fix(search): align PDF options with API

Author: VinciGit00

README.md

@@ -155,19 +155,14 @@ const res = await sgai.search({
   timeRange: "past_week",          // optional
   locationGeoCode: "us",           // optional
   fetchConfig: { /* ... */ },      // optional
-  contentTypes: [                  // optional, defaults to text-like types
-    "text/html",
-    "application/json",
-    "text/markdown",
-    "text/plain",
-  ],
+  allowedTypes: ["text/html", "application/pdf"], // optional MIME allowlist
+  processors: [{ type: "pdf", maxPages: 10 }],    // optional PDF page cap
 });
 ```
 
-By default `search` only scrapes text-like results (`text/html`, `application/json`,
-`text/markdown`, `text/plain`); PDFs, office documents and images are skipped. Pass
-`contentTypes` to change that, for example `contentTypes: ["text/html", "application/pdf"]`
-to include PDFs.
+By default `search` accepts every supported content type, including PDFs, and processes up to 25
+pages per PDF. Use `allowedTypes` to restrict accepted MIME types. Configure PDF processing
+separately with `processors`; `maxPages` accepts `1`–`500`, or `-1` for no page limit.
 
 ### crawl
 

src/schemas.ts

@@ -30,6 +30,10 @@ export const fetchContentTypeSchema = z.enum([
 	"application/rtf",
 	"application/vnd.oasis.opendocument.text",
 ]);
+export const pdfProcessorSchema = z.object({
+	type: z.literal("pdf"),
+	maxPages: z.union([z.literal(-1), z.number().int().min(1).max(500)]).default(25),
+});
 export const userPromptSchema = z.string().min(1).max(10_000);
 
 const PUBLIC_DOMAIN_RE =
@@ -250,7 +254,8 @@ export const searchRequestSchema = z
 		prompt: userPromptSchema.optional(),
 		schema: z.record(z.string(), z.unknown()).optional(),
 		locationGeoCode: z.string().max(10).optional(),
-		contentTypes: z.array(fetchContentTypeSchema).optional(),
+		allowedTypes: z.array(fetchContentTypeSchema).min(1).optional(),
+		processors: z.array(pdfProcessorSchema).min(1).optional(),
 		timeRange: z
 			.enum(["past_hour", "past_24_hours", "past_week", "past_month", "past_year"])
 			.optional(),

tests/scrapegraphai.test.ts

@@ -710,6 +710,25 @@ describe("search", () => {
 		expect(res.status).toBe("success");
 		expectRequest(0, "POST", "/search", searchParams);
 	});
+
+	test("with PDF options", async () => {
+		const body = {
+			results: [],
+			metadata: { search: {}, pages: { requested: 1, scraped: 0 } },
+		};
+		fetchSpy = spyOn(globalThis, "fetch").mockResolvedValueOnce(json(body));
+		const searchParams = {
+			query: "papers",
+			numResults: 1,
+			allowedTypes: ["application/pdf"] as const,
+			processors: [{ type: "pdf" as const, maxPages: 10 }],
+		};
+
+		const res = await sdk.search(API_KEY, searchParams);
+
+		expect(res.status).toBe("success");
+		expectRequest(0, "POST", "/search", searchParams);
+	});
 });
 
 describe("getCredits", () => {