_balance}
 ` : '';
 
-            return assignTransferProtocolMeta(`[Usage] https://r.jina.ai/YOUR_URL
-[Homepage] https://jina.ai/reader
-[Source code] https://github.com/jina-ai/reader
-${authMixin}`,
+            return assignTransferProtocolMeta(`${this.indexText}${authMixin}`,
                 { contentType: 'text/plain', envelope: null }
             );
         }
@@ -638,13 +641,13 @@ ${authMixin}`,
         return r;
     }
 
-    async *cachedScrap(urlToCrawl: URL, crawlOpts: ScrappingOptions, noCache: boolean = false) {
+    async *cachedScrap(urlToCrawl: URL, crawlOpts?: ScrappingOptions, noCache: boolean = false) {
         let cache;
-        if (!noCache && !crawlOpts.cookies?.length) {
+        if (!noCache && !crawlOpts?.cookies?.length) {
             cache = await this.queryCache(urlToCrawl);
         }
 
-        if (cache?.isFresh && (!crawlOpts.favorScreenshot || (crawlOpts.favorScreenshot && cache?.screenshotAvailable))) {
+        if (cache?.isFresh && (!crawlOpts?.favorScreenshot || (crawlOpts?.favorScreenshot && cache?.screenshotAvailable))) {
             yield cache.snapshot;
 
             return;
@@ -683,4 +686,47 @@ ${authMixin}`,
         return undefined;
     }
 
+
+    async *scrapMany(urls: URL[], options?: ScrappingOptions, noCache = false) {
+        const iterators = urls.map((url) => this.cachedScrap(url, options, noCache));
+
+        const results: (PageSnapshot | undefined)[] = iterators.map((_x)=> undefined);
+
+        let nextDeferred = Defer();
+        let concluded = false;
+
+        const handler = async (it: AsyncGenerator<PageSnapshot | undefined>, idx: number) => {
+            for await (const x of it) {
+                results[idx] = x;
+
+                if (x) {
+                    nextDeferred.resolve();
+                    nextDeferred = Defer();
+                }
+
+            }
+        };
+
+        Promise.all(
+            iterators.map((it, idx) => handler(it, idx))
+        ).finally(() => {
+            concluded = true;
+            nextDeferred.resolve();
+        });
+
+        yield results;
+
+        try {
+            while (!concluded) {
+                await nextDeferred.promise;
+
+                yield results;
+            }
+        } finally {
+            for (const x of iterators) {
+                x.return();
+            }
+        }
+    }
+
 }

backend/functions/src/cloud-functions/searcher.ts

@@ -0,0 +1,389 @@
+import {
+    assignTransferProtocolMeta, marshalErrorLike,
+    RPCHost, RPCReflection,
+    AssertionFailureError,
+    objHashMd5B64Of,
+} from 'civkit';
+import { singleton } from 'tsyringe';
+import { AsyncContext, CloudHTTPv2, Ctx, InsufficientBalanceError, Logger, OutputServerEventStream, RPCReflect } from '../shared';
+import { RateLimitControl } from '../shared/services/rate-limit';
+import _ from 'lodash';
+import { ScrappingOptions } from '../services/puppeteer';
+import { Request, Response } from 'express';
+import { JinaEmbeddingsAuthDTO } from '../shared/dto/jina-embeddings-auth';
+import { BraveSearchService } from '../services/brave-search';
+import { CrawlerHost } from './crawler';
+import { CookieParam } from 'puppeteer';
+
+import { parseString as parseSetCookieString } from 'set-cookie-parser';
+import { WebSearchQueryParams } from '../shared/3rd-party/brave-search';
+import { SearchResult } from '../db/searched';
+import { WebSearchApiResponse } from '../shared/3rd-party/brave-types';
+
+
+@singleton()
+export class SearcherHost extends RPCHost {
+    logger = this.globalLogger.child({ service: this.constructor.name });
+
+    cacheRetentionMs = 1000 * 3600 * 24 * 7;
+    cacheValidMs = 1000 * 3600;
+
+    constructor(
+        protected globalLogger: Logger,
+        protected rateLimitControl: RateLimitControl,
+        protected threadLocal: AsyncContext,
+        protected braveSearchService: BraveSearchService,
+        protected crawler: CrawlerHost,
+    ) {
+        super(...arguments);
+    }
+
+    override async init() {
+        await this.dependencyReady();
+
+        this.emit('ready');
+    }
+
+    @CloudHTTPv2({
+        name: 'search2',
+        runtime: {
+            memory: '4GiB',
+            timeoutSeconds: 300,
+            concurrency: 4,
+        },
+        tags: ['Searcher'],
+        httpMethod: ['get', 'post'],
+        returnType: [String, OutputServerEventStream],
+        exposeRoot: true,
+    })
+    @CloudHTTPv2({
+        runtime: {
+            memory: '8GiB',
+            timeoutSeconds: 300,
+            concurrency: 8,
+            maxInstances: 200,
+        },
+        openapi: {
+            operation: {
+                parameters: {
+                    'Accept': {
+                        description: `Specifies your preference for the response format. \n\n` +
+                            `Supported formats:\n` +
+                            `- text/event-stream\n` +
+                            `- application/json  or  text/json\n` +
+                            `- text/plain`
+                        ,
+                        in: 'header',
+                        schema: { type: 'string' }
+                    },
+                    'X-No-Cache': {
+                        description: `Ignores internal cache if this header is specified with a value.`,
+                        in: 'header',
+                        schema: { type: 'string' }
+                    },
+                    'X-Respond-With': {
+                        description: `Specifies the (non-default) form factor of the crawled data you prefer. \n\n` +
+                            `Supported formats:\n` +
+                            `- markdown\n` +
+                            `- html\n` +
+                            `- text\n` +
+                            `- screenshot\n`
+                        ,
+                        in: 'header',
+                        schema: { type: 'string' }
+                    },
+                    'X-Proxy-Url': {
+                        description: `Specifies your custom proxy if you prefer to use one. \n\n` +
+                            `Supported protocols:\n` +
+                            `- http\n` +
+                            `- https\n` +
+                            `- socks4\n` +
+                            `- socks5\n\n` +
+                            `For authentication, https://user:pass@host:port`,
+                        in: 'header',
+                        schema: { type: 'string' }
+                    },
+                    'X-Set-Cookie': {
+                        description: `Sets cookie(s) to the headless browser for your request. \n\n` +
+                            `Syntax is the same with standard Set-Cookie`,
+                        in: 'header',
+                        schema: { type: 'string' }
+                    },
+                    'X-With-Generated-Alt': {
+                        description: `Enable automatic alt-text generating for images without an meaningful alt-text.`,
+                        in: 'header',
+                        schema: { type: 'string' }
+                    },
+                }
+            }
+        },
+        tags: ['Searcher'],
+        httpMethod: ['get', 'post'],
+        returnType: [String, OutputServerEventStream],
+        exposeRoot: true,
+    })
+    async search(
+        @RPCReflect() rpcReflect: RPCReflection,
+        @Ctx() ctx: {
+            req: Request,
+            res: Response,
+        },
+        auth: JinaEmbeddingsAuthDTO
+    ) {
+        const uid = await auth.solveUID();
+        let chargeAmount = 0;
+        const noSlashPath = ctx.req.url.slice(1);
+        if (!noSlashPath) {
+            const latestUser = uid ? await auth.assertUser() : undefined;
+            const authMixin = latestUser ? `
+[Authenticated as] ${latestUser.user_id} (${latestUser.full_name})
+[Balance left] ${latestUser.wallet.total_balance}
+` : '';
+
+            return assignTransferProtocolMeta(`${this.crawler.indexText}${authMixin}`,
+                { contentType: 'text/plain', envelope: null }
+            );
+        }
+
+        if (uid) {
+            const user = await auth.assertUser();
+            if (!(user.wallet.total_balance > 0)) {
+                throw new InsufficientBalanceError(`Account balance not enough to run this query, please recharge.`);
+            }
+
+            await this.rateLimitControl.simpleRPCUidBasedLimit(rpcReflect, uid, ['CRAWL'],
+                [
+                    // 1000 requests per minute
+                    new Date(Date.now() - 60 * 1000), 1000
+                ]
+            );
+
+            rpcReflect.finally(() => {
+                if (chargeAmount) {
+                    auth.reportUsage(chargeAmount, 'reader-crawl').catch((err) => {
+                        this.logger.warn(`Unable to report usage for ${uid}`, { err: marshalErrorLike(err) });
+                    });
+                }
+            });
+        } else if (ctx.req.ip) {
+            this.threadLocal.set('ip', ctx.req.ip);
+            await this.rateLimitControl.simpleRpcIPBasedLimit(rpcReflect, ctx.req.ip, ['CRAWL'],
+                [
+                    // 100 requests per minute
+                    new Date(Date.now() - 60 * 1000), 100
+                ]
+            );
+        }
+
+        const customMode = ctx.req.get('x-respond-with') || 'default';
+        const withGeneratedAlt = Boolean(ctx.req.get('x-with-generated-alt'));
+        const noCache = Boolean(ctx.req.get('x-no-cache'));
+        const cookies: CookieParam[] = [];
+        const setCookieHeaders = ctx.req.headers['x-set-cookie'];
+        if (Array.isArray(setCookieHeaders)) {
+            for (const setCookie of setCookieHeaders) {
+                cookies.push({
+                    ...parseSetCookieString(setCookie, { decodeValues: false }) as CookieParam,
+                });
+            }
+        } else if (setCookieHeaders) {
+            cookies.push({
+                ...parseSetCookieString(setCookieHeaders, { decodeValues: false }) as CookieParam,
+            });
+        }
+        this.threadLocal.set('withGeneratedAlt', withGeneratedAlt);
+        const crawlOpts: ScrappingOptions = {
+            proxyUrl: ctx.req.get('x-proxy-url'),
+            cookies,
+            favorScreenshot: customMode === 'screenshot'
+        };
+
+        const searchQuery = noSlashPath;
+        const r = await this.cachedWebSearch({
+            q: searchQuery,
+            count: 5
+        });
+
+        const urls = r.web.results.map((x) => new URL(x.url));
+        const it = this.fetchSearchResults(customMode, urls, crawlOpts, noCache);
+
+        if (!ctx.req.accepts('text/plain') && ctx.req.accepts('text/event-stream')) {
+            const sseStream = new OutputServerEventStream();
+            rpcReflect.return(sseStream);
+
+            try {
+                for await (const scrapped of it) {
+                    if (!scrapped) {
+                        continue;
+                    }
+
+                    chargeAmount = this.getChargeAmount(scrapped);
+                    sseStream.write({
+                        event: 'data',
+                        data: scrapped,
+                    });
+                }
+            } catch (err: any) {
+                this.logger.error(`Failed to collect search result for query ${searchQuery}`,
+                    { err: marshalErrorLike(err) }
+                );
+                sseStream.write({
+                    event: 'error',
+                    data: marshalErrorLike(err),
+                });
+            }
+
+            sseStream.end();
+
+            return sseStream;
+        }
+
+        let lastScrapped;
+        if (!ctx.req.accepts('text/plain') && (ctx.req.accepts('text/json') || ctx.req.accepts('application/json'))) {
+            for await (const scrapped of it) {
+                lastScrapped = scrapped;
+
+                if (!this.qualified(scrapped)) {
+                    continue;
+                }
+
+                chargeAmount = this.getChargeAmount(scrapped);
+
+                return scrapped;
+            }
+
+            if (!lastScrapped) {
+                throw new AssertionFailureError(`No content available for query ${searchQuery}`);
+            }
+
+            chargeAmount = this.getChargeAmount(lastScrapped);
+
+            return lastScrapped;
+        }
+
+        for await (const scrapped of it) {
+            lastScrapped = scrapped;
+
+            if (!this.qualified(scrapped)) {
+                continue;
+            }
+            chargeAmount = this.getChargeAmount(scrapped);
+
+            return assignTransferProtocolMeta(`${scrapped}`, { contentType: 'text/plain', envelope: null });
+        }
+
+        if (!lastScrapped) {
+            throw new AssertionFailureError(`No content available for query ${searchQuery}`);
+        }
+
+        chargeAmount = this.getChargeAmount(lastScrapped);
+
+        return assignTransferProtocolMeta(`${lastScrapped}`, { contentType: 'text/plain', envelope: null });
+    }
+
+    async *fetchSearchResults(mode: string | 'markdown' | 'html' | 'text' | 'screenshot',
+        urls: URL[], options?: ScrappingOptions, noCache = false) {
+
+        for await (const scrapped of this.crawler.scrapMany(urls, options, noCache)) {
+            const mapped = scrapped.map((x, i) => {
+                if (!x) {
+                    const p = {
+                        toString() {
+                            return `[${i + 1}] No content available for ${urls[i]}`;
+                        }
+                    };
+                    const r = Object.create(p);
+                    r.url = urls[i].toString();
+
+                    return r;
+                }
+                return this.crawler.formatSnapshot(mode, x, urls[i]);
+            });
+
+            const resultArray = await Promise.all(mapped);
+            for (const [i, result] of resultArray.entries()) {
+                if (result && typeof result === 'object' && Object.hasOwn(result, 'toString')) {
+                    result.toString = function (this: any) {
+                        const mixins = [];
+                        if (this.publishedTime) {
+                            mixins.push(`[${i + 1}] Published Time: ${this.publishedTime}`);
+                        }
+
+                        if (mode === 'markdown') {
+                            return `[${i + 1}]\n${this.content}`;
+                        }
+
+                        return `[${i + 1}] Title: ${this.title}
+[${i + 1}] URL Source: ${this.url}${mixins.length ? `\n${mixins.join('\n')}` : ''}
+[${i + 1}] Markdown Content:
+${this.content}
+        `;
+                    };
+                }
+            }
+            resultArray.toString = function () {
+                return this.map((x, i) => x ? x.toString() : `[${i + 1}] No content available for ${urls[i]}`).join('\n\n').trimEnd() + '\n';
+            };
+
+            yield resultArray;
+        }
+    }
+
+    getChargeAmount(formatted: any[]) {
+        return _.sum(
+            formatted.map((x) => this.crawler.getChargeAmount(x) || 0)
+        );
+    }
+
+    qualified(scrapped: any[]) {
+        return _.every(scrapped, (x) =>
+            (x as any)?.title &&
+            (
+                (x as any).content ||
+                (x as any).screenShotUrl ||
+                (x as any).screenshot ||
+                (x as any).text ||
+                (x as any).html
+            )
+        );
+    }
+
+    async cachedWebSearch(query: WebSearchQueryParams, noCache: boolean = false) {
+        const queryDigest = objHashMd5B64Of(query);
+        let cache;
+        if (!noCache) {
+            cache = (await SearchResult.fromFirestoreQuery(
+                SearchResult.COLLECTION.where('queryDigest', '==', queryDigest)
+                    .orderBy('createdAt', 'desc')
+                    .limit(1)
+            ))[0];
+            if (cache) {
+                const age = Date.now() - cache.createdAt.valueOf();
+                const stale = cache.createdAt.valueOf() < (Date.now() - this.cacheValidMs);
+                this.logger.info(`${stale ? 'Stale cache exists' : 'Cache hit'} for search query "${query.q}", normalized digest: ${queryDigest}, ${age}ms old`, {
+                    query, digest: queryDigest, age, stale
+                });
+
+                if (!stale) {
+                    return cache.response as WebSearchApiResponse;
+                }
+            }
+        }
+
+        const r = await this.braveSearchService.webSearch(query);
+
+        const nowDate = new Date();
+        const record = SearchResult.from({
+            query,
+            queryDigest,
+            response: r,
+            createdAt: nowDate,
+            expireAt: new Date(nowDate.valueOf() + this.cacheRetentionMs)
+        });
+        SearchResult.save(record).catch((err) => {
+            this.logger.warn(`Failed to cache search result`, { err });
+        });
+
+        return r;
+    }
+}

backend/functions/src/db/searched.ts

@@ -0,0 +1,60 @@
+import { Also, parseJSONText, Prop } from 'civkit';
+import { FirestoreRecord } from '../shared/lib/firestore';
+import _ from 'lodash';
+
+@Also({
+    dictOf: Object
+})
+export class SearchResult extends FirestoreRecord {
+    static override collectionName = 'searchResults';
+
+    override _id!: string;
+
+    @Prop({
+        required: true
+    })
+    query!: any;
+
+    @Prop({
+        required: true
+    })
+    queryDigest!: string;
+
+    @Prop()
+    response?: any;
+
+    @Prop()
+    createdAt!: Date;
+
+    @Prop()
+    expireAt?: Date;
+
+    [k: string]: any;
+
+    static patchedFields = [
+        'query',
+        'response',
+    ];
+
+    static override from(input: any) {
+        for (const field of this.patchedFields) {
+            if (typeof input[field] === 'string') {
+                input[field] = parseJSONText(input[field]);
+            }
+        }
+
+        return super.from(input) as SearchResult;
+    }
+
+    override degradeForFireStore() {
+        const copy: any = { ...this };
+
+        for (const field of (this.constructor as typeof SearchResult).patchedFields) {
+            if (typeof copy[field] === 'object') {
+                copy[field] = JSON.stringify(copy[field]) as any;
+            }
+        }
+
+        return copy;
+    }
+}

backend/functions/src/services/brave-search.ts

@@ -0,0 +1,71 @@
+import { AsyncService, DownstreamServiceFailureError } from 'civkit';
+import { singleton } from 'tsyringe';
+import { Logger } from '../shared/services/logger';
+import { SecretExposer } from '../shared/services/secrets';
+import { BraveSearchHTTP, WebSearchQueryParams } from '../shared/3rd-party/brave-search';
+import { GEOIP_SUPPORTED_LANGUAGES, GeoIPService } from './geoip';
+import { AsyncContext } from '../shared';
+import { WebSearchOptionalHeaderOptions } from '../shared/3rd-party/brave-types';
+
+@singleton()
+export class BraveSearchService extends AsyncService {
+
+    logger = this.globalLogger.child({ service: this.constructor.name });
+
+    braveSearchHTTP!: BraveSearchHTTP;
+
+    constructor(
+        protected globalLogger: Logger,
+        protected secretExposer: SecretExposer,
+        protected geoipControl: GeoIPService,
+        protected threadLocal: AsyncContext,
+    ) {
+        super(...arguments);
+    }
+
+    override async init() {
+        await this.dependencyReady();
+        this.emit('ready');
+
+        this.braveSearchHTTP = new BraveSearchHTTP(this.secretExposer.BRAVE_SEARCH_API_KEY);
+    }
+
+    async webSearch(query: WebSearchQueryParams) {
+        const ip = this.threadLocal.get('ip');
+        const extraHeaders: WebSearchOptionalHeaderOptions = {};
+        if (ip) {
+            const geoip = await this.geoipControl.lookupCity(ip, GEOIP_SUPPORTED_LANGUAGES.EN);
+
+            if (geoip?.city) {
+                extraHeaders['X-Loc-City'] = geoip.city;
+            }
+            if (geoip?.country) {
+                extraHeaders['X-Loc-Country'] = geoip.country.code;
+            }
+            if (geoip?.timezone) {
+                extraHeaders['X-Loc-Timezone'] = geoip.timezone;
+            }
+            if (geoip?.coordinates) {
+                extraHeaders['X-Loc-Lat'] = `${geoip.coordinates[0]}`;
+                extraHeaders['X-Loc-Long'] = `${geoip.coordinates[1]}`;
+            }
+            if (geoip?.subdivisions?.length) {
+                extraHeaders['X-Loc-State'] = geoip.subdivisions[0].code;
+                extraHeaders['X-Loc-State-Name'] = geoip.subdivisions[0].name;
+            }
+        }
+        if (this.threadLocal.get('userAgent')) {
+            extraHeaders['User-Agent'] = this.threadLocal.get('userAgent');
+        }
+
+        try {
+            const r = await this.braveSearchHTTP.webSearch(query, { headers: extraHeaders as Record<string, string> });
+
+            return r.parsed;
+        } catch (err) {
+            throw new DownstreamServiceFailureError({ message: `Search failed`, cause: err });
+        }
+
+    }
+
+}

backend/functions/src/services/geoip.ts

@@ -0,0 +1,123 @@
+import { container, singleton } from 'tsyringe';
+import fsp from 'fs/promises';
+import { CityResponse, Reader } from 'maxmind';
+import { AsyncService, AutoCastable, Prop, runOnce } from 'civkit';
+import { Logger } from '../shared';
+import path from 'path';
+
+export enum GEOIP_SUPPORTED_LANGUAGES {
+    EN = 'en',
+    ZH_CN = 'zh-CN',
+    JA = 'ja',
+    DE = 'de',
+    FR = 'fr',
+    ES = 'es',
+    PT_BR = 'pt-BR',
+    RU = 'ru',
+}
+
+export class GeoIPInfo extends AutoCastable {
+    @Prop()
+    code?: string;
+
+    @Prop()
+    name?: string;
+}
+
+export class GeoIPCountryInfo extends GeoIPInfo {
+    @Prop()
+    eu?: boolean;
+}
+
+export class GeoIPCityResponse extends AutoCastable {
+    @Prop()
+    continent?: GeoIPInfo;
+
+    @Prop()
+    country?: GeoIPCountryInfo;
+
+    @Prop({
+        arrayOf: GeoIPInfo
+    })
+    subdivisions?: GeoIPInfo[];
+
+    @Prop()
+    city?: string;
+
+    @Prop({
+        arrayOf: Number
+    })
+    coordinates?: [number, number, number];
+
+    @Prop()
+    timezone?: string;
+}
+
+@singleton()
+export class GeoIPService extends AsyncService {
+
+    logger = this.globalLogger.child({ service: this.constructor.name });
+
+    mmdbCity!: Reader<CityResponse>;
+
+    constructor(
+        protected globalLogger: Logger,
+    ) {
+        super(...arguments);
+    }
+
+
+    override async init() {
+        await this.dependencyReady();
+
+        this.emit('ready');
+    }
+
+    @runOnce()
+    async _lazyload() {
+        const mmdpPath = path.resolve(__dirname, '..', '..', 'licensed', 'GeoLite2-City.mmdb');
+
+        const dbBuff = await fsp.readFile(mmdpPath, { flag: 'r', encoding: null });
+
+        this.mmdbCity = new Reader<CityResponse>(dbBuff);
+
+        this.logger.info(`Loaded GeoIP database, ${dbBuff.byteLength} bytes`);
+    }
+
+
+    async lookupCity(ip: string, lang: GEOIP_SUPPORTED_LANGUAGES = GEOIP_SUPPORTED_LANGUAGES.EN) {
+        await this._lazyload();
+
+        const r = this.mmdbCity.get(ip);
+
+        if (!r) {
+            return undefined;
+        }
+
+        return GeoIPCityResponse.from({
+            continent: r.continent ? {
+                code: r.continent?.code,
+                name: r.continent?.names?.[lang] || r.continent?.names?.en,
+            } : undefined,
+            country: r.country ? {
+                code: r.country?.iso_code,
+                name: r.country?.names?.[lang] || r.country?.names.en,
+                eu: r.country?.is_in_european_union,
+            } : undefined,
+            city: r.city?.names?.[lang] || r.city?.names?.en,
+            subdivisions: r.subdivisions?.map((x) => ({
+                code: x.iso_code,
+                name: x.names?.[lang] || x.names?.en,
+            })),
+            coordinates: r.location ? [
+                r.location.latitude, r.location.longitude, r.location.accuracy_radius
+            ] : undefined,
+            timezone: r.location?.time_zone,
+        });
+    }
+
+}
+
+const instance = container.resolve(GeoIPService);
+
+export default instance;

backend/functions/src/services/puppeteer.ts

@@ -278,7 +278,7 @@ document.addEventListener('load', handlePageLoad);
         return page;
     }
 
-    async *scrap(parsedUrl: URL, options: ScrappingOptions): AsyncGenerator<PageSnapshot | undefined> {
+    async *scrap(parsedUrl: URL, options?: ScrappingOptions): AsyncGenerator<PageSnapshot | undefined> {
         // parsedUrl.search = '';
         const url = parsedUrl.toString();
 
@@ -287,10 +287,10 @@ document.addEventListener('load', handlePageLoad);
         let screenshot: Buffer | undefined;
 
         const page = await this.pagePool.acquire();
-        if (options.proxyUrl) {
+        if (options?.proxyUrl) {
             await page.useProxy(options.proxyUrl);
         }
-        if (options.cookies) {
+        if (options?.cookies) {
             await page.setCookie(...options.cookies);
         }
 
@@ -353,7 +353,7 @@ document.addEventListener('load', handlePageLoad);
                     yield { ...snapshot, screenshot } as PageSnapshot;
                     break;
                 }
-                if (options.favorScreenshot && snapshot?.title && snapshot?.html !== lastHTML) {
+                if (options?.favorScreenshot && snapshot?.title && snapshot?.html !== lastHTML) {
                     screenshot = await page.screenshot();
                     lastHTML = snapshot.html;
                 }

thinapps-shared

@@ -1 +1 @@
-Subproject commit 584791b789cd483dab18735416744b4d10130993
+Subproject commit 2f2cdcff7b2738be33ee5aca858ef2d65eba29ed