n8n-nodes-crawl4ai-plus 2.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +23 -0
- package/LICENSE.md +19 -0
- package/README.md +157 -0
- package/dist/credentials/Crawl4aiApi.credentials.d.ts +7 -0
- package/dist/credentials/Crawl4aiApi.credentials.js +242 -0
- package/dist/credentials/Crawl4aiApi.credentials.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/Crawl4aiBasicCrawler.node.d.ts +5 -0
- package/dist/nodes/Crawl4aiBasicCrawler/Crawl4aiBasicCrawler.node.js +37 -0
- package/dist/nodes/Crawl4aiBasicCrawler/Crawl4aiBasicCrawler.node.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/crawlMultipleUrls.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/crawlMultipleUrls.operation.js +299 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/crawlMultipleUrls.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/crawlSingleUrl.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/crawlSingleUrl.operation.js +324 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/crawlSingleUrl.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/operations.d.ts +8 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/operations.js +67 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/operations.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/processRawHtml.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/processRawHtml.operation.js +148 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/processRawHtml.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/router.d.ts +2 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/router.js +37 -0
- package/dist/nodes/Crawl4aiBasicCrawler/actions/router.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/crawl4ai.svg +17 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/apiClient.d.ts +15 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/apiClient.js +226 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/apiClient.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/formatters.d.ts +5 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/formatters.js +81 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/formatters.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/interfaces.d.ts +189 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/interfaces.js +3 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/interfaces.js.map +1 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/utils.d.ts +8 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/utils.js +97 -0
- package/dist/nodes/Crawl4aiBasicCrawler/helpers/utils.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/Crawl4aiContentExtractor.node.d.ts +5 -0
- package/dist/nodes/Crawl4aiContentExtractor/Crawl4aiContentExtractor.node.js +38 -0
- package/dist/nodes/Crawl4aiContentExtractor/Crawl4aiContentExtractor.node.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/cssExtractor.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/cssExtractor.operation.js +336 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/cssExtractor.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/jsonExtractor.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/jsonExtractor.operation.js +369 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/jsonExtractor.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/llmExtractor.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/llmExtractor.operation.js +786 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/llmExtractor.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/operations.d.ts +8 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/operations.js +76 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/operations.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/regexExtractor.operation.d.ts +4 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/regexExtractor.operation.js +437 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/regexExtractor.operation.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/router.d.ts +2 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/router.js +37 -0
- package/dist/nodes/Crawl4aiContentExtractor/actions/router.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/crawl4ai.svg +17 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/apiClient.d.ts +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/apiClient.js +7 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/apiClient.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/formatters.d.ts +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/formatters.js +8 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/formatters.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/interfaces.d.ts +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/interfaces.js +3 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/interfaces.js.map +1 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/utils.d.ts +6 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/utils.js +89 -0
- package/dist/nodes/Crawl4aiContentExtractor/helpers/utils.js.map +1 -0
- package/dist/tsconfig.tsbuildinfo +1 -0
- package/index.js +14 -0
- package/package.json +70 -0
|
@@ -0,0 +1,189 @@
|
|
|
1
|
+
import { IDataObject } from 'n8n-workflow';
|
|
2
|
+
export interface Crawl4aiApiCredentials {
|
|
3
|
+
connectionMode: 'direct' | 'docker';
|
|
4
|
+
dockerUrl?: string;
|
|
5
|
+
authenticationType?: 'none' | 'token' | 'basic';
|
|
6
|
+
apiToken?: string;
|
|
7
|
+
username?: string;
|
|
8
|
+
password?: string;
|
|
9
|
+
enableLlm?: boolean;
|
|
10
|
+
llmProvider?: string;
|
|
11
|
+
apiKey?: string;
|
|
12
|
+
ollamaUrl?: string;
|
|
13
|
+
customProvider?: string;
|
|
14
|
+
customBaseUrl?: string;
|
|
15
|
+
customApiKey?: string;
|
|
16
|
+
cacheDir?: string;
|
|
17
|
+
}
|
|
18
|
+
export interface Crawl4aiNodeOptions extends IDataObject {
|
|
19
|
+
bypassCache?: boolean;
|
|
20
|
+
includeMedia?: boolean;
|
|
21
|
+
verboseResponse?: boolean;
|
|
22
|
+
cacheMode?: 'enabled' | 'bypass' | 'only';
|
|
23
|
+
}
|
|
24
|
+
export interface BrowserConfig {
|
|
25
|
+
browser_type?: string;
|
|
26
|
+
headless?: boolean;
|
|
27
|
+
browser_mode?: string;
|
|
28
|
+
use_managed_browser?: boolean;
|
|
29
|
+
cdp_url?: string;
|
|
30
|
+
debugging_port?: number;
|
|
31
|
+
use_persistent_context?: boolean;
|
|
32
|
+
user_data_dir?: string;
|
|
33
|
+
chrome_channel?: string;
|
|
34
|
+
channel?: string;
|
|
35
|
+
proxy?: string;
|
|
36
|
+
viewport_width?: number;
|
|
37
|
+
viewport_height?: number;
|
|
38
|
+
viewport?: {
|
|
39
|
+
width: number;
|
|
40
|
+
height: number;
|
|
41
|
+
};
|
|
42
|
+
accept_downloads?: boolean;
|
|
43
|
+
downloads_path?: string;
|
|
44
|
+
storage_state?: string | object;
|
|
45
|
+
ignore_https_errors?: boolean;
|
|
46
|
+
java_script_enabled?: boolean;
|
|
47
|
+
javaScriptEnabled?: boolean;
|
|
48
|
+
cookies?: Array<object>;
|
|
49
|
+
headers?: object;
|
|
50
|
+
user_agent?: string;
|
|
51
|
+
userAgent?: string;
|
|
52
|
+
user_agent_mode?: string;
|
|
53
|
+
user_agent_generator_config?: object;
|
|
54
|
+
text_mode?: boolean;
|
|
55
|
+
light_mode?: boolean;
|
|
56
|
+
extra_args?: Array<string>;
|
|
57
|
+
host?: string;
|
|
58
|
+
enable_stealth?: boolean;
|
|
59
|
+
}
|
|
60
|
+
export interface CrawlerRunConfig {
|
|
61
|
+
cacheMode?: 'enabled' | 'bypass' | 'only';
|
|
62
|
+
streamEnabled?: boolean;
|
|
63
|
+
pageTimeout?: number;
|
|
64
|
+
requestTimeout?: number;
|
|
65
|
+
timeout?: number;
|
|
66
|
+
waitUntil?: string;
|
|
67
|
+
waitFor?: string;
|
|
68
|
+
waitForTimeout?: number;
|
|
69
|
+
waitForImages?: boolean;
|
|
70
|
+
delayBeforeReturnHtml?: number;
|
|
71
|
+
jsCode?: string | string[];
|
|
72
|
+
jsOnly?: boolean;
|
|
73
|
+
ignoreBodyVisibility?: boolean;
|
|
74
|
+
scanFullPage?: boolean;
|
|
75
|
+
scrollDelay?: number;
|
|
76
|
+
maxScrollSteps?: number;
|
|
77
|
+
processIframes?: boolean;
|
|
78
|
+
removeOverlayElements?: boolean;
|
|
79
|
+
simulateUser?: boolean;
|
|
80
|
+
overrideNavigator?: boolean;
|
|
81
|
+
magic?: boolean;
|
|
82
|
+
adjustViewportToContent?: boolean;
|
|
83
|
+
cssSelector?: string;
|
|
84
|
+
targetElements?: string[];
|
|
85
|
+
excludedTags?: string[];
|
|
86
|
+
excludedSelector?: string;
|
|
87
|
+
keepDataAttributes?: boolean;
|
|
88
|
+
keepAttrs?: string[];
|
|
89
|
+
removeForms?: boolean;
|
|
90
|
+
onlyText?: boolean;
|
|
91
|
+
wordCountThreshold?: number;
|
|
92
|
+
excludeExternalLinks?: boolean;
|
|
93
|
+
excludeInternalLinks?: boolean;
|
|
94
|
+
excludeSocialMediaLinks?: boolean;
|
|
95
|
+
excludeDomains?: string[];
|
|
96
|
+
scoreLinks?: boolean;
|
|
97
|
+
screenshot?: boolean;
|
|
98
|
+
screenshotWaitFor?: number;
|
|
99
|
+
screenshotHeightThreshold?: number;
|
|
100
|
+
pdf?: boolean;
|
|
101
|
+
imageDescriptionMinWordThreshold?: number;
|
|
102
|
+
imageScoreThreshold?: number;
|
|
103
|
+
excludeExternalImages?: boolean;
|
|
104
|
+
excludeAllImages?: boolean;
|
|
105
|
+
tableScoreThreshold?: number;
|
|
106
|
+
checkRobotsTxt?: boolean;
|
|
107
|
+
userAgent?: string;
|
|
108
|
+
userAgentMode?: string;
|
|
109
|
+
userAgentGeneratorConfig?: object;
|
|
110
|
+
method?: string;
|
|
111
|
+
sessionId?: string;
|
|
112
|
+
sharedData?: object;
|
|
113
|
+
maxRetries?: number;
|
|
114
|
+
viewport?: {
|
|
115
|
+
width: number;
|
|
116
|
+
height: number;
|
|
117
|
+
};
|
|
118
|
+
headless?: boolean;
|
|
119
|
+
javaScriptEnabled?: boolean;
|
|
120
|
+
verbose?: boolean;
|
|
121
|
+
logConsole?: boolean;
|
|
122
|
+
captureNetworkRequests?: boolean;
|
|
123
|
+
captureConsoleMessages?: boolean;
|
|
124
|
+
extractionStrategy?: any;
|
|
125
|
+
chunkingStrategy?: any;
|
|
126
|
+
markdownGenerator?: any;
|
|
127
|
+
scrapingStrategy?: any;
|
|
128
|
+
proxyConfig?: object;
|
|
129
|
+
linkPreviewConfig?: object;
|
|
130
|
+
virtualScrollConfig?: object;
|
|
131
|
+
deepCrawlStrategy?: object;
|
|
132
|
+
experimental?: object;
|
|
133
|
+
}
|
|
134
|
+
export interface CrawlResult {
|
|
135
|
+
url: string;
|
|
136
|
+
success: boolean;
|
|
137
|
+
statusCode?: number;
|
|
138
|
+
status_code?: number;
|
|
139
|
+
title?: string;
|
|
140
|
+
markdown?: string;
|
|
141
|
+
html?: string;
|
|
142
|
+
cleaned_html?: string;
|
|
143
|
+
text?: string;
|
|
144
|
+
links?: {
|
|
145
|
+
internal: Link[];
|
|
146
|
+
external: Link[];
|
|
147
|
+
};
|
|
148
|
+
media?: {
|
|
149
|
+
images: Media[];
|
|
150
|
+
videos: Media[];
|
|
151
|
+
};
|
|
152
|
+
extracted_content?: string;
|
|
153
|
+
error_message?: string;
|
|
154
|
+
crawl_time?: number;
|
|
155
|
+
metadata?: Record<string, any>;
|
|
156
|
+
}
|
|
157
|
+
export interface Link {
|
|
158
|
+
href: string;
|
|
159
|
+
text: string;
|
|
160
|
+
title?: string;
|
|
161
|
+
}
|
|
162
|
+
export interface Media {
|
|
163
|
+
src: string;
|
|
164
|
+
alt?: string;
|
|
165
|
+
title?: string;
|
|
166
|
+
type?: string;
|
|
167
|
+
}
|
|
168
|
+
export interface CssSelectorField {
|
|
169
|
+
name: string;
|
|
170
|
+
selector: string;
|
|
171
|
+
type: 'text' | 'attribute' | 'html';
|
|
172
|
+
attribute?: string;
|
|
173
|
+
}
|
|
174
|
+
export interface CssSelectorSchema {
|
|
175
|
+
name: string;
|
|
176
|
+
baseSelector: string;
|
|
177
|
+
fields: CssSelectorField[];
|
|
178
|
+
}
|
|
179
|
+
export interface LlmSchemaField {
|
|
180
|
+
name: string;
|
|
181
|
+
type: string;
|
|
182
|
+
description?: string;
|
|
183
|
+
}
|
|
184
|
+
export interface LlmSchema {
|
|
185
|
+
title?: string;
|
|
186
|
+
type: 'object';
|
|
187
|
+
properties: Record<string, LlmSchemaField>;
|
|
188
|
+
required?: string[];
|
|
189
|
+
}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"interfaces.js","sourceRoot":"","sources":["../../../../nodes/Crawl4aiBasicCrawler/helpers/interfaces.ts"],"names":[],"mappings":""}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
import { IExecuteFunctions, IDataObject } from 'n8n-workflow';
|
|
2
|
+
import { BrowserConfig, CrawlerRunConfig } from './interfaces';
|
|
3
|
+
export declare function getCrawl4aiClient(executeFunctions: IExecuteFunctions): Promise<any>;
|
|
4
|
+
export declare function createBrowserConfig(options: IDataObject): BrowserConfig;
|
|
5
|
+
export declare function createCrawlerRunConfig(options: IDataObject): CrawlerRunConfig;
|
|
6
|
+
export declare function safeJsonParse(jsonString: string, defaultValue?: any): any;
|
|
7
|
+
export declare function cleanText(text: string): string;
|
|
8
|
+
export declare function isValidUrl(url: string): boolean;
|
|
@@ -0,0 +1,97 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.getCrawl4aiClient = getCrawl4aiClient;
|
|
4
|
+
exports.createBrowserConfig = createBrowserConfig;
|
|
5
|
+
exports.createCrawlerRunConfig = createCrawlerRunConfig;
|
|
6
|
+
exports.safeJsonParse = safeJsonParse;
|
|
7
|
+
exports.cleanText = cleanText;
|
|
8
|
+
exports.isValidUrl = isValidUrl;
|
|
9
|
+
const apiClient_1 = require("./apiClient");
|
|
10
|
+
async function getCrawl4aiClient(executeFunctions) {
|
|
11
|
+
const credentials = await executeFunctions.getCredentials('crawl4aiApi');
|
|
12
|
+
if (!credentials) {
|
|
13
|
+
throw new Error('Crawl4AI credentials are not configured!');
|
|
14
|
+
}
|
|
15
|
+
return (0, apiClient_1.createCrawlerInstance)(credentials);
|
|
16
|
+
}
|
|
17
|
+
function createBrowserConfig(options) {
|
|
18
|
+
const config = {
|
|
19
|
+
browser_type: options.browserType ? String(options.browserType) : 'chromium',
|
|
20
|
+
headless: options.headless !== false,
|
|
21
|
+
browser_mode: options.browserMode ? String(options.browserMode) : 'dedicated',
|
|
22
|
+
use_managed_browser: options.useManagedBrowser === true,
|
|
23
|
+
debugging_port: options.debuggingPort ? Number(options.debuggingPort) : 9222,
|
|
24
|
+
chrome_channel: options.chromeChannel ? String(options.chromeChannel) : 'chromium',
|
|
25
|
+
channel: options.channel ? String(options.channel) : 'chromium',
|
|
26
|
+
viewport_width: options.viewportWidth ? Number(options.viewportWidth) : 1080,
|
|
27
|
+
viewport_height: options.viewportHeight ? Number(options.viewportHeight) : 600,
|
|
28
|
+
java_script_enabled: options.javaScriptEnabled !== false,
|
|
29
|
+
user_agent: options.userAgent ? String(options.userAgent) : "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 Chrome/116.0.0.0 Safari/537.36",
|
|
30
|
+
ignore_https_errors: options.ignoreHttpsErrors !== false,
|
|
31
|
+
cookies: options.cookies ? Array.isArray(options.cookies) ? options.cookies : [] : [],
|
|
32
|
+
headers: options.headers ? (typeof options.headers === 'object' ? options.headers : {}) : {},
|
|
33
|
+
text_mode: options.textMode === true,
|
|
34
|
+
light_mode: options.lightMode === true,
|
|
35
|
+
extra_args: options.extraArgs ? Array.isArray(options.extraArgs) ? options.extraArgs : [] : [],
|
|
36
|
+
enable_stealth: options.enableStealth === true,
|
|
37
|
+
};
|
|
38
|
+
config.viewport = {
|
|
39
|
+
width: config.viewport_width || 1080,
|
|
40
|
+
height: config.viewport_height || 600,
|
|
41
|
+
};
|
|
42
|
+
config.javaScriptEnabled = config.java_script_enabled;
|
|
43
|
+
config.userAgent = config.user_agent;
|
|
44
|
+
return config;
|
|
45
|
+
}
|
|
46
|
+
function createCrawlerRunConfig(options) {
|
|
47
|
+
let excludedTags = [];
|
|
48
|
+
if (options.excludedTags) {
|
|
49
|
+
if (typeof options.excludedTags === 'string') {
|
|
50
|
+
excludedTags = options.excludedTags
|
|
51
|
+
.split(',')
|
|
52
|
+
.map(tag => tag.trim())
|
|
53
|
+
.filter(tag => tag);
|
|
54
|
+
}
|
|
55
|
+
else if (Array.isArray(options.excludedTags)) {
|
|
56
|
+
excludedTags = options.excludedTags;
|
|
57
|
+
}
|
|
58
|
+
}
|
|
59
|
+
return {
|
|
60
|
+
cacheMode: options.cacheMode || 'enabled',
|
|
61
|
+
streamEnabled: options.streamEnabled === true,
|
|
62
|
+
pageTimeout: options.pageTimeout ? Number(options.pageTimeout) : 30000,
|
|
63
|
+
requestTimeout: options.requestTimeout ? Number(options.requestTimeout) : 30000,
|
|
64
|
+
jsCode: options.jsCode ? String(options.jsCode) : undefined,
|
|
65
|
+
jsOnly: options.jsOnly === true,
|
|
66
|
+
cssSelector: options.cssSelector ? String(options.cssSelector) : undefined,
|
|
67
|
+
excludedTags,
|
|
68
|
+
excludeExternalLinks: options.excludeExternalLinks === true,
|
|
69
|
+
checkRobotsTxt: options.checkRobotsTxt === true,
|
|
70
|
+
wordCountThreshold: options.wordCountThreshold ? Number(options.wordCountThreshold) : 0,
|
|
71
|
+
sessionId: options.sessionId ? String(options.sessionId) : undefined,
|
|
72
|
+
maxRetries: options.maxRetries ? Number(options.maxRetries) : 3,
|
|
73
|
+
};
|
|
74
|
+
}
|
|
75
|
+
function safeJsonParse(jsonString, defaultValue = null) {
|
|
76
|
+
try {
|
|
77
|
+
return JSON.parse(jsonString);
|
|
78
|
+
}
|
|
79
|
+
catch (error) {
|
|
80
|
+
return defaultValue;
|
|
81
|
+
}
|
|
82
|
+
}
|
|
83
|
+
function cleanText(text) {
|
|
84
|
+
return text
|
|
85
|
+
.replace(/\s+/g, ' ')
|
|
86
|
+
.trim();
|
|
87
|
+
}
|
|
88
|
+
function isValidUrl(url) {
|
|
89
|
+
try {
|
|
90
|
+
new URL(url);
|
|
91
|
+
return true;
|
|
92
|
+
}
|
|
93
|
+
catch (error) {
|
|
94
|
+
return false;
|
|
95
|
+
}
|
|
96
|
+
}
|
|
97
|
+
//# sourceMappingURL=utils.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../../nodes/Crawl4aiBasicCrawler/helpers/utils.ts"],"names":[],"mappings":";;AAOA,8CAYC;AAOD,kDAiCC;AAOD,wDA6BC;AAKD,sCAMC;AAKD,8BAIC;AAKD,gCAOC;AA7HD,2CAAoD;AAK7C,KAAK,UAAU,iBAAiB,CACrC,gBAAmC;IAGnC,MAAM,WAAW,GAAG,MAAM,gBAAgB,CAAC,cAAc,CAAC,aAAa,CAAsC,CAAC;IAE9G,IAAI,CAAC,WAAW,EAAE,CAAC;QACjB,MAAM,IAAI,KAAK,CAAC,0CAA0C,CAAC,CAAC;IAC9D,CAAC;IAGD,OAAO,IAAA,iCAAqB,EAAC,WAAW,CAAC,CAAC;AAC5C,CAAC;AAOD,SAAgB,mBAAmB,CAAC,OAAoB;IACtD,MAAM,MAAM,GAAkB;QAC5B,YAAY,EAAE,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,UAAU;QAC5E,QAAQ,EAAE,OAAO,CAAC,QAAQ,KAAK,KAAK;QACpC,YAAY,EAAE,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,WAAW;QAC7E,mBAAmB,EAAE,OAAO,CAAC,iBAAiB,KAAK,IAAI;QACvD,cAAc,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,IAAI;QAC5E,cAAc,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,UAAU;QAClF,OAAO,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,UAAU;QAC/D,cAAc,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,IAAI;QAC5E,eAAe,EAAE,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,GAAG;QAC9E,mBAAmB,EAAE,OAAO,CAAC,iBAAiB,KAAK,KAAK;QACxD,UAAU,EAAE,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,mFAAmF;QAC/I,mBAAmB,EAAE,OAAO,CAAC,iBAAiB,KAAK,KAAK;QACxD,OAAO,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE;QACrF,OAAO,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,OAAO,CAAC,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE;QAC5F,SAAS,EAAE,OAAO,CAAC,QAAQ,KAAK,IAAI;QACpC,UAAU,EAAE,OAAO,CAAC,SAAS,KAAK,IAAI;QACtC,UAAU,EAAE,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE;QAC9F,cAAc,EAAE,OAAO,CAAC,aAAa,KAAK,IAAI;KAC/C,CAAC;IAGF,MAAM,CAAC,QAAQ,GAAG;QAChB,KAAK,EAAE,MAAM,CAAC,cAAc,IAAI,IAAI;QACpC,MAAM,EAAE,MAAM,CAAC,eAAe,IAAI,GAAG;KACtC,CAAC;IAGF,MAAM,CAAC,iBAAiB,GAAG,MAAM,CAAC,mBAAmB,CAAC;IACtD,MAAM,CAAC,SAAS,GAAG,MAAM,CAAC,UAAU,CAAC;IAErC,OAAO,MAAM,CAAC;AAChB,CAAC;AAOD,SAAgB,sBAAsB,CAAC,OAAoB;IAEzD,IAAI,YAAY,GAAa,EAAE,CAAC;IAChC,IAAI,OAAO,CAAC,YAAY,EAAE,CAAC;QACzB,IAAI,OAAO,OAAO,CAAC,YAAY,KAAK,QAAQ,EAAE,CAAC;YAC7C,YAAY,GAAI,OAAO,CAAC,YAAuB;iBAC5C,KAAK,CAAC,GAAG,CAAC;iBACV,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC;iBACtB,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;aAAM,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,YAAY,CAAC,EAAE,CAAC;YAC/C,YAAY,GAAG,OAAO,CAAC,YAAwB,CAAC;QAClD,CAAC;IACH,CAAC;IAED,OAAO;QACL,SAAS,EAAE,OAAO,CAAC,SAA0C,IAAI,SAAS;QAC1E,aAAa,EAAE,OAAO,CAAC,aAAa,KAAK,IAAI;QAC7C,WAAW,EAAE,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,KAAK;QACtE,cAAc,EAAE,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,KAAK;QAC/E,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS;QAC3D,MAAM,EAAE,OAAO,CAAC,MAAM,KAAK,IAAI;QAC/B,WAAW,EAAE,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,SAAS;QAC1E,YAAY;QACZ,oBAAoB,EAAE,OAAO,CAAC,oBAAoB,KAAK,IAAI;QAC3D,cAAc,EAAE,OAAO,CAAC,cAAc,KAAK,IAAI;QAC/C,kBAAkB,EAAE,OAAO,CAAC,kBAAkB,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC,CAAC;QACvF,SAAS,EAAE,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS;QACpE,UAAU,EAAE,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;KAChE,CAAC;AACJ,CAAC;AAKD,SAAgB,aAAa,CAAC,UAAkB,EAAE,eAAoB,IAAI;IACxE,IAAI,CAAC;QACH,OAAO,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;IAChC,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,OAAO,YAAY,CAAC;IACtB,CAAC;AACH,CAAC;AAKD,SAAgB,SAAS,CAAC,IAAY;IACpC,OAAO,IAAI;SACR,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,IAAI,EAAE,CAAC;AACZ,CAAC;AAKD,SAAgB,UAAU,CAAC,GAAW;IACpC,IAAI,CAAC;QACH,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;QACb,OAAO,IAAI,CAAC;IACd,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,5 @@
|
|
|
1
|
+
import { IExecuteFunctions, INodeExecutionData, INodeType, INodeTypeDescription } from 'n8n-workflow';
|
|
2
|
+
export declare class Crawl4aiContentExtractor implements INodeType {
|
|
3
|
+
description: INodeTypeDescription;
|
|
4
|
+
execute(this: IExecuteFunctions): Promise<INodeExecutionData[][]>;
|
|
5
|
+
}
|
|
@@ -0,0 +1,38 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.Crawl4aiContentExtractor = void 0;
|
|
4
|
+
const router_1 = require("./actions/router");
|
|
5
|
+
const operations_1 = require("./actions/operations");
|
|
6
|
+
class Crawl4aiContentExtractor {
|
|
7
|
+
constructor() {
|
|
8
|
+
this.description = {
|
|
9
|
+
displayName: 'Crawl4AI: Content Extractor',
|
|
10
|
+
name: 'crawl4aiContentExtractor',
|
|
11
|
+
icon: 'file:crawl4ai.svg',
|
|
12
|
+
group: ['transform'],
|
|
13
|
+
version: 1,
|
|
14
|
+
subtitle: '={{$parameter["operation"]}}',
|
|
15
|
+
description: 'Extract structured content from web pages using Crawl4AI',
|
|
16
|
+
defaults: {
|
|
17
|
+
name: 'Crawl4AI: Content Extractor',
|
|
18
|
+
},
|
|
19
|
+
inputs: ['main'],
|
|
20
|
+
outputs: ['main'],
|
|
21
|
+
usableAsTool: true,
|
|
22
|
+
credentials: [
|
|
23
|
+
{
|
|
24
|
+
name: 'crawl4aiApi',
|
|
25
|
+
required: true,
|
|
26
|
+
},
|
|
27
|
+
],
|
|
28
|
+
properties: [
|
|
29
|
+
...operations_1.description,
|
|
30
|
+
],
|
|
31
|
+
};
|
|
32
|
+
}
|
|
33
|
+
async execute() {
|
|
34
|
+
return await router_1.router.call(this);
|
|
35
|
+
}
|
|
36
|
+
}
|
|
37
|
+
exports.Crawl4aiContentExtractor = Crawl4aiContentExtractor;
|
|
38
|
+
//# sourceMappingURL=Crawl4aiContentExtractor.node.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"Crawl4aiContentExtractor.node.js","sourceRoot":"","sources":["../../../nodes/Crawl4aiContentExtractor/Crawl4aiContentExtractor.node.ts"],"names":[],"mappings":";;;AAQA,6CAA0C;AAC1C,qDAA4E;AAG5E,MAAa,wBAAwB;IAArC;QACC,gBAAW,GAAyB;YACnC,WAAW,EAAE,6BAA6B;YAC1C,IAAI,EAAE,0BAA0B;YAChC,IAAI,EAAE,mBAAmB;YACzB,KAAK,EAAE,CAAC,WAAW,CAAC;YACpB,OAAO,EAAE,CAAC;YACV,QAAQ,EAAE,8BAA8B;YACxC,WAAW,EAAE,0DAA0D;YACvE,QAAQ,EAAE;gBACT,IAAI,EAAE,6BAA6B;aACnC;YACD,MAAM,EAAE,CAAC,MAAM,CAAC;YAChB,OAAO,EAAE,CAAC,MAAM,CAAC;YACjB,YAAY,EAAE,IAAI;YAClB,WAAW,EAAE;gBACZ;oBACC,IAAI,EAAE,aAAa;oBACnB,QAAQ,EAAE,IAAI;iBACd;aACD;YACD,UAAU,EAAE;gBACX,GAAG,wBAAqB;aACxB;SACD,CAAC;IAMH,CAAC;IAHA,KAAK,CAAC,OAAO;QACZ,OAAO,MAAM,eAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAChC,CAAC;CACD;AA9BD,4DA8BC"}
|
|
@@ -0,0 +1,4 @@
|
|
|
1
|
+
import type { IExecuteFunctions, INodeExecutionData, INodeProperties } from 'n8n-workflow';
|
|
2
|
+
import type { Crawl4aiNodeOptions } from '../helpers/interfaces';
|
|
3
|
+
export declare const description: INodeProperties[];
|
|
4
|
+
export declare function execute(this: IExecuteFunctions, items: INodeExecutionData[], nodeOptions: Crawl4aiNodeOptions): Promise<INodeExecutionData[]>;
|