feat: migrate scrapegraph-js from github pin to npm v2.1.0

[Vikrant-Khedkar]

Summary

• Replaces broken github commit pin (ScrapeGraphAI/scrapegraph-js#096c110) with published npm package ^2.1.0
• The old commit had no dist/ and used the wrong base URL (api.scrapegraphai.com/api/v2 → 403), making the CLI broken out-of-the-box
• Updates all Api* type imports to match renamed exports in v2.1.0 (ApiScrapeRequest → ScrapeRequest, ApiHistoryEntry → HistoryEntry, etc.)

Test plan

• credits — 200, correct balance returned
• validate — {"status":"ok"}
• scrape — markdown returned for example.com
• search — results returned
• extract — structured JSON returned
• history — full history returned
• tsc --noEmit — zero errors

🤖 Generated with Claude Code

[VinciGit00]

CI/CD is broken, fix it

[VinciGit00]

Test results — commit ef5a1a8

Ran against the live API (v2.1.0 SDK). All 20 test cases pass; golden paths unchanged and error paths now emit friendly messages instead of raw SyntaxError / silent NaN.

Golden paths

#	Command	Result
1	validate	200 {status: ok}
2	credits	balance + job quotas returned
3	scrape -f markdown,html	both payloads returned
4	scrape -f links	leaf FormatConfig variant (no extras)
5	scrape -f summary	leaf variant
6	scrape -f json -p … --schema …	structured JSON
7	scrape --html-mode reader	mode propagated (history confirms "mode":"reader")
8	extract --schema …	typed JSON back
9	search --num-results 2 --country us	results returned
10	search --time-range past_week	filter applied (history confirms "timeRange":"past_week")
11	history scrape	service filter works
12	crawl --max-pages 2 --max-depth 1	completed; params echoed back correctly

Error paths (all new friendly messages)

#	Input	Output
13	search --num-results abc	--num-results: expected a number, got "abc"
14	extract --schema "not json"	--schema: invalid JSON (JSON Parse error: …)
15	crawl --max-pages xyz	--max-pages: expected a number, got "xyz"
16	crawl --include-patterns 'not-valid-json'	--include-patterns: invalid JSON (…)
17	scrape -f bogusformat	Unknown format: bogusformat. Valid: …
18	scrape -f json (no --prompt)	--prompt is required when format includes json
19	monitor update (no --id)	--id is required for update
20	monitor bogus	Unknown action: bogus. Valid: …

Not exercised

monitor create/pause/resume/delete/get/activity — free plan has monitor.limit = 1 so no live monitor was created. The code typechecks cleanly and all error branches (missing --url, missing --id, unknown action) pass.

Build

• tsc --noEmit — zero errors
• biome check — clean

[Vikrant-Khedkar]

CI/CD is broken, fix it

fixing it I thought it was breaking before also ,

[Vikrant-Khedkar]

I Fixed the CI/CD please check now @VinciGit00

[VinciGit00]

Re-test results — all green ✅

Re-ran the full test plan on feat/migrate-scrapegraph-js-npm @ 636774d with bun install + bun run src/cli.ts <cmd> on Node 22 / Bun 1.3.9. scrapegraph-js@2.1.0 resolved cleanly from npm (no github pin, no in-place build).

Check	Result
tsc --noEmit	0 errors
credits	200 · remaining: 407, plan Free Plan
validate	{"status":"ok","uptime":93321}
scrape https://example.com	markdown returned (id af038d53…), 68ms
search "what is scrapegraphai" --num-results 2	2 results (id 770f1e81…), 1.3s
extract https://example.com -p "Extract the page title and main heading"	{"page_title":"Example Domain","main_heading":"Example Domain"} (id 5a996fe5…), 688ms
history --page-size 3 --json	all 3 prior requests returned, correct renamed fields (requestParentId, elapsedMs, createdAt)

Credits delta: 407 → 397 (10 credits consumed across the run), confirming each request was actually billed server-side (not cached / stubbed).

The type rename surface (ApiScrapeRequest → ScrapeRequest, ApiHistoryEntry → HistoryEntry, etc.) compiles cleanly and the runtime shapes match the renamed types — history in particular returns the new camelCase fields without any mapping glue. No 403s, no base-URL issues — the published v2.1.0 package is wired correctly.

LGTM from my side. 🚀