fix(scrapeURL/fire-engine): default to separate US-generic proxy list…

… if no location is specified (FIR-728) (#1104) * feat(location/country): default to us-generic * add tests + fix mock
mendableai · Jan 29, 2025 · 5733b82 · 5733b82
1 parent 5c1b675
commit 5733b82
Show file tree

Hide file tree

Showing 5 changed files with 36 additions and 12 deletions.
diff --git a/apps/api/src/__tests__/snips/mocks/mocking-works-properly.json b/apps/api/src/__tests__/snips/mocks/mocking-works-properly.json
@@ -2,7 +2,7 @@
     {
         "time": 1735911273239,
         "options": {
-            "url": "http://default-fire-engine-api-service:8080/scrape",
+            "url": "<fire-engine>/scrape",
             "method": "POST",
             "body": {
                 "url": "http://firecrawl.dev",
@@ -27,7 +27,7 @@
     {
         "time": 1735911273354,
         "options": {
-            "url": "http://default-fire-engine-api-service:8080/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
+            "url": "<fire-engine>/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
             "method": "GET",
             "headers": {},
             "ignoreResponse": false,
@@ -43,7 +43,7 @@
     {
         "time": 1735911273720,
         "options": {
-            "url": "http://default-fire-engine-api-service:8080/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
+            "url": "<fire-engine>/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
             "method": "GET",
             "headers": {},
             "ignoreResponse": false,
@@ -59,7 +59,7 @@
     {
         "time": 1735911274092,
         "options": {
-            "url": "http://default-fire-engine-api-service:8080/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
+            "url": "<fire-engine>/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
             "method": "GET",
             "headers": {},
             "ignoreResponse": false,
@@ -75,7 +75,7 @@
     {
         "time": 1735911274467,
         "options": {
-            "url": "http://default-fire-engine-api-service:8080/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
+            "url": "<fire-engine>/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
             "method": "GET",
             "headers": {},
             "ignoreResponse": false,
@@ -91,7 +91,7 @@
     {
         "time": 1735911274947,
         "options": {
-            "url": "http://default-fire-engine-api-service:8080/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
+            "url": "<fire-engine>/scrape/ede37286-90db-4f60-8efb-76217dfcfa35!chrome-cdp",
             "method": "GET",
             "headers": {},
             "ignoreResponse": false,

diff --git a/apps/api/src/__tests__/snips/scrape.test.ts b/apps/api/src/__tests__/snips/scrape.test.ts
@@ -35,4 +35,24 @@ describe("Scrape tests", () => {
       "this is fake data coming from the mocking system!",
     );
   });
+
+  describe("Location API", () => {
+    it.concurrent("works without specifying an explicit location", async () => {
+      const response = await scrape({
+        url: "https://iplocation.com",
+      });
+
+      expectScrapeToSucceed(response);
+    });
+
+    it.concurrent("works with country US", async () => {
+      const response = await scrape({
+        url: "https://iplocation.com",
+        location: { country: "US" },
+      });
+
+      expectScrapeToSucceed(response);
+      expect(response.body.data.markdown).toContain("| Country | United States |");
+    });
+  })
 });
diff --git a/apps/api/src/controllers/v1/types.ts b/apps/api/src/controllers/v1/types.ts
@@ -154,13 +154,13 @@ export const scrapeOptions = z
           .string()
           .optional()
           .refine(
-            (val) => !val || Object.keys(countries).includes(val.toUpperCase()),
+            (val) => !val || Object.keys(countries).includes(val.toUpperCase()) || val === "US-generic",
             {
               message:
                 "Invalid country code. Please use a valid ISO 3166-1 alpha-2 country code.",
             },
           )
-          .transform((val) => (val ? val.toUpperCase() : "US")),
+          .transform((val) => (val ? val.toUpperCase() : "US-generic")),
         languages: z.string().array().optional(),
       })
       .optional(),
@@ -178,7 +178,7 @@ export const scrapeOptions = z
                 "Invalid country code. Please use a valid ISO 3166-1 alpha-2 country code.",
             },
           )
-          .transform((val) => (val ? val.toUpperCase() : "US")),
+          .transform((val) => (val ? val.toUpperCase() : "US-generic")),
         languages: z.string().array().optional(),
       })
       .optional(),

diff --git a/apps/api/src/scraper/scrapeURL/lib/fetch.ts b/apps/api/src/scraper/scrapeURL/lib/fetch.ts
@@ -126,10 +126,14 @@ export async function robustFetch<
     const makeRequestTypeId = (
       request: (typeof mock)["requests"][number]["options"],
     ) => {
-      let out = request.url + ";" + request.method;
+      let trueUrl = (process.env.FIRE_ENGINE_BETA_URL && request.url.startsWith(process.env.FIRE_ENGINE_BETA_URL))
+        ? request.url.replace(process.env.FIRE_ENGINE_BETA_URL, "<fire-engine>")
+        : request.url;
+
+      let out = trueUrl + ";" + request.method;
       if (
         process.env.FIRE_ENGINE_BETA_URL &&
-        url.startsWith(process.env.FIRE_ENGINE_BETA_URL) &&
+        (trueUrl.startsWith("<fire-engine>")) &&
         request.method === "POST"
       ) {
         out += "f-e;" + request.body?.engine + ";" + request.body?.url;

diff --git a/apps/api/src/scraper/scrapeURL/lib/mock.ts b/apps/api/src/scraper/scrapeURL/lib/mock.ts
@@ -3,7 +3,7 @@ import * as path from "path";
 import { logger as _logger } from "../../../lib/logger";
 import { Logger } from "winston";
 const saveMocksDirPath = path.join(__dirname, "../mocks/").replace("dist/", "");
-const loadMocksDirPath = path.join(__dirname, "../../../__tests__/snips/mocks");
+const loadMocksDirPath = path.join(__dirname, "../../../__tests__/snips/mocks").replace("dist/", "");
 
 export async function saveMock(options: unknown, result: unknown) {
   if (process.env.FIRECRAWL_SAVE_MOCKS !== "true") return;