Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 0 additions & 5 deletions .github/workflows/ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -14,10 +14,6 @@ jobs:
- uses: actions/checkout@v4
- uses: oven-sh/setup-bun@v2
- run: bun install
# scrapegraph-js is pinned to a GitHub commit (PR #13 head, not yet on npm)
# and ships without a prebuilt dist/ — build it in-place so module
# resolution works.
- run: cd node_modules/scrapegraph-js && bun install && bun run build
- run: bun test

lint:
Expand All @@ -27,5 +23,4 @@ jobs:
- uses: actions/checkout@v4
- uses: oven-sh/setup-bun@v2
- run: bun install
- run: cd node_modules/scrapegraph-js && bun install && bun run build
- run: bun run check
4 changes: 2 additions & 2 deletions bun.lock

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

2 changes: 1 addition & 1 deletion package.json
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@
"chalk": "^5.4.1",
"citty": "^0.1.6",
"dotenv": "^17.2.4",
"scrapegraph-js": "github:ScrapeGraphAI/scrapegraph-js#096c110"
"scrapegraph-js": "^2.1.0"
},
"devDependencies": {
"@biomejs/biome": "^1.9.4",
Expand Down
64 changes: 33 additions & 31 deletions src/commands/crawl.ts
Original file line number Diff line number Diff line change
@@ -1,27 +1,13 @@
import { defineCommand } from "citty";
import { crawl } from "scrapegraph-js";
import type { ApiCrawlRequest, ApiScrapeFormatEntry } from "scrapegraph-js";
import type { CrawlRequest, FetchConfig } from "scrapegraph-js";
import { resolveApiKey } from "../lib/folders.js";
import { BASE_FORMATS, type BaseFormat, buildBaseFormat } from "../lib/formats.js";
import * as log from "../lib/log.js";

const FORMATS = [
"markdown",
"html",
"screenshot",
"branding",
"links",
"images",
"summary",
] as const;
type Format = (typeof FORMATS)[number];
import { parseIntArg, parseJsonArg } from "../lib/parse.js";

const POLL_INTERVAL_MS = 3000;

function buildFormat(f: Format): ApiScrapeFormatEntry {
if (f === "markdown" || f === "html") return { type: f, mode: "normal" };
return { type: f } as ApiScrapeFormatEntry;
}

export default defineCommand({
meta: {
name: "crawl",
Expand All @@ -36,7 +22,7 @@ export default defineCommand({
format: {
type: "string",
alias: "f",
description: `Per-page format(s), comma-separated: ${FORMATS.join(", ")} (default: markdown)`,
description: `Per-page format(s), comma-separated: ${BASE_FORMATS.join(", ")} (default: markdown)`,
},
"max-pages": { type: "string", description: "Maximum pages to crawl (default 50, max 1000)" },
"max-depth": { type: "string", description: "Crawl depth (default 2)" },
Expand Down Expand Up @@ -64,24 +50,40 @@ export default defineCommand({
.map((f) => f.trim())
.filter(Boolean);
for (const f of requested) {
if (!FORMATS.includes(f as Format))
out.error(`Unknown format: ${f}. Valid: ${FORMATS.join(", ")}`);
if (!BASE_FORMATS.includes(f as BaseFormat))
out.error(`Unknown format: ${f}. Valid: ${BASE_FORMATS.join(", ")}`);
}
const formats = requested.map((f) => buildFormat(f as Format));

const params: ApiCrawlRequest = { url: args.url, formats };
const mut = params as Record<string, unknown>;
if (args["max-pages"]) mut.maxPages = Number(args["max-pages"]);
if (args["max-depth"]) mut.maxDepth = Number(args["max-depth"]);
if (args["max-links-per-page"]) mut.maxLinksPerPage = Number(args["max-links-per-page"]);
if (args["allow-external"]) mut.allowExternal = true;
if (args["include-patterns"]) mut.includePatterns = JSON.parse(args["include-patterns"]);
if (args["exclude-patterns"]) mut.excludePatterns = JSON.parse(args["exclude-patterns"]);
const formats = requested.map((f) => buildBaseFormat(f as BaseFormat));

const fetchConfig: Record<string, unknown> = {};
if (args.mode) fetchConfig.mode = args.mode;
if (args.stealth) fetchConfig.stealth = true;
if (Object.keys(fetchConfig).length > 0) mut.fetchConfig = fetchConfig;

const params: CrawlRequest = {
url: args.url,
formats,
...(args["max-pages"] && { maxPages: parseIntArg(args["max-pages"], "max-pages", out) }),
...(args["max-depth"] && { maxDepth: parseIntArg(args["max-depth"], "max-depth", out) }),
...(args["max-links-per-page"] && {
maxLinksPerPage: parseIntArg(args["max-links-per-page"], "max-links-per-page", out),
}),
...(args["allow-external"] && { allowExternal: true }),
...(args["include-patterns"] && {
includePatterns: parseJsonArg(
args["include-patterns"],
"include-patterns",
out,
) as string[],
}),
...(args["exclude-patterns"] && {
excludePatterns: parseJsonArg(
args["exclude-patterns"],
"exclude-patterns",
out,
) as string[],
}),
...(Object.keys(fetchConfig).length > 0 && { fetchConfig: fetchConfig as FetchConfig }),
};

out.start("Starting crawl");
const job = await crawl.start(apiKey, params);
Expand Down
23 changes: 14 additions & 9 deletions src/commands/extract.ts
Original file line number Diff line number Diff line change
@@ -1,8 +1,9 @@
import { defineCommand } from "citty";
import { extract } from "scrapegraph-js";
import type { ApiExtractRequest } from "scrapegraph-js";
import type { ExtractRequest, FetchConfig } from "scrapegraph-js";
import { resolveApiKey } from "../lib/folders.js";
import * as log from "../lib/log.js";
import { parseIntArg, parseJsonArg } from "../lib/parse.js";

export default defineCommand({
meta: {
Expand Down Expand Up @@ -42,16 +43,20 @@ export default defineCommand({
const fetchConfig: Record<string, unknown> = {};
if (args.mode) fetchConfig.mode = args.mode;
if (args.stealth) fetchConfig.stealth = true;
if (args.scrolls) fetchConfig.scrolls = Number(args.scrolls);
if (args.cookies) fetchConfig.cookies = JSON.parse(args.cookies);
if (args.headers) fetchConfig.headers = JSON.parse(args.headers);
if (args.scrolls) fetchConfig.scrolls = parseIntArg(args.scrolls, "scrolls", out);
if (args.cookies) fetchConfig.cookies = parseJsonArg(args.cookies, "cookies", out);
if (args.headers) fetchConfig.headers = parseJsonArg(args.headers, "headers", out);
if (args.country) fetchConfig.country = args.country;

const params: ApiExtractRequest = { url: args.url, prompt: args.prompt };
if (args.schema) (params as Record<string, unknown>).schema = JSON.parse(args.schema);
if (args["html-mode"]) (params as Record<string, unknown>).mode = args["html-mode"];
if (Object.keys(fetchConfig).length > 0)
(params as Record<string, unknown>).fetchConfig = fetchConfig;
const params: ExtractRequest = {
url: args.url,
prompt: args.prompt,
...(args.schema && {
schema: parseJsonArg(args.schema, "schema", out) as Record<string, unknown>,
}),
...(args["html-mode"] && { mode: args["html-mode"] as "normal" | "reader" | "prune" }),
...(Object.keys(fetchConfig).length > 0 && { fetchConfig: fetchConfig as FetchConfig }),
};

out.start("Extracting");
const result = await extract(apiKey, params);
Expand Down
22 changes: 13 additions & 9 deletions src/commands/history.ts
Original file line number Diff line number Diff line change
Expand Up @@ -2,20 +2,21 @@ import * as p from "@clack/prompts";
import chalk from "chalk";
import { defineCommand } from "citty";
import { history } from "scrapegraph-js";
import type { ApiHistoryEntry, ApiHistoryService } from "scrapegraph-js";
import type { HistoryEntry, Service } from "scrapegraph-js";
import { resolveApiKey } from "../lib/folders.js";
import * as log from "../lib/log.js";
import { parseIntArg } from "../lib/parse.js";

const SERVICES = ["scrape", "extract", "search", "monitor", "crawl"] as const;
const VALID = SERVICES.join(", ");
const LOAD_MORE = "__load_more__";

function entryUrl(e: ApiHistoryEntry): string {
function entryUrl(e: HistoryEntry): string {
const params = e.params as Record<string, unknown>;
return String(params.url ?? params.query ?? "");
}

function entryLabel(e: ApiHistoryEntry): string {
function entryLabel(e: HistoryEntry): string {
const short = e.id.length > 12 ? `${e.id.slice(0, 12)}…` : e.id;
const url = entryUrl(e);
const urlShort = url.length > 50 ? `${url.slice(0, 49)}…` : url;
Expand All @@ -24,7 +25,7 @@ function entryLabel(e: ApiHistoryEntry): string {
return `${chalk.dim(short)} ${color(e.status)} ${urlShort}`;
}

function entryHint(e: ApiHistoryEntry): string {
function entryHint(e: HistoryEntry): string {
if (!e.createdAt) return "";
const d = new Date(e.createdAt);
return Number.isNaN(d.getTime()) ? e.createdAt : d.toLocaleString();
Expand All @@ -49,11 +50,14 @@ export default defineCommand({
const quiet = !!args.json;
const out = log.create(quiet);
const apiKey = await resolveApiKey(quiet);
const service = args.service as ApiHistoryService | undefined;
if (service && !SERVICES.includes(service)) out.error(`Invalid service. Valid: ${VALID}`);
const rawService = args.service;
if (rawService && !SERVICES.includes(rawService as Service)) {
out.error(`Invalid service. Valid: ${VALID}`);
}
const service = rawService as Service | undefined;
const requestId = (args as { _: string[] })._.at(1);
const limit = args["page-size"] ? Number(args["page-size"]) : 20;
let page = args.page ? Number(args.page) : 1;
const limit = args["page-size"] ? parseIntArg(args["page-size"], "page-size", out) : 20;
let page = args.page ? parseIntArg(args.page, "page", out) : 1;

const fetchPage = async (pg: number) => {
const r = await history.list(apiKey, {
Expand All @@ -62,7 +66,7 @@ export default defineCommand({
...(service ? { service } : {}),
});
if (!r.data) out.error(r.error);
const d = r.data as { data: ApiHistoryEntry[]; pagination: { total: number } };
const d = r.data as { data: HistoryEntry[]; pagination: { total: number } };
return {
rows: d.data ?? [],
hasMore: (d.pagination?.total ?? 0) > pg * limit,
Expand Down
68 changes: 31 additions & 37 deletions src/commands/monitor.ts
Original file line number Diff line number Diff line change
Expand Up @@ -3,12 +3,15 @@ import chalk from "chalk";
import { defineCommand } from "citty";
import { monitor } from "scrapegraph-js";
import type {
ApiMonitorCreateInput,
ApiMonitorUpdateInput,
ApiScrapeFormatEntry,
FetchConfig,
FormatConfig,
MonitorCreateRequest,
MonitorUpdateRequest,
} from "scrapegraph-js";
import { resolveApiKey } from "../lib/folders.js";
import { BASE_FORMATS, type BaseFormat, buildBaseFormat } from "../lib/formats.js";
import * as log from "../lib/log.js";
import { parseIntArg } from "../lib/parse.js";

const ACTIONS = [
"create",
Expand All @@ -22,24 +25,14 @@ const ACTIONS = [
] as const;
type Action = (typeof ACTIONS)[number];

const FORMATS = [
"markdown",
"html",
"screenshot",
"branding",
"links",
"images",
"summary",
] as const;

function buildFormats(raw: string): ApiScrapeFormatEntry[] {
function buildFormats(raw: string, onInvalid: (f: string) => never): FormatConfig[] {
return raw
.split(",")
.map((f) => f.trim())
.filter(Boolean)
.map((f) => {
if (f === "markdown" || f === "html") return { type: f, mode: "normal" as const };
return { type: f } as ApiScrapeFormatEntry;
if (!BASE_FORMATS.includes(f as BaseFormat)) onInvalid(f);
return buildBaseFormat(f as BaseFormat);
});
}

Expand All @@ -64,7 +57,7 @@ export default defineCommand({
format: {
type: "string",
alias: "f",
description: `Formats to track, comma-separated: ${FORMATS.join(", ")} (default: markdown)`,
description: `Formats to track, comma-separated: ${BASE_FORMATS.join(", ")} (default: markdown)`,
},
"webhook-url": { type: "string", description: "Webhook URL for change notifications" },
mode: { type: "string", alias: "m", description: "Fetch mode: auto (default), fast, js" },
Expand All @@ -81,33 +74,33 @@ export default defineCommand({

if (!ACTIONS.includes(action)) {
out.error(`Unknown action: ${action}. Valid: ${ACTIONS.join(", ")}`);
return;
}

const needsId: Action[] = ["get", "update", "delete", "pause", "resume", "activity"];
if (needsId.includes(action) && !args.id) {
out.error(`--id is required for ${action}`);
return;
}

const fetchConfig: Record<string, unknown> = {};
if (args.mode) fetchConfig.mode = args.mode;
if (args.stealth) fetchConfig.stealth = true;

const onInvalidFormat = (f: string): never =>
out.error(`Unknown format: ${f}. Valid: ${BASE_FORMATS.join(", ")}`);

switch (action) {
case "create": {
if (!args.url) return out.error("--url is required for create");
if (!args.interval) return out.error("--interval is required for create");
if (!args.url) out.error("--url is required for create");
if (!args.interval) out.error("--interval is required for create");

const params: ApiMonitorCreateInput = {
const params: MonitorCreateRequest = {
url: args.url,
interval: args.interval,
formats: buildFormats(args.format ?? "markdown"),
formats: buildFormats(args.format ?? "markdown", onInvalidFormat),
...(args.name && { name: args.name }),
...(args["webhook-url"] && { webhookUrl: args["webhook-url"] }),
...(Object.keys(fetchConfig).length > 0 && { fetchConfig: fetchConfig as FetchConfig }),
};
const mut = params as Record<string, unknown>;
if (args.name) mut.name = args.name;
if (args["webhook-url"]) mut.webhookUrl = args["webhook-url"];
if (Object.keys(fetchConfig).length > 0) mut.fetchConfig = fetchConfig;

out.start("Creating monitor");
const result = await monitor.create(apiKey, params);
Expand Down Expand Up @@ -146,13 +139,13 @@ export default defineCommand({
}

case "update": {
const params: ApiMonitorUpdateInput = {};
const mut = params as Record<string, unknown>;
if (args.name) mut.name = args.name;
if (args.interval) mut.interval = args.interval;
if (args["webhook-url"]) mut.webhookUrl = args["webhook-url"];
if (args.format) mut.formats = buildFormats(args.format);
if (Object.keys(fetchConfig).length > 0) mut.fetchConfig = fetchConfig;
const params: MonitorUpdateRequest = {
...(args.name && { name: args.name }),
...(args.interval && { interval: args.interval }),
...(args["webhook-url"] && { webhookUrl: args["webhook-url"] }),
...(args.format && { formats: buildFormats(args.format, onInvalidFormat) }),
...(Object.keys(fetchConfig).length > 0 && { fetchConfig: fetchConfig as FetchConfig }),
};

out.start("Updating monitor");
const result = await monitor.update(apiKey, args.id as string, params);
Expand Down Expand Up @@ -190,9 +183,10 @@ export default defineCommand({
}

case "activity": {
const qp: { limit?: number; cursor?: string } = {};
if (args.limit) qp.limit = Number(args.limit);
if (args.cursor) qp.cursor = args.cursor;
const qp: { limit?: number; cursor?: string } = {
...(args.limit && { limit: parseIntArg(args.limit, "limit", out) }),
...(args.cursor && { cursor: args.cursor }),
};

out.start("Fetching monitor activity");
const result = await monitor.activity(apiKey, args.id as string, qp);
Expand Down
Loading
Loading