Merge pull request #13 from scrapfly/extraction-scrape-config

Granitosaurus · web-flow · commit 29a99266c22e · 2024-11-09T13:19:49.000+07:00
add scrape extraction params and rename extraction template options
diff --git a/__tests__/client/extraction.test.ts b/__tests__/client/extraction.test.ts
@@ -39,8 +39,8 @@ Deno.test('extract: fails due to invalid config', async () => {
                 new ExtractionConfig({
                     body: html,
                     content_type: 'text/html',
-                    ephemeral_template: { source: 'html' },
-                    template: 'template',
+                    extraction_ephemeral_template: { source: 'html' },
+                    extraction_template: 'template',
                 }),
             );
         },
diff --git a/__tests__/config/extraction.test.ts b/__tests__/config/extraction.test.ts
@@ -58,11 +58,11 @@ Deno.test('url param generation: sets charset', async () => {
     });
 });
 
-Deno.test('url param generation: sets template', async () => {
+Deno.test('url param generation: sets extraction_template', async () => {
     const config = new ExtractionConfig({
         body: input_html,
         content_type: input_content_type,
-        template: 'my_template',
+        extraction_template: 'my_template',
     });
     const params = config.toApiParams({ key: '1234' });
     assertEquals(params, {
@@ -72,11 +72,11 @@ Deno.test('url param generation: sets template', async () => {
     });
 });
 
-Deno.test('url param generation: sets ephemeral_template', async () => {
+Deno.test('url param generation: sets extraction_ephemeral_template', async () => {
     const config = new ExtractionConfig({
         body: input_html,
         content_type: input_content_type,
-        ephemeral_template: { source: 'html', selectors: [] },
+        extraction_ephemeral_template: { source: 'html', selectors: [] },
     });
     const params = config.toApiParams({ key: '1234' });
     assertEquals(params, {
diff --git a/__tests__/config/scrape.test.ts b/__tests__/config/scrape.test.ts
@@ -3,6 +3,7 @@ import { HttpMethod } from '../../src/types.ts';
 import { ScrapeConfigError } from '../../src/errors.ts';
 import { assertEquals, assertThrows } from "https://deno.land/std@0.224.0/assert/mod.ts";
 
+const input_content_type = 'text/html';
 
 Deno.test('scrapeconfig loads', () => {
     const config = new ScrapeConfig({ url: 'http://httpbin.dev/get' });
@@ -15,8 +16,6 @@ Deno.test('scrapeconfig throws on unknown options', () => {
     }, ScrapeConfigError, "Invalid option provided: foobar");
 });
 
-
-
 Deno.test('scrapeconfig allowed methods', () => {
     (['GET', 'POST', 'PUT', 'PATCH', 'HEAD'] as HttpMethod[]).forEach((method) => {
         const config = new ScrapeConfig({
@@ -360,6 +359,58 @@ Deno.test('url param generation: proxy_pool sets', () => {
     });
 });
 
+Deno.test('url param generation: sets extraction_template', async () => {
+    const config = new ScrapeConfig({
+        url: 'http://httpbin.dev/get',
+        extraction_template: 'my_template',
+    });
+    const params = config.toApiParams({ key: '1234' });
+    assertEquals(params, {
+        key: '1234',
+        url: 'http://httpbin.dev/get',
+        extraction_template: 'my_template',
+    });
+});
+
+Deno.test('url param generation: sets extraction_ephemeral_template', async () => {
+    const config = new ScrapeConfig({
+        url: 'http://httpbin.dev/get',
+        extraction_ephemeral_template: { source: 'html', selectors: [] },
+    });
+    const params = config.toApiParams({ key: '1234' });
+    assertEquals(params, {
+        key: '1234',
+        url: 'http://httpbin.dev/get',
+        extraction_template: 'ephemeral:eyJzb3VyY2UiOiJodG1sIiwic2VsZWN0b3JzIjpbXX0',
+    });
+});
+
+Deno.test('url param generation: sets extraction_prompt', async () => {
+    const config = new ScrapeConfig({
+        url: 'http://httpbin.dev/get',
+        extraction_prompt: 'summarize the document',
+    });
+    const params = config.toApiParams({ key: '1234' });
+    assertEquals(params, {
+        key: '1234',
+        url: 'http://httpbin.dev/get',
+        extraction_prompt: 'summarize the document',
+    });
+});
+
+Deno.test('url param generation: sets extraction_model', async () => {
+    const config = new ScrapeConfig({
+        url: 'http://httpbin.dev/get',
+        extraction_model: 'review_list',
+    });
+    const params = config.toApiParams({ key: '1234' });
+    assertEquals(params, {
+        key: '1234',
+        url: 'http://httpbin.dev/get',
+        extraction_model: 'review_list',
+    });
+});
+
 Deno.test('url param generation: session sets', () => {
     const config = new ScrapeConfig({
         url: 'http://httpbin.dev/get',
diff --git a/examples/deno/deno_examples.ts b/examples/deno/deno_examples.ts
@@ -81,6 +81,28 @@ export async function JSRender(apiKey: string) {
   console.log(scrape_result.result.browser_data);
 }
 
+/* Use AI extraction capabilities with the the web scraping API
+ * all Extraction API methods are supported, see below examples for more
+ */
+export async function scrapeExtraction(apiKey: string) {
+  const client = new ScrapflyClient({ key: apiKey});
+
+  let scrape_result = await client.scrape(
+    new ScrapeConfig({
+      url: 'https://web-scraping.dev/product/1',
+      // enable browsers:
+      render_js: true,
+      // use LLM prompt for auto parsing
+      extraction_prompt: "Extract the product specification in json format",
+    })
+  );
+
+  // access the extraction result
+  console.log("extraction result:");
+  console.log(scrape_result.result.extracted_data);
+}
+
+
 /* Scrapfly Extraction API offers LLM (Language Learning Model) based extraction
  * This example demonstrates how to use LLM query HTML files
  * https://scrapfly.io/docs/extraction-api/llm-prompt
@@ -190,7 +212,7 @@ export async function extractionTemplates(apiKey: string){
       body: html,
       content_type: "text/html",
       // provide template:
-      ephemeral_template: template,
+      extraction_ephemeral_template: template,
     })
   );
   console.log('product  extract');
diff --git a/examples/node_commonjs/commonjs_examples.cjs b/examples/node_commonjs/commonjs_examples.cjs
@@ -81,6 +81,27 @@ async function JSRender(apiKey) {
   console.log(scrape_result.result.browser_data);
 }
 
+/* Use AI extraction capabilities with the the web scraping API
+ * all Extraction API methods are supported, see below examples for more
+ */
+async function scrapeExtraction(apiKey) {
+  const client = new ScrapflyClient({ key: apiKey});
+
+  let scrape_result = await client.scrape(
+    new ScrapeConfig({
+      url: 'https://web-scraping.dev/product/1',
+      // enable browsers:
+      render_js: true,
+      // use LLM prompt for auto parsing
+      extraction_prompt: "Extract the product specification in json format",
+    })
+  );
+
+  // access the extraction result
+  console.log("extraction result:");
+  console.log(scrape_result.result.extracted_data);
+}
+
 /* Scrapfly Extraction API offers LLM (Language Learning Model) based extraction
  * This example demonstrates how to use LLM query HTML files
  * https://scrapfly.io/docs/extraction-api/llm-prompt
@@ -190,7 +211,7 @@ async function extractionTemplates(apiKey){
       body: html,
       content_type: "text/html",
       // provide template:
-      ephemeral_template: template,
+      extraction_ephemeral_template: template,
     })
   );
   console.log('product  extract');
@@ -233,6 +254,7 @@ module.exports = {
   getAccount,
   basicGet,
   JSRender,
+  scrapeExtraction,
   extractionLLM,
   extractionAutoExtract,
   extractionTemplates,
diff --git a/src/extractionconfig.ts b/src/extractionconfig.ts
@@ -1,5 +1,6 @@
 import * as errors from './errors.ts';
 import { urlsafe_b64encode } from './utils.ts';
+import { ExtractionConfigError } from './errors.ts';
 
 export enum CompressionFormat {
   /**
@@ -21,30 +22,55 @@ type ExtractionConfigOptions = {
   content_type: string;
   url?: string;
   charset?: string;
-  template?: string; // saved template name
-  ephemeral_template?: object; // ephemeraly declared json template
+  extraction_template?: string; // saved template name
+  extraction_ephemeral_template?: object; // ephemeraly declared json template
   extraction_prompt?: string;
   extraction_model?: string;
   is_document_compressed?: boolean;
   document_compression_format?: 'gzip' | 'zstd' | 'deflate' | CompressionFormat;
   webhook?: string;
+
+  // deprecated options
+  template?: string;
+  ephemeral_template?: object;
 };
 
 export class ExtractionConfig {
   body: string | Uint8Array;
   content_type: string;
   url?: string;
   charset?: string;
-  template?: string; // saved template name
-  ephemeral_template?: object; // ephemeraly declared json template
+  extraction_template?: string; // saved template name
+  extraction_ephemeral_template?: object; // ephemeraly declared json template
   extraction_prompt?: string;
   extraction_model?: string;
   is_document_compressed?: boolean;
   document_compression_format?: 'gzip' | 'zstd' | 'deflate' | CompressionFormat;
   webhook?: string;
 
+  // // deprecated options
+  template?: string;
+  ephemeral_template?: object;
+
   constructor(options: ExtractionConfigOptions) {
     this.validateOptions(options);
+    if (options.template) {
+      console.warn(
+        `Deprecation warning: 'template' is deprecated. Use 'extraction_template' instead.`
+      );
+      this.extraction_template = options.template;
+    } else {
+      this.extraction_template = options.extraction_template;
+    }
+    if (options.ephemeral_template) {
+      console.warn(
+        `Deprecation warning: 'ephemeral_template' is deprecated. Use 'extraction_ephemeral_template' instead.`
+      );
+      this.extraction_ephemeral_template = options.ephemeral_template;
+    } else {
+      this.extraction_ephemeral_template = options.extraction_ephemeral_template;
+    }
+
     if (
       options.document_compression_format &&
       !Object.values(CompressionFormat).includes(options.document_compression_format as CompressionFormat)
@@ -57,8 +83,8 @@ export class ExtractionConfig {
     this.content_type = options.content_type;
     this.url = options.url ?? this.url;
     this.charset = options.charset ?? this.charset;
-    this.template = options.template ?? this.template;
-    this.ephemeral_template = options.ephemeral_template ?? this.ephemeral_template;
+    this.extraction_template = options.extraction_template ?? this.extraction_template;
+    this.extraction_ephemeral_template = options.extraction_ephemeral_template ?? this.extraction_ephemeral_template;
     this.extraction_prompt = options.extraction_prompt ?? this.extraction_prompt;
     this.extraction_model = options.extraction_model ?? this.extraction_model;
     this.is_document_compressed = options.is_document_compressed ?? this.is_document_compressed;
@@ -90,18 +116,18 @@ export class ExtractionConfig {
       params.charset = this.charset;
     }
 
-    if (this.template && this.ephemeral_template) {
-      throw new errors.ExtractionConfigError(
-        'You cannot pass both parameters template and ephemeral_template. You must choose',
+    if (this.extraction_template && this.extraction_ephemeral_template) {
+      throw new ExtractionConfigError(
+        'You cannot pass both parameters extraction_template and extraction_ephemeral_template. You must choose',
       );
     }
 
-    if (this.template) {
-      params.extraction_template = this.template;
+    if (this.extraction_template) {
+      params.extraction_template = this.extraction_template;
     }
 
-    if (this.ephemeral_template) {
-      params.extraction_template = 'ephemeral:' + urlsafe_b64encode(JSON.stringify(this.ephemeral_template));
+    if (this.extraction_ephemeral_template) {
+      params.extraction_template = 'ephemeral:' + urlsafe_b64encode(JSON.stringify(this.extraction_ephemeral_template));
     }
 
     if (this.extraction_prompt) {
diff --git a/src/scrapeconfig.ts b/src/scrapeconfig.ts
@@ -61,6 +61,10 @@ type ScrapeConfigOptions = {
   tags?: string[];
   format?: 'json' | 'text' | 'markdown' | 'clean_html' | 'raw' | Format;
   format_options?: ('no_links' | 'no_images' | 'only_content' | FormatOption)[];
+  extraction_template?: string; // saved template name
+  extraction_ephemeral_template?: object; // ephemeraly declared json template
+  extraction_prompt?: string;
+  extraction_model?: string;  
   correlation_id?: string;
   cookies?: Rec<string>;
   body?: string;
@@ -104,6 +108,10 @@ export class ScrapeConfig {
   tags: Set<string> = new Set<string>();
   format?: 'json' | 'text' | 'markdown' | 'clean_html' | 'raw' | Format;
   format_options?: ('no_links' | 'no_images' | 'only_content' | FormatOption)[];
+  extraction_template?: string; // saved template name
+  extraction_ephemeral_template?: object; // ephemeraly declared json template
+  extraction_prompt?: string;
+  extraction_model?: string;  
   correlation_id?: string;
   cookies?: Rec<string>;
   body?: string;
@@ -163,6 +171,10 @@ export class ScrapeConfig {
     this.tags = new Set(options.tags) ?? this.tags;
     this.format = options.format ?? this.format;
     this.format_options = options.format_options ?? this.format_options;
+    this.extraction_template = options.extraction_template ?? this.extraction_template;
+    this.extraction_ephemeral_template = options.extraction_ephemeral_template ?? this.extraction_ephemeral_template;
+    this.extraction_prompt = options.extraction_prompt ?? this.extraction_prompt;
+    this.extraction_model = options.extraction_model ?? this.extraction_model;    
     this.correlation_id = options.correlation_id ?? this.correlation_id;
     this.cookies = options.cookies
       ? Object.fromEntries(Object.entries(options.cookies).map(([k, v]) => [k.toLowerCase(), v]))
@@ -338,6 +350,27 @@ export class ScrapeConfig {
         params.format += ':' + this.format_options.join(',');
       }
     }
+    if (this.extraction_template && this.extraction_ephemeral_template) {
+      throw new ScrapeConfigError(
+        'You cannot pass both parameters extraction_template and extraction_ephemeral_template. You must choose',
+      );
+    }
+
+    if (this.extraction_template) {
+      params.extraction_template = this.extraction_template;
+    }
+
+    if (this.extraction_ephemeral_template) {
+      params.extraction_template = 'ephemeral:' + urlsafe_b64encode(JSON.stringify(this.extraction_ephemeral_template));
+    }
+
+    if (this.extraction_prompt) {
+      params.extraction_prompt = this.extraction_prompt;
+    }
+
+    if (this.extraction_model) {
+      params.extraction_model = this.extraction_model;
+    }
     if (this.correlation_id) {
       params.correlation_id = this.correlation_id;
     }
diff --git a/tsconfig.json b/tsconfig.json
@@ -5,7 +5,7 @@
         "types": ["node", "jest"],
         "skipLibCheck": true,
         "module": "ESNext",
-        "lib": ["ES2022", "ESNext"],
+        "lib": ["ES2022", "ESNext", "dom"],
         "moduleResolution": "Node",
         "rootDir": ".",
         "outDir": "build",
