@nimble-way/nimble-js 0.20.0 → 1.0.0
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/CHANGELOG.md +9 -0
- package/README.md +10 -10
- package/client.d.mts +10 -39
- package/client.d.mts.map +1 -1
- package/client.d.ts +10 -39
- package/client.d.ts.map +1 -1
- package/client.js +8 -43
- package/client.js.map +1 -1
- package/client.mjs +8 -43
- package/client.mjs.map +1 -1
- package/package.json +1 -1
- package/resources/{task-agent/task-agent.d.mts → agents/agents.d.mts} +51 -213
- package/resources/agents/agents.d.mts.map +1 -0
- package/resources/{task-agent/task-agent.d.ts → agents/agents.d.ts} +51 -213
- package/resources/agents/agents.d.ts.map +1 -0
- package/resources/agents/agents.js +67 -0
- package/resources/agents/agents.js.map +1 -0
- package/resources/agents/agents.mjs +62 -0
- package/resources/agents/agents.mjs.map +1 -0
- package/resources/agents/index.d.mts +4 -0
- package/resources/agents/index.d.mts.map +1 -0
- package/resources/agents/index.d.ts +4 -0
- package/resources/agents/index.d.ts.map +1 -0
- package/resources/{task-agent → agents}/index.js +3 -3
- package/resources/agents/index.js.map +1 -0
- package/resources/{task-agent → agents}/index.mjs +1 -1
- package/resources/agents/index.mjs.map +1 -0
- package/resources/{task-agent → agents}/runs.d.mts +192 -43
- package/resources/agents/runs.d.mts.map +1 -0
- package/resources/{task-agent → agents}/runs.d.ts +192 -43
- package/resources/agents/runs.d.ts.map +1 -0
- package/resources/agents/runs.js +63 -0
- package/resources/agents/runs.js.map +1 -0
- package/resources/agents/runs.mjs +59 -0
- package/resources/agents/runs.mjs.map +1 -0
- package/resources/{task-agent → agents}/templates.d.mts +3 -6
- package/resources/agents/templates.d.mts.map +1 -0
- package/resources/{task-agent → agents}/templates.d.ts +3 -6
- package/resources/agents/templates.d.ts.map +1 -0
- package/resources/{task-agent → agents}/templates.js +5 -8
- package/resources/agents/templates.js.map +1 -0
- package/resources/agents/templates.mjs +19 -0
- package/resources/agents/templates.mjs.map +1 -0
- package/resources/agents.d.mts +2 -0
- package/resources/agents.d.mts.map +1 -0
- package/resources/agents.d.ts +2 -0
- package/resources/agents.d.ts.map +1 -0
- package/resources/{task-agent.js → agents.js} +2 -2
- package/resources/agents.js.map +1 -0
- package/resources/{task-agent.mjs → agents.mjs} +2 -2
- package/resources/agents.mjs.map +1 -0
- package/resources/batches.js +3 -3
- package/resources/batches.mjs +3 -3
- package/resources/crawl.js +4 -4
- package/resources/crawl.mjs +4 -4
- package/resources/domain-knowledge.js +1 -1
- package/resources/domain-knowledge.mjs +1 -1
- package/resources/extract/extract.d.mts +1243 -0
- package/resources/extract/extract.d.mts.map +1 -0
- package/resources/extract/extract.d.ts +1243 -0
- package/resources/extract/extract.d.ts.map +1 -0
- package/resources/extract/extract.js +52 -0
- package/resources/extract/extract.js.map +1 -0
- package/resources/extract/extract.mjs +47 -0
- package/resources/extract/extract.mjs.map +1 -0
- package/resources/extract/index.d.mts +3 -0
- package/resources/extract/index.d.mts.map +1 -0
- package/resources/extract/index.d.ts +3 -0
- package/resources/extract/index.d.ts.map +1 -0
- package/resources/extract/index.js +9 -0
- package/resources/extract/index.js.map +1 -0
- package/resources/extract/index.mjs +4 -0
- package/resources/extract/index.mjs.map +1 -0
- package/resources/extract/templates/generations.d.mts +364 -0
- package/resources/extract/templates/generations.d.mts.map +1 -0
- package/resources/extract/templates/generations.d.ts +364 -0
- package/resources/extract/templates/generations.d.ts.map +1 -0
- package/resources/extract/templates/generations.js +39 -0
- package/resources/extract/templates/generations.js.map +1 -0
- package/resources/extract/templates/generations.mjs +35 -0
- package/resources/extract/templates/generations.mjs.map +1 -0
- package/resources/extract/templates/index.d.mts +4 -0
- package/resources/extract/templates/index.d.mts.map +1 -0
- package/resources/extract/templates/index.d.ts +4 -0
- package/resources/extract/templates/index.d.ts.map +1 -0
- package/resources/extract/templates/index.js +11 -0
- package/resources/extract/templates/index.js.map +1 -0
- package/resources/extract/templates/index.mjs +5 -0
- package/resources/extract/templates/index.mjs.map +1 -0
- package/resources/extract/templates/templates.d.mts +870 -0
- package/resources/extract/templates/templates.d.mts.map +1 -0
- package/resources/extract/templates/templates.d.ts +870 -0
- package/resources/extract/templates/templates.d.ts.map +1 -0
- package/resources/extract/templates/templates.js +120 -0
- package/resources/extract/templates/templates.js.map +1 -0
- package/resources/extract/templates/templates.mjs +115 -0
- package/resources/extract/templates/templates.mjs.map +1 -0
- package/resources/extract/templates/versions.d.mts +226 -0
- package/resources/extract/templates/versions.d.mts.map +1 -0
- package/resources/extract/templates/versions.d.ts +226 -0
- package/resources/extract/templates/versions.d.ts.map +1 -0
- package/resources/extract/templates/versions.js +42 -0
- package/resources/extract/templates/versions.js.map +1 -0
- package/resources/extract/templates/versions.mjs +38 -0
- package/resources/extract/templates/versions.mjs.map +1 -0
- package/resources/extract/templates.d.mts +2 -0
- package/resources/extract/templates.d.mts.map +1 -0
- package/resources/extract/templates.d.ts +2 -0
- package/resources/extract/templates.d.ts.map +1 -0
- package/resources/extract/templates.js +6 -0
- package/resources/extract/templates.js.map +1 -0
- package/resources/extract/templates.mjs +3 -0
- package/resources/extract/templates.mjs.map +1 -0
- package/resources/extract.d.mts +2 -0
- package/resources/extract.d.mts.map +1 -0
- package/resources/extract.d.ts +2 -0
- package/resources/extract.d.ts.map +1 -0
- package/resources/extract.js +6 -0
- package/resources/extract.js.map +1 -0
- package/resources/extract.mjs +3 -0
- package/resources/extract.mjs.map +1 -0
- package/resources/fast-serp.js +1 -1
- package/resources/fast-serp.mjs +1 -1
- package/resources/index.d.mts +4 -4
- package/resources/index.d.mts.map +1 -1
- package/resources/index.d.ts +4 -4
- package/resources/index.d.ts.map +1 -1
- package/resources/index.js +5 -5
- package/resources/index.js.map +1 -1
- package/resources/index.mjs +2 -2
- package/resources/index.mjs.map +1 -1
- package/resources/jobs/index.d.mts +2 -2
- package/resources/jobs/index.d.mts.map +1 -1
- package/resources/jobs/index.d.ts +2 -2
- package/resources/jobs/index.d.ts.map +1 -1
- package/resources/jobs/index.js.map +1 -1
- package/resources/jobs/index.mjs.map +1 -1
- package/resources/jobs/jobs.d.mts +35 -103
- package/resources/jobs/jobs.d.mts.map +1 -1
- package/resources/jobs/jobs.d.ts +35 -103
- package/resources/jobs/jobs.d.ts.map +1 -1
- package/resources/jobs/jobs.js +10 -28
- package/resources/jobs/jobs.js.map +1 -1
- package/resources/jobs/jobs.mjs +11 -29
- package/resources/jobs/jobs.mjs.map +1 -1
- package/resources/jobs/runs/artifacts.d.mts +4 -12
- package/resources/jobs/runs/artifacts.d.mts.map +1 -1
- package/resources/jobs/runs/artifacts.d.ts +4 -12
- package/resources/jobs/runs/artifacts.d.ts.map +1 -1
- package/resources/jobs/runs/artifacts.js +8 -16
- package/resources/jobs/runs/artifacts.js.map +1 -1
- package/resources/jobs/runs/artifacts.mjs +8 -16
- package/resources/jobs/runs/artifacts.mjs.map +1 -1
- package/resources/jobs/runs/index.d.mts +1 -1
- package/resources/jobs/runs/index.d.mts.map +1 -1
- package/resources/jobs/runs/index.d.ts +1 -1
- package/resources/jobs/runs/index.d.ts.map +1 -1
- package/resources/jobs/runs/index.js.map +1 -1
- package/resources/jobs/runs/index.mjs.map +1 -1
- package/resources/jobs/runs/runs.d.mts +60 -28
- package/resources/jobs/runs/runs.d.mts.map +1 -1
- package/resources/jobs/runs/runs.d.ts +60 -28
- package/resources/jobs/runs/runs.d.ts.map +1 -1
- package/resources/jobs/runs/runs.js +12 -12
- package/resources/jobs/runs/runs.js.map +1 -1
- package/resources/jobs/runs/runs.mjs +12 -12
- package/resources/jobs/runs/runs.mjs.map +1 -1
- package/resources/media.js +2 -2
- package/resources/media.mjs +2 -2
- package/resources/serp.js +3 -3
- package/resources/serp.mjs +3 -3
- package/resources/tasks.js +3 -3
- package/resources/tasks.mjs +3 -3
- package/resources/top-level.d.mts +1 -1202
- package/resources/top-level.d.mts.map +1 -1
- package/resources/top-level.d.ts +1 -1202
- package/resources/top-level.d.ts.map +1 -1
- package/src/client.ts +41 -123
- package/src/resources/{task-agent/task-agent.ts → agents/agents.ts} +71 -258
- package/src/resources/{task-agent → agents}/index.ts +14 -14
- package/src/resources/{task-agent → agents}/runs.ts +228 -54
- package/src/resources/{task-agent → agents}/templates.ts +5 -8
- package/src/resources/{task-agent.ts → agents.ts} +1 -1
- package/src/resources/batches.ts +3 -3
- package/src/resources/crawl.ts +4 -4
- package/src/resources/domain-knowledge.ts +1 -1
- package/src/resources/extract/extract.ts +5182 -0
- package/src/resources/extract/index.ts +25 -0
- package/src/resources/extract/templates/generations.ts +435 -0
- package/src/resources/extract/templates/index.ts +29 -0
- package/src/resources/extract/templates/templates.ts +1176 -0
- package/src/resources/extract/templates/versions.ts +284 -0
- package/src/resources/extract/templates.ts +3 -0
- package/src/resources/extract.ts +3 -0
- package/src/resources/fast-serp.ts +1 -1
- package/src/resources/index.ts +16 -36
- package/src/resources/jobs/index.ts +1 -1
- package/src/resources/jobs/jobs.ts +50 -125
- package/src/resources/jobs/runs/artifacts.ts +8 -16
- package/src/resources/jobs/runs/index.ts +1 -0
- package/src/resources/jobs/runs/runs.ts +75 -31
- package/src/resources/media.ts +2 -2
- package/src/resources/serp.ts +3 -3
- package/src/resources/tasks.ts +3 -3
- package/src/resources/top-level.ts +72 -5171
- package/src/version.ts +1 -1
- package/version.d.mts +1 -1
- package/version.d.mts.map +1 -1
- package/version.d.ts +1 -1
- package/version.d.ts.map +1 -1
- package/version.js +1 -1
- package/version.js.map +1 -1
- package/version.mjs +1 -1
- package/version.mjs.map +1 -1
- package/resources/agent.d.mts +0 -641
- package/resources/agent.d.mts.map +0 -1
- package/resources/agent.d.ts +0 -641
- package/resources/agent.d.ts.map +0 -1
- package/resources/agent.js +0 -84
- package/resources/agent.js.map +0 -1
- package/resources/agent.mjs +0 -80
- package/resources/agent.mjs.map +0 -1
- package/resources/task-agent/index.d.mts +0 -4
- package/resources/task-agent/index.d.mts.map +0 -1
- package/resources/task-agent/index.d.ts +0 -4
- package/resources/task-agent/index.d.ts.map +0 -1
- package/resources/task-agent/index.js.map +0 -1
- package/resources/task-agent/index.mjs.map +0 -1
- package/resources/task-agent/runs.d.mts.map +0 -1
- package/resources/task-agent/runs.d.ts.map +0 -1
- package/resources/task-agent/runs.js +0 -63
- package/resources/task-agent/runs.js.map +0 -1
- package/resources/task-agent/runs.mjs +0 -59
- package/resources/task-agent/runs.mjs.map +0 -1
- package/resources/task-agent/task-agent.d.mts.map +0 -1
- package/resources/task-agent/task-agent.d.ts.map +0 -1
- package/resources/task-agent/task-agent.js +0 -80
- package/resources/task-agent/task-agent.js.map +0 -1
- package/resources/task-agent/task-agent.mjs +0 -75
- package/resources/task-agent/task-agent.mjs.map +0 -1
- package/resources/task-agent/templates.d.mts.map +0 -1
- package/resources/task-agent/templates.d.ts.map +0 -1
- package/resources/task-agent/templates.js.map +0 -1
- package/resources/task-agent/templates.mjs +0 -22
- package/resources/task-agent/templates.mjs.map +0 -1
- package/resources/task-agent.d.mts +0 -2
- package/resources/task-agent.d.mts.map +0 -1
- package/resources/task-agent.d.ts +0 -2
- package/resources/task-agent.d.ts.map +0 -1
- package/resources/task-agent.js.map +0 -1
- package/resources/task-agent.mjs.map +0 -1
- package/src/resources/agent.ts +0 -935
|
@@ -0,0 +1,1243 @@
|
|
|
1
|
+
import { APIResource } from "../../core/resource.mjs";
|
|
2
|
+
import * as Shared from "../shared.mjs";
|
|
3
|
+
import * as TemplatesAPI from "./templates/templates.mjs";
|
|
4
|
+
import { TemplateAsyncParams, TemplateAsyncResponse, TemplateBatchParams, TemplateBatchResponse, TemplateGetResponse, TemplateListParams, TemplateListResponse, TemplateRunParams, TemplateRunResponse, TemplateUpdateParams, TemplateUpdateResponse, Templates } from "./templates/templates.mjs";
|
|
5
|
+
import { APIPromise } from "../../core/api-promise.mjs";
|
|
6
|
+
import { RequestOptions } from "../../internal/request-options.mjs";
|
|
7
|
+
export declare class Extract extends APIResource {
|
|
8
|
+
templates: TemplatesAPI.Templates;
|
|
9
|
+
/**
|
|
10
|
+
* Extract Async Endpoint
|
|
11
|
+
*
|
|
12
|
+
* @example
|
|
13
|
+
* ```ts
|
|
14
|
+
* const response = await client.extract.async({ url: 'url' });
|
|
15
|
+
* ```
|
|
16
|
+
*/
|
|
17
|
+
async(body: ExtractAsyncParams, options?: RequestOptions): APIPromise<ExtractAsyncResponse>;
|
|
18
|
+
/**
|
|
19
|
+
* Extract Batch Endpoint
|
|
20
|
+
*
|
|
21
|
+
* @example
|
|
22
|
+
* ```ts
|
|
23
|
+
* const response = await client.extract.batch({
|
|
24
|
+
* inputs: [{}],
|
|
25
|
+
* });
|
|
26
|
+
* ```
|
|
27
|
+
*/
|
|
28
|
+
batch(body: ExtractBatchParams, options?: RequestOptions): APIPromise<ExtractBatchResponse>;
|
|
29
|
+
/**
|
|
30
|
+
* Extract
|
|
31
|
+
*
|
|
32
|
+
* @example
|
|
33
|
+
* ```ts
|
|
34
|
+
* const response = await client.extract.run({ url: 'url' });
|
|
35
|
+
* ```
|
|
36
|
+
*/
|
|
37
|
+
run(body: ExtractRunParams, options?: RequestOptions): APIPromise<ExtractRunResponse>;
|
|
38
|
+
}
|
|
39
|
+
/**
|
|
40
|
+
* Response when an async extract task is created successfully.
|
|
41
|
+
*/
|
|
42
|
+
export interface ExtractAsyncResponse {
|
|
43
|
+
/**
|
|
44
|
+
* Status indicating the async task was created successfully.
|
|
45
|
+
*/
|
|
46
|
+
status: 'success';
|
|
47
|
+
/**
|
|
48
|
+
* The created async task details.
|
|
49
|
+
*/
|
|
50
|
+
task: ExtractAsyncResponse.Task;
|
|
51
|
+
}
|
|
52
|
+
export declare namespace ExtractAsyncResponse {
|
|
53
|
+
/**
|
|
54
|
+
* The created async task details.
|
|
55
|
+
*/
|
|
56
|
+
interface Task {
|
|
57
|
+
/**
|
|
58
|
+
* Unique task identifier.
|
|
59
|
+
*/
|
|
60
|
+
id: string;
|
|
61
|
+
_query: unknown;
|
|
62
|
+
/**
|
|
63
|
+
* Timestamp when the task was created.
|
|
64
|
+
*/
|
|
65
|
+
created_at: string;
|
|
66
|
+
/**
|
|
67
|
+
* Original input data for the task.
|
|
68
|
+
*/
|
|
69
|
+
input: unknown;
|
|
70
|
+
/**
|
|
71
|
+
* Current state of the task.
|
|
72
|
+
*/
|
|
73
|
+
state: 'pending' | 'queued' | 'in_progress' | 'success' | 'error';
|
|
74
|
+
/**
|
|
75
|
+
* URL for checking the task status.
|
|
76
|
+
*/
|
|
77
|
+
status_url: string;
|
|
78
|
+
/**
|
|
79
|
+
* Account name that owns the task.
|
|
80
|
+
*/
|
|
81
|
+
account_name?: string;
|
|
82
|
+
api_type?: 'web' | 'serp' | 'ecommerce' | 'social' | 'media' | 'agent' | 'extract' | 'fast-serp' | 'labs';
|
|
83
|
+
/**
|
|
84
|
+
* Batch ID if this task is part of a batch.
|
|
85
|
+
*/
|
|
86
|
+
batch_id?: string | null;
|
|
87
|
+
/**
|
|
88
|
+
* URL for downloading the task results.
|
|
89
|
+
*/
|
|
90
|
+
download_url?: string | null;
|
|
91
|
+
/**
|
|
92
|
+
* Error message if the task failed.
|
|
93
|
+
*/
|
|
94
|
+
error?: string | null;
|
|
95
|
+
/**
|
|
96
|
+
* Classification of the error type.
|
|
97
|
+
*/
|
|
98
|
+
error_type?: string | null;
|
|
99
|
+
/**
|
|
100
|
+
* Timestamp when the task was last modified.
|
|
101
|
+
*/
|
|
102
|
+
modified_at?: string;
|
|
103
|
+
/**
|
|
104
|
+
* Storage location of the output data.
|
|
105
|
+
*/
|
|
106
|
+
output_url?: string | null;
|
|
107
|
+
/**
|
|
108
|
+
* Queue name the task was submitted to.
|
|
109
|
+
*/
|
|
110
|
+
queue?: string;
|
|
111
|
+
/**
|
|
112
|
+
* HTTP status code from the task execution.
|
|
113
|
+
*/
|
|
114
|
+
status_code?: number;
|
|
115
|
+
}
|
|
116
|
+
}
|
|
117
|
+
/**
|
|
118
|
+
* Response when a batch of extract tasks is created successfully.
|
|
119
|
+
*/
|
|
120
|
+
export interface ExtractBatchResponse {
|
|
121
|
+
/**
|
|
122
|
+
* Unique identifier for the batch.
|
|
123
|
+
*/
|
|
124
|
+
batch_id: string;
|
|
125
|
+
/**
|
|
126
|
+
* Number of tasks in the batch.
|
|
127
|
+
*/
|
|
128
|
+
batch_size: number;
|
|
129
|
+
/**
|
|
130
|
+
* List of created tasks.
|
|
131
|
+
*/
|
|
132
|
+
tasks: Array<ExtractBatchResponse.Task>;
|
|
133
|
+
}
|
|
134
|
+
export declare namespace ExtractBatchResponse {
|
|
135
|
+
interface Task {
|
|
136
|
+
/**
|
|
137
|
+
* Unique task identifier.
|
|
138
|
+
*/
|
|
139
|
+
id: string;
|
|
140
|
+
_query: unknown;
|
|
141
|
+
/**
|
|
142
|
+
* Timestamp when the task was created.
|
|
143
|
+
*/
|
|
144
|
+
created_at: string;
|
|
145
|
+
/**
|
|
146
|
+
* Original input data for the task.
|
|
147
|
+
*/
|
|
148
|
+
input: unknown;
|
|
149
|
+
/**
|
|
150
|
+
* Current state of the task.
|
|
151
|
+
*/
|
|
152
|
+
state: 'pending' | 'queued' | 'in_progress' | 'success' | 'error';
|
|
153
|
+
/**
|
|
154
|
+
* URL for checking the task status.
|
|
155
|
+
*/
|
|
156
|
+
status_url: string;
|
|
157
|
+
/**
|
|
158
|
+
* Account name that owns the task.
|
|
159
|
+
*/
|
|
160
|
+
account_name?: string;
|
|
161
|
+
api_type?: 'web' | 'serp' | 'ecommerce' | 'social' | 'media' | 'agent' | 'extract' | 'fast-serp' | 'labs';
|
|
162
|
+
/**
|
|
163
|
+
* Batch ID if this task is part of a batch.
|
|
164
|
+
*/
|
|
165
|
+
batch_id?: string | null;
|
|
166
|
+
/**
|
|
167
|
+
* URL for downloading the task results.
|
|
168
|
+
*/
|
|
169
|
+
download_url?: string | null;
|
|
170
|
+
/**
|
|
171
|
+
* Error message if the task failed.
|
|
172
|
+
*/
|
|
173
|
+
error?: string | null;
|
|
174
|
+
/**
|
|
175
|
+
* Classification of the error type.
|
|
176
|
+
*/
|
|
177
|
+
error_type?: string | null;
|
|
178
|
+
/**
|
|
179
|
+
* Timestamp when the task was last modified.
|
|
180
|
+
*/
|
|
181
|
+
modified_at?: string;
|
|
182
|
+
/**
|
|
183
|
+
* Storage location of the output data.
|
|
184
|
+
*/
|
|
185
|
+
output_url?: string | null;
|
|
186
|
+
/**
|
|
187
|
+
* Queue name the task was submitted to.
|
|
188
|
+
*/
|
|
189
|
+
queue?: string;
|
|
190
|
+
/**
|
|
191
|
+
* HTTP status code from the task execution.
|
|
192
|
+
*/
|
|
193
|
+
status_code?: number;
|
|
194
|
+
}
|
|
195
|
+
}
|
|
196
|
+
export interface ExtractRunResponse {
|
|
197
|
+
data: ExtractRunResponse.Data;
|
|
198
|
+
metadata: ExtractRunResponse.Metadata;
|
|
199
|
+
/**
|
|
200
|
+
* The status of the task.
|
|
201
|
+
*/
|
|
202
|
+
status: 'success' | 'skipped' | 'fatal' | 'error' | 'postponed' | 'ignored' | 'rejected' | 'blocked';
|
|
203
|
+
/**
|
|
204
|
+
* Unique identifier for the task.
|
|
205
|
+
*/
|
|
206
|
+
task_id: string;
|
|
207
|
+
/**
|
|
208
|
+
* The final URL.
|
|
209
|
+
*/
|
|
210
|
+
url: string;
|
|
211
|
+
debug?: ExtractRunResponse.Debug;
|
|
212
|
+
/**
|
|
213
|
+
* Pagination information if applicable.
|
|
214
|
+
*/
|
|
215
|
+
pagination?: ExtractRunResponse.NextPageParams | Array<ExtractRunResponse.UnionMember1>;
|
|
216
|
+
/**
|
|
217
|
+
* The HTTP status code of the task.
|
|
218
|
+
*/
|
|
219
|
+
status_code?: number;
|
|
220
|
+
/**
|
|
221
|
+
* List of warnings generated during the task.
|
|
222
|
+
*/
|
|
223
|
+
warnings?: Array<string>;
|
|
224
|
+
}
|
|
225
|
+
export declare namespace ExtractRunResponse {
|
|
226
|
+
interface Data {
|
|
227
|
+
/**
|
|
228
|
+
* Browser actions execution results. Present only when browser_actions were
|
|
229
|
+
* specified in the request.
|
|
230
|
+
*/
|
|
231
|
+
browser_actions?: Data.BrowserActions;
|
|
232
|
+
/**
|
|
233
|
+
* The cookies collected from browser actions during the task.
|
|
234
|
+
*/
|
|
235
|
+
cookies?: Array<unknown>;
|
|
236
|
+
/**
|
|
237
|
+
* The evaluation results from browser actions during the task.
|
|
238
|
+
*/
|
|
239
|
+
eval?: Array<unknown>;
|
|
240
|
+
/**
|
|
241
|
+
* The http requests from browser actions made during the task.
|
|
242
|
+
*/
|
|
243
|
+
fetch?: Array<unknown>;
|
|
244
|
+
/**
|
|
245
|
+
* The headers received during the task.
|
|
246
|
+
*/
|
|
247
|
+
headers?: {
|
|
248
|
+
[key: string]: string;
|
|
249
|
+
};
|
|
250
|
+
/**
|
|
251
|
+
* The HTML content of the page.
|
|
252
|
+
*/
|
|
253
|
+
html?: string;
|
|
254
|
+
/**
|
|
255
|
+
* List of all unique URLs found on the page.
|
|
256
|
+
*/
|
|
257
|
+
links?: Array<string>;
|
|
258
|
+
/**
|
|
259
|
+
* The Markdown version of the HTML content.
|
|
260
|
+
*/
|
|
261
|
+
markdown?: string;
|
|
262
|
+
/**
|
|
263
|
+
* The network capture data collected during the task.
|
|
264
|
+
*/
|
|
265
|
+
network_capture?: Array<Data.NetworkCapture>;
|
|
266
|
+
/**
|
|
267
|
+
* Individual HTML content of each pagination page, before merging.
|
|
268
|
+
*/
|
|
269
|
+
pages_html?: Array<string>;
|
|
270
|
+
/**
|
|
271
|
+
* The parsing results extracted from the HTML & network content.
|
|
272
|
+
*/
|
|
273
|
+
parsing?: Data.ParsingSuccessResult | Data.ParsingErrorResult | {
|
|
274
|
+
[key: string]: unknown;
|
|
275
|
+
};
|
|
276
|
+
/**
|
|
277
|
+
* The list of redirects that occurred during the task.
|
|
278
|
+
*/
|
|
279
|
+
redirects?: Array<Data.Redirect>;
|
|
280
|
+
/**
|
|
281
|
+
* Screenshots taken during the task, from browser actions, or the screenshot
|
|
282
|
+
* format.
|
|
283
|
+
*/
|
|
284
|
+
screenshots?: Array<unknown>;
|
|
285
|
+
}
|
|
286
|
+
namespace Data {
|
|
287
|
+
/**
|
|
288
|
+
* Browser actions execution results. Present only when browser_actions were
|
|
289
|
+
* specified in the request.
|
|
290
|
+
*/
|
|
291
|
+
interface BrowserActions {
|
|
292
|
+
results: Array<BrowserActions.Result>;
|
|
293
|
+
success: boolean;
|
|
294
|
+
total_duration: number;
|
|
295
|
+
}
|
|
296
|
+
namespace BrowserActions {
|
|
297
|
+
interface Result {
|
|
298
|
+
duration: number;
|
|
299
|
+
name: 'goto' | 'wait' | 'wait_for_element' | 'wait_for_navigation' | 'click' | 'fill' | 'press' | 'scroll' | 'auto_scroll' | 'screenshot' | 'get_cookies' | 'eval' | 'fetch';
|
|
300
|
+
status: 'no-run' | 'in-progress' | 'done' | 'error' | 'skipped';
|
|
301
|
+
error?: string;
|
|
302
|
+
result?: unknown;
|
|
303
|
+
}
|
|
304
|
+
}
|
|
305
|
+
interface NetworkCapture {
|
|
306
|
+
filter: NetworkCapture.Filter;
|
|
307
|
+
results: Array<NetworkCapture.Result>;
|
|
308
|
+
errorMessage?: string;
|
|
309
|
+
}
|
|
310
|
+
namespace NetworkCapture {
|
|
311
|
+
interface Filter {
|
|
312
|
+
validation: boolean;
|
|
313
|
+
wait_for_requests_count: number;
|
|
314
|
+
method?: 'GET' | 'HEAD' | 'POST' | 'PUT' | 'DELETE' | 'CONNECT' | 'OPTIONS' | 'TRACE' | 'PATCH';
|
|
315
|
+
/**
|
|
316
|
+
* Resource type for network capture filtering
|
|
317
|
+
*/
|
|
318
|
+
resource_type?: 'document' | 'stylesheet' | 'image' | 'media' | 'font' | 'script' | 'texttrack' | 'xhr' | 'fetch' | 'prefetch' | 'eventsource' | 'websocket' | 'manifest' | 'signedexchange' | 'ping' | 'cspviolationreport' | 'preflight' | 'other' | 'fedcm' | Array<'document' | 'stylesheet' | 'image' | 'media' | 'font' | 'script' | 'texttrack' | 'xhr' | 'fetch' | 'prefetch' | 'eventsource' | 'websocket' | 'manifest' | 'signedexchange' | 'ping' | 'cspviolationreport' | 'preflight' | 'other' | 'fedcm'>;
|
|
319
|
+
status_code?: number | Array<number>;
|
|
320
|
+
url?: Filter.URL;
|
|
321
|
+
wait_for_requests_count_timeout?: number;
|
|
322
|
+
}
|
|
323
|
+
namespace Filter {
|
|
324
|
+
interface URL {
|
|
325
|
+
type: 'exact' | 'contains';
|
|
326
|
+
value: string;
|
|
327
|
+
}
|
|
328
|
+
}
|
|
329
|
+
interface Result {
|
|
330
|
+
request: Result.Request;
|
|
331
|
+
response: Result.Response;
|
|
332
|
+
}
|
|
333
|
+
namespace Result {
|
|
334
|
+
interface Request {
|
|
335
|
+
headers: {
|
|
336
|
+
[key: string]: string;
|
|
337
|
+
};
|
|
338
|
+
method: string;
|
|
339
|
+
/**
|
|
340
|
+
* Resource type for network capture filtering
|
|
341
|
+
*/
|
|
342
|
+
resource_type: 'document' | 'stylesheet' | 'image' | 'media' | 'font' | 'script' | 'texttrack' | 'xhr' | 'fetch' | 'prefetch' | 'eventsource' | 'websocket' | 'manifest' | 'signedexchange' | 'ping' | 'cspviolationreport' | 'preflight' | 'other' | 'fedcm';
|
|
343
|
+
url: string;
|
|
344
|
+
body?: string;
|
|
345
|
+
}
|
|
346
|
+
interface Response {
|
|
347
|
+
body: string;
|
|
348
|
+
headers: {
|
|
349
|
+
[key: string]: string;
|
|
350
|
+
};
|
|
351
|
+
serialization: 'none' | 'base64';
|
|
352
|
+
status: number;
|
|
353
|
+
status_text: string;
|
|
354
|
+
}
|
|
355
|
+
}
|
|
356
|
+
}
|
|
357
|
+
interface ParsingSuccessResult {
|
|
358
|
+
entities: {
|
|
359
|
+
[key: string]: unknown;
|
|
360
|
+
};
|
|
361
|
+
status: 'success';
|
|
362
|
+
}
|
|
363
|
+
interface ParsingErrorResult {
|
|
364
|
+
error: string;
|
|
365
|
+
status: 'error';
|
|
366
|
+
}
|
|
367
|
+
interface Redirect {
|
|
368
|
+
status_code: number;
|
|
369
|
+
url: string;
|
|
370
|
+
}
|
|
371
|
+
}
|
|
372
|
+
interface Metadata {
|
|
373
|
+
/**
|
|
374
|
+
* The name of the agent used for the query.
|
|
375
|
+
*/
|
|
376
|
+
agent?: string;
|
|
377
|
+
/**
|
|
378
|
+
* The driver used for the task.
|
|
379
|
+
*/
|
|
380
|
+
driver?: string;
|
|
381
|
+
/**
|
|
382
|
+
* The localization identifier for the query.
|
|
383
|
+
*/
|
|
384
|
+
localization_id?: string;
|
|
385
|
+
/**
|
|
386
|
+
* The duration in milliseconds of the query processing.
|
|
387
|
+
*/
|
|
388
|
+
query_duration?: number;
|
|
389
|
+
/**
|
|
390
|
+
* The time when the query was received.
|
|
391
|
+
*/
|
|
392
|
+
query_time?: string;
|
|
393
|
+
/**
|
|
394
|
+
* Additional response parameters.
|
|
395
|
+
*/
|
|
396
|
+
response_parameters?: unknown;
|
|
397
|
+
/**
|
|
398
|
+
* A tag associated with the query.
|
|
399
|
+
*/
|
|
400
|
+
tag?: string;
|
|
401
|
+
}
|
|
402
|
+
interface Debug {
|
|
403
|
+
/**
|
|
404
|
+
* Performance metrics collected during the task.
|
|
405
|
+
*/
|
|
406
|
+
performance_metrics?: {
|
|
407
|
+
[key: string]: number;
|
|
408
|
+
};
|
|
409
|
+
/**
|
|
410
|
+
* Total bytes used by the proxy during the task.
|
|
411
|
+
*/
|
|
412
|
+
proxy_total_bytes_usage?: number;
|
|
413
|
+
/**
|
|
414
|
+
* The transformed output after applying any transformations.
|
|
415
|
+
*/
|
|
416
|
+
transformed_output?: unknown;
|
|
417
|
+
/**
|
|
418
|
+
* The userbrowser instance using during the task.
|
|
419
|
+
*/
|
|
420
|
+
userbrowser?: unknown;
|
|
421
|
+
}
|
|
422
|
+
interface NextPageParams {
|
|
423
|
+
next_page_params: {
|
|
424
|
+
[key: string]: unknown;
|
|
425
|
+
};
|
|
426
|
+
}
|
|
427
|
+
interface UnionMember1 {
|
|
428
|
+
next_page_params: {
|
|
429
|
+
[key: string]: unknown;
|
|
430
|
+
};
|
|
431
|
+
}
|
|
432
|
+
}
|
|
433
|
+
export interface ExtractAsyncParams {
|
|
434
|
+
/**
|
|
435
|
+
* Target URL to scrape
|
|
436
|
+
*/
|
|
437
|
+
url: string;
|
|
438
|
+
/**
|
|
439
|
+
* Custom flow for the optimization engine: maps candidate names to the number of
|
|
440
|
+
* attempts to spend on each candidate before advancing (0 skips it). Key order
|
|
441
|
+
* defines the flow order. Providing it opts the request into 'auto' driver
|
|
442
|
+
* selection.
|
|
443
|
+
*/
|
|
444
|
+
auto_driver_configuration?: {
|
|
445
|
+
[key: string]: number;
|
|
446
|
+
};
|
|
447
|
+
/**
|
|
448
|
+
* Request body for POST, PUT, PATCH methods
|
|
449
|
+
*/
|
|
450
|
+
body?: unknown;
|
|
451
|
+
/**
|
|
452
|
+
* Browser type to emulate
|
|
453
|
+
*/
|
|
454
|
+
browser?: 'chrome' | 'firefox' | ExtractAsyncParams.UnionMember1;
|
|
455
|
+
/**
|
|
456
|
+
* Array of browser automation actions to execute sequentially
|
|
457
|
+
*/
|
|
458
|
+
browser_actions?: Array<Shared.AutoScrollAction | Shared.ClickAction | Shared.EvalAction | Shared.FetchAction | Shared.FillAction | Shared.GetCookiesAction | Shared.GotoAction | Shared.PressAction | Shared.ScreenshotAction | Shared.ScrollAction | Shared.WaitAction | Shared.WaitForElementAction | Shared.WaitForNavigationAction>;
|
|
459
|
+
/**
|
|
460
|
+
* URL to call back when async operation completes
|
|
461
|
+
*/
|
|
462
|
+
callback_url?: string;
|
|
463
|
+
/**
|
|
464
|
+
* City for geolocation
|
|
465
|
+
*/
|
|
466
|
+
city?: string;
|
|
467
|
+
/**
|
|
468
|
+
* Whether to automatically handle cookie consent headers
|
|
469
|
+
*/
|
|
470
|
+
consent_header?: boolean;
|
|
471
|
+
/**
|
|
472
|
+
* Browser cookies as array of cookie objects
|
|
473
|
+
*/
|
|
474
|
+
cookies?: Array<ExtractAsyncParams.UnionMember0> | string;
|
|
475
|
+
/**
|
|
476
|
+
* Country code for geolocation and proxy selection
|
|
477
|
+
*/
|
|
478
|
+
country?: 'AD' | 'AE' | 'AF' | 'AG' | 'AI' | 'AL' | 'AM' | 'AO' | 'AQ' | 'AR' | 'AS' | 'AT' | 'AU' | 'AW' | 'AX' | 'AZ' | 'BA' | 'BB' | 'BD' | 'BE' | 'BF' | 'BG' | 'BH' | 'BI' | 'BJ' | 'BL' | 'BM' | 'BN' | 'BO' | 'BQ' | 'BR' | 'BS' | 'BT' | 'BV' | 'BW' | 'BY' | 'BZ' | 'CA' | 'CC' | 'CD' | 'CF' | 'CG' | 'CH' | 'CI' | 'CK' | 'CL' | 'CM' | 'CN' | 'CO' | 'CR' | 'CU' | 'CV' | 'CW' | 'CX' | 'CY' | 'CZ' | 'DE' | 'DJ' | 'DK' | 'DM' | 'DO' | 'DZ' | 'EC' | 'EE' | 'EG' | 'EH' | 'ER' | 'ES' | 'ET' | 'FI' | 'FJ' | 'FK' | 'FM' | 'FO' | 'FR' | 'GA' | 'GB' | 'GD' | 'GE' | 'GF' | 'GG' | 'GH' | 'GI' | 'GL' | 'GM' | 'GN' | 'GP' | 'GQ' | 'GR' | 'GS' | 'GT' | 'GU' | 'GW' | 'GY' | 'HK' | 'HM' | 'HN' | 'HR' | 'HT' | 'HU' | 'ID' | 'IE' | 'IL' | 'IM' | 'IN' | 'IO' | 'IQ' | 'IR' | 'IS' | 'IT' | 'JE' | 'JM' | 'JO' | 'JP' | 'KE' | 'KG' | 'KH' | 'KI' | 'KM' | 'KN' | 'KP' | 'KR' | 'KW' | 'KY' | 'KZ' | 'LA' | 'LB' | 'LC' | 'LI' | 'LK' | 'LR' | 'LS' | 'LT' | 'LU' | 'LV' | 'LY' | 'MA' | 'MC' | 'MD' | 'ME' | 'MF' | 'MG' | 'MH' | 'MK' | 'ML' | 'MM' | 'MN' | 'MO' | 'MP' | 'MQ' | 'MR' | 'MS' | 'MT' | 'MU' | 'MV' | 'MW' | 'MX' | 'MY' | 'MZ' | 'NA' | 'NC' | 'NE' | 'NF' | 'NG' | 'NI' | 'NL' | 'NO' | 'NP' | 'NR' | 'NU' | 'NZ' | 'OM' | 'PA' | 'PE' | 'PF' | 'PG' | 'PH' | 'PK' | 'PL' | 'PM' | 'PN' | 'PR' | 'PS' | 'PT' | 'PW' | 'PY' | 'QA' | 'RE' | 'RO' | 'RS' | 'RU' | 'RW' | 'SA' | 'SB' | 'SC' | 'SD' | 'SE' | 'SG' | 'SH' | 'SI' | 'SJ' | 'SK' | 'SL' | 'SM' | 'SN' | 'SO' | 'SR' | 'SS' | 'ST' | 'SV' | 'SX' | 'SY' | 'SZ' | 'TC' | 'TD' | 'TF' | 'TG' | 'TH' | 'TJ' | 'TK' | 'TL' | 'TM' | 'TN' | 'TO' | 'TR' | 'TT' | 'TV' | 'TW' | 'TZ' | 'UA' | 'UG' | 'UM' | 'US' | 'UY' | 'UZ' | 'VA' | 'VC' | 'VE' | 'VG' | 'VI' | 'VN' | 'VU' | 'WF' | 'WS' | 'XK' | 'YE' | 'YT' | 'ZA' | 'ZM' | 'ZW' | 'ALL';
|
|
479
|
+
/**
|
|
480
|
+
* Device type for browser emulation
|
|
481
|
+
*/
|
|
482
|
+
device?: 'desktop' | 'mobile' | 'tablet';
|
|
483
|
+
/**
|
|
484
|
+
* Browser driver to use
|
|
485
|
+
*/
|
|
486
|
+
driver?: 'auto' | 'vx6' | 'vx8' | 'vx8-pro' | 'vx10' | 'vx10-pro' | 'vx12' | 'vx12-pro' | 'media-vx6' | 'fast-vx6';
|
|
487
|
+
/**
|
|
488
|
+
* Expected HTTP status codes for successful requests
|
|
489
|
+
*/
|
|
490
|
+
expected_status_codes?: Array<number>;
|
|
491
|
+
/**
|
|
492
|
+
* List of acceptable response formats in order of preference
|
|
493
|
+
*/
|
|
494
|
+
formats?: Array<'html' | 'markdown' | 'screenshot' | 'headers' | 'links'>;
|
|
495
|
+
/**
|
|
496
|
+
* Custom HTTP headers to include in the request
|
|
497
|
+
*/
|
|
498
|
+
headers?: {
|
|
499
|
+
[key: string]: string | Array<string> | null;
|
|
500
|
+
};
|
|
501
|
+
/**
|
|
502
|
+
* Whether to use HTTP/2 protocol
|
|
503
|
+
*/
|
|
504
|
+
http2?: boolean;
|
|
505
|
+
/**
|
|
506
|
+
* Whether to emulate XMLHttpRequest behavior
|
|
507
|
+
*/
|
|
508
|
+
is_xhr?: boolean;
|
|
509
|
+
/**
|
|
510
|
+
* Locale for browser language and region settings
|
|
511
|
+
*/
|
|
512
|
+
locale?: 'aa-DJ' | 'aa-ER' | 'aa-ET' | 'af' | 'af-NA' | 'af-ZA' | 'ak' | 'ak-GH' | 'am' | 'am-ET' | 'an-ES' | 'ar' | 'ar-AE' | 'ar-BH' | 'ar-DZ' | 'ar-EG' | 'ar-IN' | 'ar-IQ' | 'ar-JO' | 'ar-KW' | 'ar-LB' | 'ar-LY' | 'ar-MA' | 'ar-OM' | 'ar-QA' | 'ar-SA' | 'ar-SD' | 'ar-SY' | 'ar-TN' | 'ar-YE' | 'as' | 'as-IN' | 'asa' | 'asa-TZ' | 'ast-ES' | 'az' | 'az-AZ' | 'az-Cyrl' | 'az-Cyrl-AZ' | 'az-Latn' | 'az-Latn-AZ' | 'be' | 'be-BY' | 'bem' | 'bem-ZM' | 'ber-DZ' | 'ber-MA' | 'bez' | 'bez-TZ' | 'bg' | 'bg-BG' | 'bho-IN' | 'bm' | 'bm-ML' | 'bn' | 'bn-BD' | 'bn-IN' | 'bo' | 'bo-CN' | 'bo-IN' | 'br-FR' | 'brx-IN' | 'bs' | 'bs-BA' | 'byn-ER' | 'ca' | 'ca-AD' | 'ca-ES' | 'ca-FR' | 'ca-IT' | 'cgg' | 'cgg-UG' | 'chr' | 'chr-US' | 'crh-UA' | 'cs' | 'cs-CZ' | 'csb-PL' | 'cv-RU' | 'cy' | 'cy-GB' | 'da' | 'da-DK' | 'dav' | 'dav-KE' | 'de' | 'de-AT' | 'de-BE' | 'de-CH' | 'de-DE' | 'de-LI' | 'de-LU' | 'dv-MV' | 'dz-BT' | 'ebu' | 'ebu-KE' | 'ee' | 'ee-GH' | 'ee-TG' | 'el' | 'el-CY' | 'el-GR' | 'en' | 'en-AG' | 'en-AS' | 'en-AU' | 'en-BE' | 'en-BW' | 'en-BZ' | 'en-CA' | 'en-DK' | 'en-GB' | 'en-GU' | 'en-HK' | 'en-IE' | 'en-IN' | 'en-JM' | 'en-MH' | 'en-MP' | 'en-MT' | 'en-MU' | 'en-NA' | 'en-NG' | 'en-NZ' | 'en-PH' | 'en-PK' | 'en-SG' | 'en-TT' | 'en-UM' | 'en-US' | 'en-VI' | 'en-ZA' | 'en-ZM' | 'en-ZW' | 'eo' | 'es' | 'es-419' | 'es-AR' | 'es-BO' | 'es-CL' | 'es-CO' | 'es-CR' | 'es-CU' | 'es-DO' | 'es-EC' | 'es-ES' | 'es-GQ' | 'es-GT' | 'es-HN' | 'es-MX' | 'es-NI' | 'es-PA' | 'es-PE' | 'es-PR' | 'es-PY' | 'es-SV' | 'es-US' | 'es-UY' | 'es-VE' | 'et' | 'et-EE' | 'eu' | 'eu-ES' | 'fa' | 'fa-AF' | 'fa-IR' | 'ff' | 'ff-SN' | 'fi' | 'fi-FI' | 'fil' | 'fil-PH' | 'fo' | 'fo-FO' | 'fr' | 'fr-BE' | 'fr-BF' | 'fr-BI' | 'fr-BJ' | 'fr-BL' | 'fr-CA' | 'fr-CD' | 'fr-CF' | 'fr-CG' | 'fr-CH' | 'fr-CI' | 'fr-CM' | 'fr-DJ' | 'fr-FR' | 'fr-GA' | 'fr-GN' | 'fr-GP' | 'fr-GQ' | 'fr-KM' | 'fr-LU' | 'fr-MC' | 'fr-MF' | 'fr-MG' | 'fr-ML' | 'fr-MQ' | 'fr-NE' | 'fr-RE' | 'fr-RW' | 'fr-SN' | 'fr-TD' | 'fr-TG' | 'fur-IT' | 'fy-DE' | 'fy-NL' | 'ga' | 'ga-IE' | 'gd-GB' | 'gez-ER' | 'gez-ET' | 'gl' | 'gl-ES' | 'gsw' | 'gsw-CH' | 'gu' | 'gu-IN' | 'guz' | 'guz-KE' | 'gv' | 'gv-GB' | 'ha' | 'ha-Latn' | 'ha-Latn-GH' | 'ha-Latn-NE' | 'ha-Latn-NG' | 'ha-NG' | 'haw' | 'haw-US' | 'he' | 'he-IL' | 'hi' | 'hi-IN' | 'hne-IN' | 'hr' | 'hr-HR' | 'hsb-DE' | 'ht-HT' | 'hu' | 'hu-HU' | 'hy' | 'hy-AM' | 'id' | 'id-ID' | 'ig' | 'ig-NG' | 'ii' | 'ii-CN' | 'ik-CA' | 'is' | 'is-IS' | 'it' | 'it-CH' | 'it-IT' | 'iu-CA' | 'iw-IL' | 'ja' | 'ja-JP' | 'jmc' | 'jmc-TZ' | 'ka' | 'ka-GE' | 'kab' | 'kab-DZ' | 'kam' | 'kam-KE' | 'kde' | 'kde-TZ' | 'kea' | 'kea-CV' | 'khq' | 'khq-ML' | 'ki' | 'ki-KE' | 'kk' | 'kk-Cyrl' | 'kk-Cyrl-KZ' | 'kk-KZ' | 'kl' | 'kl-GL' | 'kln' | 'kln-KE' | 'km' | 'km-KH' | 'kn' | 'kn-IN' | 'ko' | 'ko-KR' | 'kok' | 'kok-IN' | 'ks-IN' | 'ku-TR' | 'kw' | 'kw-GB' | 'ky-KG' | 'lag' | 'lag-TZ' | 'lb-LU' | 'lg' | 'lg-UG' | 'li-BE' | 'li-NL' | 'lij-IT' | 'lo-LA' | 'lt' | 'lt-LT' | 'luo' | 'luo-KE' | 'luy' | 'luy-KE' | 'lv' | 'lv-LV' | 'mag-IN' | 'mai-IN' | 'mas' | 'mas-KE' | 'mas-TZ' | 'mer' | 'mer-KE' | 'mfe' | 'mfe-MU' | 'mg' | 'mg-MG' | 'mhr-RU' | 'mi-NZ' | 'mk' | 'mk-MK' | 'ml' | 'ml-IN' | 'mn-MN' | 'mr' | 'mr-IN' | 'ms' | 'ms-BN' | 'ms-MY' | 'mt' | 'mt-MT' | 'my' | 'my-MM' | 'nan-TW' | 'naq' | 'naq-NA' | 'nb' | 'nb-NO' | 'nd' | 'nd-ZW' | 'nds-DE' | 'nds-NL' | 'ne' | 'ne-IN' | 'ne-NP' | 'nl' | 'nl-AW' | 'nl-BE' | 'nl-NL' | 'nn' | 'nn-NO' | 'nr-ZA' | 'nso-ZA' | 'nyn' | 'nyn-UG' | 'oc-FR' | 'om' | 'om-ET' | 'om-KE' | 'or' | 'or-IN' | 'os-RU' | 'pa' | 'pa-Arab' | 'pa-Arab-PK' | 'pa-Guru' | 'pa-Guru-IN' | 'pa-IN' | 'pa-PK' | 'pap-AN' | 'pl' | 'pl-PL' | 'ps' | 'ps-AF' | 'pt' | 'pt-BR' | 'pt-GW' | 'pt-MZ' | 'pt-PT' | 'rm' | 'rm-CH' | 'ro' | 'ro-MD' | 'ro-RO' | 'rof' | 'rof-TZ' | 'ru' | 'ru-MD' | 'ru-RU' | 'ru-UA' | 'rw' | 'rw-RW' | 'rwk' | 'rwk-TZ' | 'sa-IN' | 'saq' | 'saq-KE' | 'sc-IT' | 'sd-IN' | 'se-NO' | 'seh' | 'seh-MZ' | 'ses' | 'ses-ML' | 'sg' | 'sg-CF' | 'shi' | 'shi-Latn' | 'shi-Latn-MA' | 'shi-Tfng' | 'shi-Tfng-MA' | 'shs-CA' | 'si' | 'si-LK' | 'sid-ET' | 'sk' | 'sk-SK' | 'sl' | 'sl-SI' | 'sn' | 'sn-ZW' | 'so' | 'so-DJ' | 'so-ET' | 'so-KE' | 'so-SO' | 'sq' | 'sq-AL' | 'sq-MK' | 'sr' | 'sr-Cyrl' | 'sr-Cyrl-BA' | 'sr-Cyrl-ME' | 'sr-Cyrl-RS' | 'sr-Latn' | 'sr-Latn-BA' | 'sr-Latn-ME' | 'sr-Latn-RS' | 'sr-ME' | 'sr-RS' | 'ss-ZA' | 'st-ZA' | 'sv' | 'sv-FI' | 'sv-SE' | 'sw' | 'sw-KE' | 'sw-TZ' | 'ta' | 'ta-IN' | 'ta-LK' | 'te' | 'te-IN' | 'teo' | 'teo-KE' | 'teo-UG' | 'tg-TJ' | 'th' | 'th-TH' | 'ti' | 'ti-ER' | 'ti-ET' | 'tig-ER' | 'tk-TM' | 'tl-PH' | 'tn-ZA' | 'to' | 'to-TO' | 'tr' | 'tr-CY' | 'tr-TR' | 'ts-ZA' | 'tt-RU' | 'tzm' | 'tzm-Latn' | 'tzm-Latn-MA' | 'ug-CN' | 'uk' | 'uk-UA' | 'unm-US' | 'ur' | 'ur-IN' | 'ur-PK' | 'uz' | 'uz-Arab' | 'uz-Arab-AF' | 'uz-Cyrl' | 'uz-Cyrl-UZ' | 'uz-Latn' | 'uz-Latn-UZ' | 'uz-UZ' | 've-ZA' | 'vi' | 'vi-VN' | 'vun' | 'vun-TZ' | 'wa-BE' | 'wae-CH' | 'wal-ET' | 'wo-SN' | 'xh-ZA' | 'xog' | 'xog-UG' | 'yi-US' | 'yo' | 'yo-NG' | 'yue-HK' | 'zh' | 'zh-CN' | 'zh-HK' | 'zh-Hans' | 'zh-Hans-CN' | 'zh-Hans-HK' | 'zh-Hans-MO' | 'zh-Hans-SG' | 'zh-Hant' | 'zh-Hant-HK' | 'zh-Hant-MO' | 'zh-Hant-TW' | 'zh-SG' | 'zh-TW' | 'zu' | 'zu-ZA' | 'auto';
|
|
513
|
+
/**
|
|
514
|
+
* Selects which markdown conversion strategy to use. "full_page" converts the
|
|
515
|
+
* entire HTML page. "main_content" uses Mozilla Readability to extract the main
|
|
516
|
+
* article content before converting.
|
|
517
|
+
*/
|
|
518
|
+
markdown_backend?: 'full_page' | 'main_content';
|
|
519
|
+
/**
|
|
520
|
+
* HTTP method for the request
|
|
521
|
+
*/
|
|
522
|
+
method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
|
|
523
|
+
/**
|
|
524
|
+
* Filters for capturing network traffic
|
|
525
|
+
*/
|
|
526
|
+
network_capture?: Array<ExtractAsyncParams.NetworkCapture>;
|
|
527
|
+
/**
|
|
528
|
+
* Operating system to emulate
|
|
529
|
+
*/
|
|
530
|
+
os?: 'windows' | 'mac os' | 'linux' | 'android' | 'ios';
|
|
531
|
+
/**
|
|
532
|
+
* Whether to parse the response content
|
|
533
|
+
*/
|
|
534
|
+
parse?: boolean;
|
|
535
|
+
/**
|
|
536
|
+
* Custom parser configuration as a key-value map
|
|
537
|
+
*/
|
|
538
|
+
parser?: {
|
|
539
|
+
[key: string]: unknown;
|
|
540
|
+
} | string;
|
|
541
|
+
/**
|
|
542
|
+
* Referrer policy for the request
|
|
543
|
+
*/
|
|
544
|
+
referrer_type?: 'random' | 'no-referer' | 'same-origin' | 'google' | 'bing' | 'facebook' | 'twitter' | 'instagram';
|
|
545
|
+
/**
|
|
546
|
+
* Whether to render JavaScript content using a browser
|
|
547
|
+
*/
|
|
548
|
+
render?: boolean | 'auto';
|
|
549
|
+
/**
|
|
550
|
+
* Request timeout in milliseconds
|
|
551
|
+
*/
|
|
552
|
+
request_timeout?: number;
|
|
553
|
+
session?: ExtractAsyncParams.Session;
|
|
554
|
+
/**
|
|
555
|
+
* Skills or capabilities required for the request
|
|
556
|
+
*/
|
|
557
|
+
skill?: string | Array<string>;
|
|
558
|
+
/**
|
|
559
|
+
* US state for geolocation (only valid when country is US)
|
|
560
|
+
*/
|
|
561
|
+
state?: 'AL' | 'AK' | 'AS' | 'AZ' | 'AR' | 'CA' | 'CO' | 'CT' | 'DE' | 'DC' | 'FL' | 'GA' | 'GU' | 'HI' | 'ID' | 'IL' | 'IN' | 'IA' | 'KS' | 'KY' | 'LA' | 'ME' | 'MD' | 'MA' | 'MI' | 'MN' | 'MS' | 'MO' | 'MT' | 'NE' | 'NV' | 'NH' | 'NJ' | 'NM' | 'NY' | 'NC' | 'ND' | 'MP' | 'OH' | 'OK' | 'OR' | 'PA' | 'PR' | 'RI' | 'SC' | 'SD' | 'TN' | 'TX' | 'UT' | 'VT' | 'VA' | 'VI' | 'WA' | 'WV' | 'WI' | 'WY';
|
|
562
|
+
/**
|
|
563
|
+
* Whether to compress stored data
|
|
564
|
+
*/
|
|
565
|
+
storage_compress?: boolean;
|
|
566
|
+
/**
|
|
567
|
+
* Custom name for the stored object
|
|
568
|
+
*/
|
|
569
|
+
storage_object_name?: string;
|
|
570
|
+
/**
|
|
571
|
+
* Type of storage to use for results
|
|
572
|
+
*/
|
|
573
|
+
storage_type?: string;
|
|
574
|
+
/**
|
|
575
|
+
* URL for storage location
|
|
576
|
+
*/
|
|
577
|
+
storage_url?: string;
|
|
578
|
+
/**
|
|
579
|
+
* User-defined tag for request identification
|
|
580
|
+
*/
|
|
581
|
+
tag?: string;
|
|
582
|
+
}
|
|
583
|
+
export declare namespace ExtractAsyncParams {
|
|
584
|
+
interface UnionMember1 {
|
|
585
|
+
name: 'chrome' | 'firefox';
|
|
586
|
+
/**
|
|
587
|
+
* Specific browser version to emulate
|
|
588
|
+
*/
|
|
589
|
+
version?: string;
|
|
590
|
+
}
|
|
591
|
+
interface UnionMember0 {
|
|
592
|
+
creation?: string | null;
|
|
593
|
+
domain?: string | null;
|
|
594
|
+
expires?: string;
|
|
595
|
+
extensions?: Array<string> | null;
|
|
596
|
+
hostOnly?: boolean | null;
|
|
597
|
+
httpOnly?: boolean | null;
|
|
598
|
+
lastAccessed?: string | null;
|
|
599
|
+
maxAge?: 'Infinity' | '-Infinity' | number | null;
|
|
600
|
+
name?: string;
|
|
601
|
+
path?: string | null;
|
|
602
|
+
pathIsDefault?: boolean | null;
|
|
603
|
+
sameSite?: 'strict' | 'lax' | 'none';
|
|
604
|
+
secure?: boolean;
|
|
605
|
+
value?: string;
|
|
606
|
+
[k: string]: unknown;
|
|
607
|
+
}
|
|
608
|
+
interface NetworkCapture {
|
|
609
|
+
method?: 'GET' | 'HEAD' | 'POST' | 'PUT' | 'DELETE' | 'CONNECT' | 'OPTIONS' | 'TRACE' | 'PATCH';
|
|
610
|
+
/**
|
|
611
|
+
* Resource type for network capture filtering
|
|
612
|
+
*/
|
|
613
|
+
resource_type?: string | Array<string>;
|
|
614
|
+
status_code?: number | Array<number>;
|
|
615
|
+
url?: NetworkCapture.URL;
|
|
616
|
+
validation?: boolean;
|
|
617
|
+
wait_for_requests_count?: number;
|
|
618
|
+
wait_for_requests_count_timeout?: number;
|
|
619
|
+
}
|
|
620
|
+
namespace NetworkCapture {
|
|
621
|
+
interface URL {
|
|
622
|
+
value: string;
|
|
623
|
+
type?: 'exact' | 'contains';
|
|
624
|
+
}
|
|
625
|
+
}
|
|
626
|
+
interface Session {
|
|
627
|
+
id?: string;
|
|
628
|
+
prefetch_userbrowser?: boolean;
|
|
629
|
+
renew_on_blocked?: boolean;
|
|
630
|
+
retry?: boolean;
|
|
631
|
+
timeout?: number;
|
|
632
|
+
}
|
|
633
|
+
}
|
|
634
|
+
export interface ExtractBatchParams {
|
|
635
|
+
/**
|
|
636
|
+
* Array of extraction requests. Each object can include extraction parameters and
|
|
637
|
+
* async/storage settings.
|
|
638
|
+
*/
|
|
639
|
+
inputs: Array<ExtractBatchParams.Input>;
|
|
640
|
+
/**
|
|
641
|
+
* Shared parameters applied to the entire batch. Can include extraction parameters
|
|
642
|
+
* and async/storage settings.
|
|
643
|
+
*/
|
|
644
|
+
shared_inputs?: ExtractBatchParams.SharedInputs;
|
|
645
|
+
}
|
|
646
|
+
export declare namespace ExtractBatchParams {
|
|
647
|
+
interface Input {
|
|
648
|
+
/**
|
|
649
|
+
* Custom flow for the optimization engine: maps candidate names to the number of
|
|
650
|
+
* attempts to spend on each candidate before advancing (0 skips it). Key order
|
|
651
|
+
* defines the flow order. Providing it opts the request into 'auto' driver
|
|
652
|
+
* selection.
|
|
653
|
+
*/
|
|
654
|
+
auto_driver_configuration?: {
|
|
655
|
+
[key: string]: number;
|
|
656
|
+
};
|
|
657
|
+
/**
|
|
658
|
+
* Request body for POST, PUT, PATCH methods
|
|
659
|
+
*/
|
|
660
|
+
body?: unknown;
|
|
661
|
+
/**
|
|
662
|
+
* Browser type to emulate
|
|
663
|
+
*/
|
|
664
|
+
browser?: 'chrome' | 'firefox' | Input.UnionMember1;
|
|
665
|
+
/**
|
|
666
|
+
* Array of browser automation actions to execute sequentially
|
|
667
|
+
*/
|
|
668
|
+
browser_actions?: Array<Shared.AutoScrollAction | Shared.ClickAction | Shared.EvalAction | Shared.FetchAction | Shared.FillAction | Shared.GetCookiesAction | Shared.GotoAction | Shared.PressAction | Shared.ScreenshotAction | Shared.ScrollAction | Shared.WaitAction | Shared.WaitForElementAction | Shared.WaitForNavigationAction>;
|
|
669
|
+
/**
|
|
670
|
+
* URL to call back when async operation completes
|
|
671
|
+
*/
|
|
672
|
+
callback_url?: string;
|
|
673
|
+
/**
|
|
674
|
+
* City for geolocation
|
|
675
|
+
*/
|
|
676
|
+
city?: string;
|
|
677
|
+
/**
|
|
678
|
+
* Whether to automatically handle cookie consent headers
|
|
679
|
+
*/
|
|
680
|
+
consent_header?: boolean;
|
|
681
|
+
/**
|
|
682
|
+
* Browser cookies as array of cookie objects
|
|
683
|
+
*/
|
|
684
|
+
cookies?: Array<Input.UnionMember0> | string;
|
|
685
|
+
/**
|
|
686
|
+
* Country code for geolocation and proxy selection
|
|
687
|
+
*/
|
|
688
|
+
country?: 'AD' | 'AE' | 'AF' | 'AG' | 'AI' | 'AL' | 'AM' | 'AO' | 'AQ' | 'AR' | 'AS' | 'AT' | 'AU' | 'AW' | 'AX' | 'AZ' | 'BA' | 'BB' | 'BD' | 'BE' | 'BF' | 'BG' | 'BH' | 'BI' | 'BJ' | 'BL' | 'BM' | 'BN' | 'BO' | 'BQ' | 'BR' | 'BS' | 'BT' | 'BV' | 'BW' | 'BY' | 'BZ' | 'CA' | 'CC' | 'CD' | 'CF' | 'CG' | 'CH' | 'CI' | 'CK' | 'CL' | 'CM' | 'CN' | 'CO' | 'CR' | 'CU' | 'CV' | 'CW' | 'CX' | 'CY' | 'CZ' | 'DE' | 'DJ' | 'DK' | 'DM' | 'DO' | 'DZ' | 'EC' | 'EE' | 'EG' | 'EH' | 'ER' | 'ES' | 'ET' | 'FI' | 'FJ' | 'FK' | 'FM' | 'FO' | 'FR' | 'GA' | 'GB' | 'GD' | 'GE' | 'GF' | 'GG' | 'GH' | 'GI' | 'GL' | 'GM' | 'GN' | 'GP' | 'GQ' | 'GR' | 'GS' | 'GT' | 'GU' | 'GW' | 'GY' | 'HK' | 'HM' | 'HN' | 'HR' | 'HT' | 'HU' | 'ID' | 'IE' | 'IL' | 'IM' | 'IN' | 'IO' | 'IQ' | 'IR' | 'IS' | 'IT' | 'JE' | 'JM' | 'JO' | 'JP' | 'KE' | 'KG' | 'KH' | 'KI' | 'KM' | 'KN' | 'KP' | 'KR' | 'KW' | 'KY' | 'KZ' | 'LA' | 'LB' | 'LC' | 'LI' | 'LK' | 'LR' | 'LS' | 'LT' | 'LU' | 'LV' | 'LY' | 'MA' | 'MC' | 'MD' | 'ME' | 'MF' | 'MG' | 'MH' | 'MK' | 'ML' | 'MM' | 'MN' | 'MO' | 'MP' | 'MQ' | 'MR' | 'MS' | 'MT' | 'MU' | 'MV' | 'MW' | 'MX' | 'MY' | 'MZ' | 'NA' | 'NC' | 'NE' | 'NF' | 'NG' | 'NI' | 'NL' | 'NO' | 'NP' | 'NR' | 'NU' | 'NZ' | 'OM' | 'PA' | 'PE' | 'PF' | 'PG' | 'PH' | 'PK' | 'PL' | 'PM' | 'PN' | 'PR' | 'PS' | 'PT' | 'PW' | 'PY' | 'QA' | 'RE' | 'RO' | 'RS' | 'RU' | 'RW' | 'SA' | 'SB' | 'SC' | 'SD' | 'SE' | 'SG' | 'SH' | 'SI' | 'SJ' | 'SK' | 'SL' | 'SM' | 'SN' | 'SO' | 'SR' | 'SS' | 'ST' | 'SV' | 'SX' | 'SY' | 'SZ' | 'TC' | 'TD' | 'TF' | 'TG' | 'TH' | 'TJ' | 'TK' | 'TL' | 'TM' | 'TN' | 'TO' | 'TR' | 'TT' | 'TV' | 'TW' | 'TZ' | 'UA' | 'UG' | 'UM' | 'US' | 'UY' | 'UZ' | 'VA' | 'VC' | 'VE' | 'VG' | 'VI' | 'VN' | 'VU' | 'WF' | 'WS' | 'XK' | 'YE' | 'YT' | 'ZA' | 'ZM' | 'ZW' | 'ALL';
|
|
689
|
+
/**
|
|
690
|
+
* Device type for browser emulation
|
|
691
|
+
*/
|
|
692
|
+
device?: 'desktop' | 'mobile' | 'tablet';
|
|
693
|
+
/**
|
|
694
|
+
* Browser driver to use. Use 'auto' to let the engine select the candidate config
|
|
695
|
+
* per domain.
|
|
696
|
+
*/
|
|
697
|
+
driver?: 'auto' | 'vx6' | 'vx8' | 'vx8-pro' | 'vx10' | 'vx10-pro' | 'vx12' | 'vx12-pro' | 'media-vx6' | 'fast-vx6';
|
|
698
|
+
/**
|
|
699
|
+
* Expected HTTP status codes for successful requests
|
|
700
|
+
*/
|
|
701
|
+
expected_status_codes?: Array<number>;
|
|
702
|
+
/**
|
|
703
|
+
* List of acceptable response formats in order of preference
|
|
704
|
+
*/
|
|
705
|
+
formats?: Array<'html' | 'markdown' | 'screenshot' | 'headers' | 'links'>;
|
|
706
|
+
/**
|
|
707
|
+
* Custom HTTP headers to include in the request
|
|
708
|
+
*/
|
|
709
|
+
headers?: {
|
|
710
|
+
[key: string]: string | Array<string> | null;
|
|
711
|
+
};
|
|
712
|
+
/**
|
|
713
|
+
* Whether to use HTTP/2 protocol
|
|
714
|
+
*/
|
|
715
|
+
http2?: boolean;
|
|
716
|
+
/**
|
|
717
|
+
* Whether to emulate XMLHttpRequest behavior
|
|
718
|
+
*/
|
|
719
|
+
is_xhr?: boolean;
|
|
720
|
+
/**
|
|
721
|
+
* Locale for browser language and region settings
|
|
722
|
+
*/
|
|
723
|
+
locale?: 'aa-DJ' | 'aa-ER' | 'aa-ET' | 'af' | 'af-NA' | 'af-ZA' | 'ak' | 'ak-GH' | 'am' | 'am-ET' | 'an-ES' | 'ar' | 'ar-AE' | 'ar-BH' | 'ar-DZ' | 'ar-EG' | 'ar-IN' | 'ar-IQ' | 'ar-JO' | 'ar-KW' | 'ar-LB' | 'ar-LY' | 'ar-MA' | 'ar-OM' | 'ar-QA' | 'ar-SA' | 'ar-SD' | 'ar-SY' | 'ar-TN' | 'ar-YE' | 'as' | 'as-IN' | 'asa' | 'asa-TZ' | 'ast-ES' | 'az' | 'az-AZ' | 'az-Cyrl' | 'az-Cyrl-AZ' | 'az-Latn' | 'az-Latn-AZ' | 'be' | 'be-BY' | 'bem' | 'bem-ZM' | 'ber-DZ' | 'ber-MA' | 'bez' | 'bez-TZ' | 'bg' | 'bg-BG' | 'bho-IN' | 'bm' | 'bm-ML' | 'bn' | 'bn-BD' | 'bn-IN' | 'bo' | 'bo-CN' | 'bo-IN' | 'br-FR' | 'brx-IN' | 'bs' | 'bs-BA' | 'byn-ER' | 'ca' | 'ca-AD' | 'ca-ES' | 'ca-FR' | 'ca-IT' | 'cgg' | 'cgg-UG' | 'chr' | 'chr-US' | 'crh-UA' | 'cs' | 'cs-CZ' | 'csb-PL' | 'cv-RU' | 'cy' | 'cy-GB' | 'da' | 'da-DK' | 'dav' | 'dav-KE' | 'de' | 'de-AT' | 'de-BE' | 'de-CH' | 'de-DE' | 'de-LI' | 'de-LU' | 'dv-MV' | 'dz-BT' | 'ebu' | 'ebu-KE' | 'ee' | 'ee-GH' | 'ee-TG' | 'el' | 'el-CY' | 'el-GR' | 'en' | 'en-AG' | 'en-AS' | 'en-AU' | 'en-BE' | 'en-BW' | 'en-BZ' | 'en-CA' | 'en-DK' | 'en-GB' | 'en-GU' | 'en-HK' | 'en-IE' | 'en-IN' | 'en-JM' | 'en-MH' | 'en-MP' | 'en-MT' | 'en-MU' | 'en-NA' | 'en-NG' | 'en-NZ' | 'en-PH' | 'en-PK' | 'en-SG' | 'en-TT' | 'en-UM' | 'en-US' | 'en-VI' | 'en-ZA' | 'en-ZM' | 'en-ZW' | 'eo' | 'es' | 'es-419' | 'es-AR' | 'es-BO' | 'es-CL' | 'es-CO' | 'es-CR' | 'es-CU' | 'es-DO' | 'es-EC' | 'es-ES' | 'es-GQ' | 'es-GT' | 'es-HN' | 'es-MX' | 'es-NI' | 'es-PA' | 'es-PE' | 'es-PR' | 'es-PY' | 'es-SV' | 'es-US' | 'es-UY' | 'es-VE' | 'et' | 'et-EE' | 'eu' | 'eu-ES' | 'fa' | 'fa-AF' | 'fa-IR' | 'ff' | 'ff-SN' | 'fi' | 'fi-FI' | 'fil' | 'fil-PH' | 'fo' | 'fo-FO' | 'fr' | 'fr-BE' | 'fr-BF' | 'fr-BI' | 'fr-BJ' | 'fr-BL' | 'fr-CA' | 'fr-CD' | 'fr-CF' | 'fr-CG' | 'fr-CH' | 'fr-CI' | 'fr-CM' | 'fr-DJ' | 'fr-FR' | 'fr-GA' | 'fr-GN' | 'fr-GP' | 'fr-GQ' | 'fr-KM' | 'fr-LU' | 'fr-MC' | 'fr-MF' | 'fr-MG' | 'fr-ML' | 'fr-MQ' | 'fr-NE' | 'fr-RE' | 'fr-RW' | 'fr-SN' | 'fr-TD' | 'fr-TG' | 'fur-IT' | 'fy-DE' | 'fy-NL' | 'ga' | 'ga-IE' | 'gd-GB' | 'gez-ER' | 'gez-ET' | 'gl' | 'gl-ES' | 'gsw' | 'gsw-CH' | 'gu' | 'gu-IN' | 'guz' | 'guz-KE' | 'gv' | 'gv-GB' | 'ha' | 'ha-Latn' | 'ha-Latn-GH' | 'ha-Latn-NE' | 'ha-Latn-NG' | 'ha-NG' | 'haw' | 'haw-US' | 'he' | 'he-IL' | 'hi' | 'hi-IN' | 'hne-IN' | 'hr' | 'hr-HR' | 'hsb-DE' | 'ht-HT' | 'hu' | 'hu-HU' | 'hy' | 'hy-AM' | 'id' | 'id-ID' | 'ig' | 'ig-NG' | 'ii' | 'ii-CN' | 'ik-CA' | 'is' | 'is-IS' | 'it' | 'it-CH' | 'it-IT' | 'iu-CA' | 'iw-IL' | 'ja' | 'ja-JP' | 'jmc' | 'jmc-TZ' | 'ka' | 'ka-GE' | 'kab' | 'kab-DZ' | 'kam' | 'kam-KE' | 'kde' | 'kde-TZ' | 'kea' | 'kea-CV' | 'khq' | 'khq-ML' | 'ki' | 'ki-KE' | 'kk' | 'kk-Cyrl' | 'kk-Cyrl-KZ' | 'kk-KZ' | 'kl' | 'kl-GL' | 'kln' | 'kln-KE' | 'km' | 'km-KH' | 'kn' | 'kn-IN' | 'ko' | 'ko-KR' | 'kok' | 'kok-IN' | 'ks-IN' | 'ku-TR' | 'kw' | 'kw-GB' | 'ky-KG' | 'lag' | 'lag-TZ' | 'lb-LU' | 'lg' | 'lg-UG' | 'li-BE' | 'li-NL' | 'lij-IT' | 'lo-LA' | 'lt' | 'lt-LT' | 'luo' | 'luo-KE' | 'luy' | 'luy-KE' | 'lv' | 'lv-LV' | 'mag-IN' | 'mai-IN' | 'mas' | 'mas-KE' | 'mas-TZ' | 'mer' | 'mer-KE' | 'mfe' | 'mfe-MU' | 'mg' | 'mg-MG' | 'mhr-RU' | 'mi-NZ' | 'mk' | 'mk-MK' | 'ml' | 'ml-IN' | 'mn-MN' | 'mr' | 'mr-IN' | 'ms' | 'ms-BN' | 'ms-MY' | 'mt' | 'mt-MT' | 'my' | 'my-MM' | 'nan-TW' | 'naq' | 'naq-NA' | 'nb' | 'nb-NO' | 'nd' | 'nd-ZW' | 'nds-DE' | 'nds-NL' | 'ne' | 'ne-IN' | 'ne-NP' | 'nl' | 'nl-AW' | 'nl-BE' | 'nl-NL' | 'nn' | 'nn-NO' | 'nr-ZA' | 'nso-ZA' | 'nyn' | 'nyn-UG' | 'oc-FR' | 'om' | 'om-ET' | 'om-KE' | 'or' | 'or-IN' | 'os-RU' | 'pa' | 'pa-Arab' | 'pa-Arab-PK' | 'pa-Guru' | 'pa-Guru-IN' | 'pa-IN' | 'pa-PK' | 'pap-AN' | 'pl' | 'pl-PL' | 'ps' | 'ps-AF' | 'pt' | 'pt-BR' | 'pt-GW' | 'pt-MZ' | 'pt-PT' | 'rm' | 'rm-CH' | 'ro' | 'ro-MD' | 'ro-RO' | 'rof' | 'rof-TZ' | 'ru' | 'ru-MD' | 'ru-RU' | 'ru-UA' | 'rw' | 'rw-RW' | 'rwk' | 'rwk-TZ' | 'sa-IN' | 'saq' | 'saq-KE' | 'sc-IT' | 'sd-IN' | 'se-NO' | 'seh' | 'seh-MZ' | 'ses' | 'ses-ML' | 'sg' | 'sg-CF' | 'shi' | 'shi-Latn' | 'shi-Latn-MA' | 'shi-Tfng' | 'shi-Tfng-MA' | 'shs-CA' | 'si' | 'si-LK' | 'sid-ET' | 'sk' | 'sk-SK' | 'sl' | 'sl-SI' | 'sn' | 'sn-ZW' | 'so' | 'so-DJ' | 'so-ET' | 'so-KE' | 'so-SO' | 'sq' | 'sq-AL' | 'sq-MK' | 'sr' | 'sr-Cyrl' | 'sr-Cyrl-BA' | 'sr-Cyrl-ME' | 'sr-Cyrl-RS' | 'sr-Latn' | 'sr-Latn-BA' | 'sr-Latn-ME' | 'sr-Latn-RS' | 'sr-ME' | 'sr-RS' | 'ss-ZA' | 'st-ZA' | 'sv' | 'sv-FI' | 'sv-SE' | 'sw' | 'sw-KE' | 'sw-TZ' | 'ta' | 'ta-IN' | 'ta-LK' | 'te' | 'te-IN' | 'teo' | 'teo-KE' | 'teo-UG' | 'tg-TJ' | 'th' | 'th-TH' | 'ti' | 'ti-ER' | 'ti-ET' | 'tig-ER' | 'tk-TM' | 'tl-PH' | 'tn-ZA' | 'to' | 'to-TO' | 'tr' | 'tr-CY' | 'tr-TR' | 'ts-ZA' | 'tt-RU' | 'tzm' | 'tzm-Latn' | 'tzm-Latn-MA' | 'ug-CN' | 'uk' | 'uk-UA' | 'unm-US' | 'ur' | 'ur-IN' | 'ur-PK' | 'uz' | 'uz-Arab' | 'uz-Arab-AF' | 'uz-Cyrl' | 'uz-Cyrl-UZ' | 'uz-Latn' | 'uz-Latn-UZ' | 'uz-UZ' | 've-ZA' | 'vi' | 'vi-VN' | 'vun' | 'vun-TZ' | 'wa-BE' | 'wae-CH' | 'wal-ET' | 'wo-SN' | 'xh-ZA' | 'xog' | 'xog-UG' | 'yi-US' | 'yo' | 'yo-NG' | 'yue-HK' | 'zh' | 'zh-CN' | 'zh-HK' | 'zh-Hans' | 'zh-Hans-CN' | 'zh-Hans-HK' | 'zh-Hans-MO' | 'zh-Hans-SG' | 'zh-Hant' | 'zh-Hant-HK' | 'zh-Hant-MO' | 'zh-Hant-TW' | 'zh-SG' | 'zh-TW' | 'zu' | 'zu-ZA' | 'auto';
|
|
724
|
+
/**
|
|
725
|
+
* Selects which markdown conversion strategy to use. "full_page" converts the
|
|
726
|
+
* entire HTML page. "main_content" uses Mozilla Readability to extract the main
|
|
727
|
+
* article content before converting.
|
|
728
|
+
*/
|
|
729
|
+
markdown_backend?: 'full_page' | 'main_content';
|
|
730
|
+
/**
|
|
731
|
+
* HTTP method for the request
|
|
732
|
+
*/
|
|
733
|
+
method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
|
|
734
|
+
/**
|
|
735
|
+
* Filters for capturing network traffic
|
|
736
|
+
*/
|
|
737
|
+
network_capture?: Array<Input.NetworkCapture>;
|
|
738
|
+
/**
|
|
739
|
+
* Operating system to emulate
|
|
740
|
+
*/
|
|
741
|
+
os?: 'windows' | 'mac os' | 'linux' | 'android' | 'ios';
|
|
742
|
+
/**
|
|
743
|
+
* Whether to parse the response content
|
|
744
|
+
*/
|
|
745
|
+
parse?: boolean;
|
|
746
|
+
/**
|
|
747
|
+
* Custom parser configuration as a key-value map
|
|
748
|
+
*/
|
|
749
|
+
parser?: {
|
|
750
|
+
[key: string]: unknown;
|
|
751
|
+
} | string;
|
|
752
|
+
/**
|
|
753
|
+
* Referrer policy for the request
|
|
754
|
+
*/
|
|
755
|
+
referrer_type?: 'random' | 'no-referer' | 'same-origin' | 'google' | 'bing' | 'facebook' | 'twitter' | 'instagram';
|
|
756
|
+
/**
|
|
757
|
+
* Whether to render JavaScript content using a browser. Use 'auto' to let the
|
|
758
|
+
* engine select the candidate config per domain.
|
|
759
|
+
*/
|
|
760
|
+
render?: boolean | 'auto';
|
|
761
|
+
/**
|
|
762
|
+
* Request timeout in milliseconds
|
|
763
|
+
*/
|
|
764
|
+
request_timeout?: number;
|
|
765
|
+
session?: Input.Session;
|
|
766
|
+
/**
|
|
767
|
+
* Skills or capabilities required for the request
|
|
768
|
+
*/
|
|
769
|
+
skill?: string | Array<string>;
|
|
770
|
+
/**
|
|
771
|
+
* US state for geolocation (only valid when country is US)
|
|
772
|
+
*/
|
|
773
|
+
state?: 'AL' | 'AK' | 'AS' | 'AZ' | 'AR' | 'CA' | 'CO' | 'CT' | 'DE' | 'DC' | 'FL' | 'GA' | 'GU' | 'HI' | 'ID' | 'IL' | 'IN' | 'IA' | 'KS' | 'KY' | 'LA' | 'ME' | 'MD' | 'MA' | 'MI' | 'MN' | 'MS' | 'MO' | 'MT' | 'NE' | 'NV' | 'NH' | 'NJ' | 'NM' | 'NY' | 'NC' | 'ND' | 'MP' | 'OH' | 'OK' | 'OR' | 'PA' | 'PR' | 'RI' | 'SC' | 'SD' | 'TN' | 'TX' | 'UT' | 'VT' | 'VA' | 'VI' | 'WA' | 'WV' | 'WI' | 'WY';
|
|
774
|
+
/**
|
|
775
|
+
* Whether to compress stored data
|
|
776
|
+
*/
|
|
777
|
+
storage_compress?: boolean;
|
|
778
|
+
/**
|
|
779
|
+
* Custom name for the stored object
|
|
780
|
+
*/
|
|
781
|
+
storage_object_name?: string;
|
|
782
|
+
/**
|
|
783
|
+
* Type of storage to use for results
|
|
784
|
+
*/
|
|
785
|
+
storage_type?: string;
|
|
786
|
+
/**
|
|
787
|
+
* URL for storage location
|
|
788
|
+
*/
|
|
789
|
+
storage_url?: string;
|
|
790
|
+
/**
|
|
791
|
+
* User-defined tag for request identification
|
|
792
|
+
*/
|
|
793
|
+
tag?: string;
|
|
794
|
+
/**
|
|
795
|
+
* Target URL to scrape
|
|
796
|
+
*/
|
|
797
|
+
url?: string;
|
|
798
|
+
}
|
|
799
|
+
namespace Input {
|
|
800
|
+
interface UnionMember1 {
|
|
801
|
+
name: 'chrome' | 'firefox';
|
|
802
|
+
/**
|
|
803
|
+
* Specific browser version to emulate
|
|
804
|
+
*/
|
|
805
|
+
version?: string;
|
|
806
|
+
}
|
|
807
|
+
interface UnionMember0 {
|
|
808
|
+
creation?: string | null;
|
|
809
|
+
domain?: string | null;
|
|
810
|
+
expires?: string;
|
|
811
|
+
extensions?: Array<string> | null;
|
|
812
|
+
hostOnly?: boolean | null;
|
|
813
|
+
httpOnly?: boolean | null;
|
|
814
|
+
lastAccessed?: string | null;
|
|
815
|
+
maxAge?: 'Infinity' | '-Infinity' | number | null;
|
|
816
|
+
name?: string;
|
|
817
|
+
path?: string | null;
|
|
818
|
+
pathIsDefault?: boolean | null;
|
|
819
|
+
sameSite?: 'strict' | 'lax' | 'none';
|
|
820
|
+
secure?: boolean;
|
|
821
|
+
value?: string;
|
|
822
|
+
[k: string]: unknown;
|
|
823
|
+
}
|
|
824
|
+
interface NetworkCapture {
|
|
825
|
+
method?: 'GET' | 'HEAD' | 'POST' | 'PUT' | 'DELETE' | 'CONNECT' | 'OPTIONS' | 'TRACE' | 'PATCH';
|
|
826
|
+
/**
|
|
827
|
+
* Resource type for network capture filtering
|
|
828
|
+
*/
|
|
829
|
+
resource_type?: string | Array<string>;
|
|
830
|
+
status_code?: number | Array<number>;
|
|
831
|
+
url?: NetworkCapture.URL;
|
|
832
|
+
validation?: boolean;
|
|
833
|
+
wait_for_requests_count?: number;
|
|
834
|
+
wait_for_requests_count_timeout?: number;
|
|
835
|
+
}
|
|
836
|
+
namespace NetworkCapture {
|
|
837
|
+
interface URL {
|
|
838
|
+
value: string;
|
|
839
|
+
type?: 'exact' | 'contains';
|
|
840
|
+
}
|
|
841
|
+
}
|
|
842
|
+
interface Session {
|
|
843
|
+
id?: string;
|
|
844
|
+
prefetch_userbrowser?: boolean;
|
|
845
|
+
renew_on_blocked?: boolean;
|
|
846
|
+
retry?: boolean;
|
|
847
|
+
timeout?: number;
|
|
848
|
+
}
|
|
849
|
+
}
|
|
850
|
+
/**
|
|
851
|
+
* Shared parameters applied to the entire batch. Can include extraction parameters
|
|
852
|
+
* and async/storage settings.
|
|
853
|
+
*/
|
|
854
|
+
interface SharedInputs {
|
|
855
|
+
/**
|
|
856
|
+
* Custom flow for the optimization engine: maps candidate names to the number of
|
|
857
|
+
* attempts to spend on each candidate before advancing (0 skips it). Key order
|
|
858
|
+
* defines the flow order. Providing it opts the request into 'auto' driver
|
|
859
|
+
* selection.
|
|
860
|
+
*/
|
|
861
|
+
auto_driver_configuration?: {
|
|
862
|
+
[key: string]: number;
|
|
863
|
+
};
|
|
864
|
+
/**
|
|
865
|
+
* Request body for POST, PUT, PATCH methods
|
|
866
|
+
*/
|
|
867
|
+
body?: unknown;
|
|
868
|
+
/**
|
|
869
|
+
* Browser type to emulate
|
|
870
|
+
*/
|
|
871
|
+
browser?: 'chrome' | 'firefox' | SharedInputs.UnionMember1;
|
|
872
|
+
/**
|
|
873
|
+
* Array of browser automation actions to execute sequentially
|
|
874
|
+
*/
|
|
875
|
+
browser_actions?: Array<Shared.AutoScrollAction | Shared.ClickAction | Shared.EvalAction | Shared.FetchAction | Shared.FillAction | Shared.GetCookiesAction | Shared.GotoAction | Shared.PressAction | Shared.ScreenshotAction | Shared.ScrollAction | Shared.WaitAction | Shared.WaitForElementAction | Shared.WaitForNavigationAction>;
|
|
876
|
+
/**
|
|
877
|
+
* URL to call back when async operation completes
|
|
878
|
+
*/
|
|
879
|
+
callback_url?: string;
|
|
880
|
+
/**
|
|
881
|
+
* City for geolocation
|
|
882
|
+
*/
|
|
883
|
+
city?: string;
|
|
884
|
+
/**
|
|
885
|
+
* Whether to automatically handle cookie consent headers
|
|
886
|
+
*/
|
|
887
|
+
consent_header?: boolean;
|
|
888
|
+
/**
|
|
889
|
+
* Browser cookies as array of cookie objects
|
|
890
|
+
*/
|
|
891
|
+
cookies?: Array<SharedInputs.UnionMember0> | string;
|
|
892
|
+
/**
|
|
893
|
+
* Country code for geolocation and proxy selection
|
|
894
|
+
*/
|
|
895
|
+
country?: 'AD' | 'AE' | 'AF' | 'AG' | 'AI' | 'AL' | 'AM' | 'AO' | 'AQ' | 'AR' | 'AS' | 'AT' | 'AU' | 'AW' | 'AX' | 'AZ' | 'BA' | 'BB' | 'BD' | 'BE' | 'BF' | 'BG' | 'BH' | 'BI' | 'BJ' | 'BL' | 'BM' | 'BN' | 'BO' | 'BQ' | 'BR' | 'BS' | 'BT' | 'BV' | 'BW' | 'BY' | 'BZ' | 'CA' | 'CC' | 'CD' | 'CF' | 'CG' | 'CH' | 'CI' | 'CK' | 'CL' | 'CM' | 'CN' | 'CO' | 'CR' | 'CU' | 'CV' | 'CW' | 'CX' | 'CY' | 'CZ' | 'DE' | 'DJ' | 'DK' | 'DM' | 'DO' | 'DZ' | 'EC' | 'EE' | 'EG' | 'EH' | 'ER' | 'ES' | 'ET' | 'FI' | 'FJ' | 'FK' | 'FM' | 'FO' | 'FR' | 'GA' | 'GB' | 'GD' | 'GE' | 'GF' | 'GG' | 'GH' | 'GI' | 'GL' | 'GM' | 'GN' | 'GP' | 'GQ' | 'GR' | 'GS' | 'GT' | 'GU' | 'GW' | 'GY' | 'HK' | 'HM' | 'HN' | 'HR' | 'HT' | 'HU' | 'ID' | 'IE' | 'IL' | 'IM' | 'IN' | 'IO' | 'IQ' | 'IR' | 'IS' | 'IT' | 'JE' | 'JM' | 'JO' | 'JP' | 'KE' | 'KG' | 'KH' | 'KI' | 'KM' | 'KN' | 'KP' | 'KR' | 'KW' | 'KY' | 'KZ' | 'LA' | 'LB' | 'LC' | 'LI' | 'LK' | 'LR' | 'LS' | 'LT' | 'LU' | 'LV' | 'LY' | 'MA' | 'MC' | 'MD' | 'ME' | 'MF' | 'MG' | 'MH' | 'MK' | 'ML' | 'MM' | 'MN' | 'MO' | 'MP' | 'MQ' | 'MR' | 'MS' | 'MT' | 'MU' | 'MV' | 'MW' | 'MX' | 'MY' | 'MZ' | 'NA' | 'NC' | 'NE' | 'NF' | 'NG' | 'NI' | 'NL' | 'NO' | 'NP' | 'NR' | 'NU' | 'NZ' | 'OM' | 'PA' | 'PE' | 'PF' | 'PG' | 'PH' | 'PK' | 'PL' | 'PM' | 'PN' | 'PR' | 'PS' | 'PT' | 'PW' | 'PY' | 'QA' | 'RE' | 'RO' | 'RS' | 'RU' | 'RW' | 'SA' | 'SB' | 'SC' | 'SD' | 'SE' | 'SG' | 'SH' | 'SI' | 'SJ' | 'SK' | 'SL' | 'SM' | 'SN' | 'SO' | 'SR' | 'SS' | 'ST' | 'SV' | 'SX' | 'SY' | 'SZ' | 'TC' | 'TD' | 'TF' | 'TG' | 'TH' | 'TJ' | 'TK' | 'TL' | 'TM' | 'TN' | 'TO' | 'TR' | 'TT' | 'TV' | 'TW' | 'TZ' | 'UA' | 'UG' | 'UM' | 'US' | 'UY' | 'UZ' | 'VA' | 'VC' | 'VE' | 'VG' | 'VI' | 'VN' | 'VU' | 'WF' | 'WS' | 'XK' | 'YE' | 'YT' | 'ZA' | 'ZM' | 'ZW' | 'ALL';
|
|
896
|
+
/**
|
|
897
|
+
* Device type for browser emulation
|
|
898
|
+
*/
|
|
899
|
+
device?: 'desktop' | 'mobile' | 'tablet';
|
|
900
|
+
/**
|
|
901
|
+
* Browser driver to use. Use 'auto' to let the engine select the candidate config
|
|
902
|
+
* per domain.
|
|
903
|
+
*/
|
|
904
|
+
driver?: 'auto' | 'vx6' | 'vx8' | 'vx8-pro' | 'vx10' | 'vx10-pro' | 'vx12' | 'vx12-pro' | 'media-vx6' | 'fast-vx6';
|
|
905
|
+
/**
|
|
906
|
+
* Expected HTTP status codes for successful requests
|
|
907
|
+
*/
|
|
908
|
+
expected_status_codes?: Array<number>;
|
|
909
|
+
/**
|
|
910
|
+
* List of acceptable response formats in order of preference
|
|
911
|
+
*/
|
|
912
|
+
formats?: Array<'html' | 'markdown' | 'screenshot' | 'headers' | 'links'>;
|
|
913
|
+
/**
|
|
914
|
+
* Custom HTTP headers to include in the request
|
|
915
|
+
*/
|
|
916
|
+
headers?: {
|
|
917
|
+
[key: string]: string | Array<string> | null;
|
|
918
|
+
};
|
|
919
|
+
/**
|
|
920
|
+
* Whether to use HTTP/2 protocol
|
|
921
|
+
*/
|
|
922
|
+
http2?: boolean;
|
|
923
|
+
/**
|
|
924
|
+
* Whether to emulate XMLHttpRequest behavior
|
|
925
|
+
*/
|
|
926
|
+
is_xhr?: boolean;
|
|
927
|
+
/**
|
|
928
|
+
* Locale for browser language and region settings
|
|
929
|
+
*/
|
|
930
|
+
locale?: 'aa-DJ' | 'aa-ER' | 'aa-ET' | 'af' | 'af-NA' | 'af-ZA' | 'ak' | 'ak-GH' | 'am' | 'am-ET' | 'an-ES' | 'ar' | 'ar-AE' | 'ar-BH' | 'ar-DZ' | 'ar-EG' | 'ar-IN' | 'ar-IQ' | 'ar-JO' | 'ar-KW' | 'ar-LB' | 'ar-LY' | 'ar-MA' | 'ar-OM' | 'ar-QA' | 'ar-SA' | 'ar-SD' | 'ar-SY' | 'ar-TN' | 'ar-YE' | 'as' | 'as-IN' | 'asa' | 'asa-TZ' | 'ast-ES' | 'az' | 'az-AZ' | 'az-Cyrl' | 'az-Cyrl-AZ' | 'az-Latn' | 'az-Latn-AZ' | 'be' | 'be-BY' | 'bem' | 'bem-ZM' | 'ber-DZ' | 'ber-MA' | 'bez' | 'bez-TZ' | 'bg' | 'bg-BG' | 'bho-IN' | 'bm' | 'bm-ML' | 'bn' | 'bn-BD' | 'bn-IN' | 'bo' | 'bo-CN' | 'bo-IN' | 'br-FR' | 'brx-IN' | 'bs' | 'bs-BA' | 'byn-ER' | 'ca' | 'ca-AD' | 'ca-ES' | 'ca-FR' | 'ca-IT' | 'cgg' | 'cgg-UG' | 'chr' | 'chr-US' | 'crh-UA' | 'cs' | 'cs-CZ' | 'csb-PL' | 'cv-RU' | 'cy' | 'cy-GB' | 'da' | 'da-DK' | 'dav' | 'dav-KE' | 'de' | 'de-AT' | 'de-BE' | 'de-CH' | 'de-DE' | 'de-LI' | 'de-LU' | 'dv-MV' | 'dz-BT' | 'ebu' | 'ebu-KE' | 'ee' | 'ee-GH' | 'ee-TG' | 'el' | 'el-CY' | 'el-GR' | 'en' | 'en-AG' | 'en-AS' | 'en-AU' | 'en-BE' | 'en-BW' | 'en-BZ' | 'en-CA' | 'en-DK' | 'en-GB' | 'en-GU' | 'en-HK' | 'en-IE' | 'en-IN' | 'en-JM' | 'en-MH' | 'en-MP' | 'en-MT' | 'en-MU' | 'en-NA' | 'en-NG' | 'en-NZ' | 'en-PH' | 'en-PK' | 'en-SG' | 'en-TT' | 'en-UM' | 'en-US' | 'en-VI' | 'en-ZA' | 'en-ZM' | 'en-ZW' | 'eo' | 'es' | 'es-419' | 'es-AR' | 'es-BO' | 'es-CL' | 'es-CO' | 'es-CR' | 'es-CU' | 'es-DO' | 'es-EC' | 'es-ES' | 'es-GQ' | 'es-GT' | 'es-HN' | 'es-MX' | 'es-NI' | 'es-PA' | 'es-PE' | 'es-PR' | 'es-PY' | 'es-SV' | 'es-US' | 'es-UY' | 'es-VE' | 'et' | 'et-EE' | 'eu' | 'eu-ES' | 'fa' | 'fa-AF' | 'fa-IR' | 'ff' | 'ff-SN' | 'fi' | 'fi-FI' | 'fil' | 'fil-PH' | 'fo' | 'fo-FO' | 'fr' | 'fr-BE' | 'fr-BF' | 'fr-BI' | 'fr-BJ' | 'fr-BL' | 'fr-CA' | 'fr-CD' | 'fr-CF' | 'fr-CG' | 'fr-CH' | 'fr-CI' | 'fr-CM' | 'fr-DJ' | 'fr-FR' | 'fr-GA' | 'fr-GN' | 'fr-GP' | 'fr-GQ' | 'fr-KM' | 'fr-LU' | 'fr-MC' | 'fr-MF' | 'fr-MG' | 'fr-ML' | 'fr-MQ' | 'fr-NE' | 'fr-RE' | 'fr-RW' | 'fr-SN' | 'fr-TD' | 'fr-TG' | 'fur-IT' | 'fy-DE' | 'fy-NL' | 'ga' | 'ga-IE' | 'gd-GB' | 'gez-ER' | 'gez-ET' | 'gl' | 'gl-ES' | 'gsw' | 'gsw-CH' | 'gu' | 'gu-IN' | 'guz' | 'guz-KE' | 'gv' | 'gv-GB' | 'ha' | 'ha-Latn' | 'ha-Latn-GH' | 'ha-Latn-NE' | 'ha-Latn-NG' | 'ha-NG' | 'haw' | 'haw-US' | 'he' | 'he-IL' | 'hi' | 'hi-IN' | 'hne-IN' | 'hr' | 'hr-HR' | 'hsb-DE' | 'ht-HT' | 'hu' | 'hu-HU' | 'hy' | 'hy-AM' | 'id' | 'id-ID' | 'ig' | 'ig-NG' | 'ii' | 'ii-CN' | 'ik-CA' | 'is' | 'is-IS' | 'it' | 'it-CH' | 'it-IT' | 'iu-CA' | 'iw-IL' | 'ja' | 'ja-JP' | 'jmc' | 'jmc-TZ' | 'ka' | 'ka-GE' | 'kab' | 'kab-DZ' | 'kam' | 'kam-KE' | 'kde' | 'kde-TZ' | 'kea' | 'kea-CV' | 'khq' | 'khq-ML' | 'ki' | 'ki-KE' | 'kk' | 'kk-Cyrl' | 'kk-Cyrl-KZ' | 'kk-KZ' | 'kl' | 'kl-GL' | 'kln' | 'kln-KE' | 'km' | 'km-KH' | 'kn' | 'kn-IN' | 'ko' | 'ko-KR' | 'kok' | 'kok-IN' | 'ks-IN' | 'ku-TR' | 'kw' | 'kw-GB' | 'ky-KG' | 'lag' | 'lag-TZ' | 'lb-LU' | 'lg' | 'lg-UG' | 'li-BE' | 'li-NL' | 'lij-IT' | 'lo-LA' | 'lt' | 'lt-LT' | 'luo' | 'luo-KE' | 'luy' | 'luy-KE' | 'lv' | 'lv-LV' | 'mag-IN' | 'mai-IN' | 'mas' | 'mas-KE' | 'mas-TZ' | 'mer' | 'mer-KE' | 'mfe' | 'mfe-MU' | 'mg' | 'mg-MG' | 'mhr-RU' | 'mi-NZ' | 'mk' | 'mk-MK' | 'ml' | 'ml-IN' | 'mn-MN' | 'mr' | 'mr-IN' | 'ms' | 'ms-BN' | 'ms-MY' | 'mt' | 'mt-MT' | 'my' | 'my-MM' | 'nan-TW' | 'naq' | 'naq-NA' | 'nb' | 'nb-NO' | 'nd' | 'nd-ZW' | 'nds-DE' | 'nds-NL' | 'ne' | 'ne-IN' | 'ne-NP' | 'nl' | 'nl-AW' | 'nl-BE' | 'nl-NL' | 'nn' | 'nn-NO' | 'nr-ZA' | 'nso-ZA' | 'nyn' | 'nyn-UG' | 'oc-FR' | 'om' | 'om-ET' | 'om-KE' | 'or' | 'or-IN' | 'os-RU' | 'pa' | 'pa-Arab' | 'pa-Arab-PK' | 'pa-Guru' | 'pa-Guru-IN' | 'pa-IN' | 'pa-PK' | 'pap-AN' | 'pl' | 'pl-PL' | 'ps' | 'ps-AF' | 'pt' | 'pt-BR' | 'pt-GW' | 'pt-MZ' | 'pt-PT' | 'rm' | 'rm-CH' | 'ro' | 'ro-MD' | 'ro-RO' | 'rof' | 'rof-TZ' | 'ru' | 'ru-MD' | 'ru-RU' | 'ru-UA' | 'rw' | 'rw-RW' | 'rwk' | 'rwk-TZ' | 'sa-IN' | 'saq' | 'saq-KE' | 'sc-IT' | 'sd-IN' | 'se-NO' | 'seh' | 'seh-MZ' | 'ses' | 'ses-ML' | 'sg' | 'sg-CF' | 'shi' | 'shi-Latn' | 'shi-Latn-MA' | 'shi-Tfng' | 'shi-Tfng-MA' | 'shs-CA' | 'si' | 'si-LK' | 'sid-ET' | 'sk' | 'sk-SK' | 'sl' | 'sl-SI' | 'sn' | 'sn-ZW' | 'so' | 'so-DJ' | 'so-ET' | 'so-KE' | 'so-SO' | 'sq' | 'sq-AL' | 'sq-MK' | 'sr' | 'sr-Cyrl' | 'sr-Cyrl-BA' | 'sr-Cyrl-ME' | 'sr-Cyrl-RS' | 'sr-Latn' | 'sr-Latn-BA' | 'sr-Latn-ME' | 'sr-Latn-RS' | 'sr-ME' | 'sr-RS' | 'ss-ZA' | 'st-ZA' | 'sv' | 'sv-FI' | 'sv-SE' | 'sw' | 'sw-KE' | 'sw-TZ' | 'ta' | 'ta-IN' | 'ta-LK' | 'te' | 'te-IN' | 'teo' | 'teo-KE' | 'teo-UG' | 'tg-TJ' | 'th' | 'th-TH' | 'ti' | 'ti-ER' | 'ti-ET' | 'tig-ER' | 'tk-TM' | 'tl-PH' | 'tn-ZA' | 'to' | 'to-TO' | 'tr' | 'tr-CY' | 'tr-TR' | 'ts-ZA' | 'tt-RU' | 'tzm' | 'tzm-Latn' | 'tzm-Latn-MA' | 'ug-CN' | 'uk' | 'uk-UA' | 'unm-US' | 'ur' | 'ur-IN' | 'ur-PK' | 'uz' | 'uz-Arab' | 'uz-Arab-AF' | 'uz-Cyrl' | 'uz-Cyrl-UZ' | 'uz-Latn' | 'uz-Latn-UZ' | 'uz-UZ' | 've-ZA' | 'vi' | 'vi-VN' | 'vun' | 'vun-TZ' | 'wa-BE' | 'wae-CH' | 'wal-ET' | 'wo-SN' | 'xh-ZA' | 'xog' | 'xog-UG' | 'yi-US' | 'yo' | 'yo-NG' | 'yue-HK' | 'zh' | 'zh-CN' | 'zh-HK' | 'zh-Hans' | 'zh-Hans-CN' | 'zh-Hans-HK' | 'zh-Hans-MO' | 'zh-Hans-SG' | 'zh-Hant' | 'zh-Hant-HK' | 'zh-Hant-MO' | 'zh-Hant-TW' | 'zh-SG' | 'zh-TW' | 'zu' | 'zu-ZA' | 'auto';
|
|
931
|
+
/**
|
|
932
|
+
* Selects which markdown conversion strategy to use. "full_page" converts the
|
|
933
|
+
* entire HTML page. "main_content" uses Mozilla Readability to extract the main
|
|
934
|
+
* article content before converting.
|
|
935
|
+
*/
|
|
936
|
+
markdown_backend?: 'full_page' | 'main_content';
|
|
937
|
+
/**
|
|
938
|
+
* HTTP method for the request
|
|
939
|
+
*/
|
|
940
|
+
method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
|
|
941
|
+
/**
|
|
942
|
+
* Filters for capturing network traffic
|
|
943
|
+
*/
|
|
944
|
+
network_capture?: Array<SharedInputs.NetworkCapture>;
|
|
945
|
+
/**
|
|
946
|
+
* Operating system to emulate
|
|
947
|
+
*/
|
|
948
|
+
os?: 'windows' | 'mac os' | 'linux' | 'android' | 'ios';
|
|
949
|
+
/**
|
|
950
|
+
* Whether to parse the response content
|
|
951
|
+
*/
|
|
952
|
+
parse?: boolean;
|
|
953
|
+
/**
|
|
954
|
+
* Custom parser configuration as a key-value map
|
|
955
|
+
*/
|
|
956
|
+
parser?: {
|
|
957
|
+
[key: string]: unknown;
|
|
958
|
+
} | string;
|
|
959
|
+
/**
|
|
960
|
+
* Referrer policy for the request
|
|
961
|
+
*/
|
|
962
|
+
referrer_type?: 'random' | 'no-referer' | 'same-origin' | 'google' | 'bing' | 'facebook' | 'twitter' | 'instagram';
|
|
963
|
+
/**
|
|
964
|
+
* Whether to render JavaScript content using a browser. Use 'auto' to let the
|
|
965
|
+
* engine select the candidate config per domain.
|
|
966
|
+
*/
|
|
967
|
+
render?: boolean | 'auto';
|
|
968
|
+
/**
|
|
969
|
+
* Request timeout in milliseconds
|
|
970
|
+
*/
|
|
971
|
+
request_timeout?: number;
|
|
972
|
+
session?: SharedInputs.Session;
|
|
973
|
+
/**
|
|
974
|
+
* Skills or capabilities required for the request
|
|
975
|
+
*/
|
|
976
|
+
skill?: string | Array<string>;
|
|
977
|
+
/**
|
|
978
|
+
* US state for geolocation (only valid when country is US)
|
|
979
|
+
*/
|
|
980
|
+
state?: 'AL' | 'AK' | 'AS' | 'AZ' | 'AR' | 'CA' | 'CO' | 'CT' | 'DE' | 'DC' | 'FL' | 'GA' | 'GU' | 'HI' | 'ID' | 'IL' | 'IN' | 'IA' | 'KS' | 'KY' | 'LA' | 'ME' | 'MD' | 'MA' | 'MI' | 'MN' | 'MS' | 'MO' | 'MT' | 'NE' | 'NV' | 'NH' | 'NJ' | 'NM' | 'NY' | 'NC' | 'ND' | 'MP' | 'OH' | 'OK' | 'OR' | 'PA' | 'PR' | 'RI' | 'SC' | 'SD' | 'TN' | 'TX' | 'UT' | 'VT' | 'VA' | 'VI' | 'WA' | 'WV' | 'WI' | 'WY';
|
|
981
|
+
/**
|
|
982
|
+
* Whether to compress stored data
|
|
983
|
+
*/
|
|
984
|
+
storage_compress?: boolean;
|
|
985
|
+
/**
|
|
986
|
+
* Custom name for the stored object
|
|
987
|
+
*/
|
|
988
|
+
storage_object_name?: string;
|
|
989
|
+
/**
|
|
990
|
+
* Type of storage to use for results
|
|
991
|
+
*/
|
|
992
|
+
storage_type?: string;
|
|
993
|
+
/**
|
|
994
|
+
* URL for storage location
|
|
995
|
+
*/
|
|
996
|
+
storage_url?: string;
|
|
997
|
+
/**
|
|
998
|
+
* User-defined tag for request identification
|
|
999
|
+
*/
|
|
1000
|
+
tag?: string;
|
|
1001
|
+
/**
|
|
1002
|
+
* Target URL to scrape
|
|
1003
|
+
*/
|
|
1004
|
+
url?: string;
|
|
1005
|
+
}
|
|
1006
|
+
namespace SharedInputs {
|
|
1007
|
+
interface UnionMember1 {
|
|
1008
|
+
name: 'chrome' | 'firefox';
|
|
1009
|
+
/**
|
|
1010
|
+
* Specific browser version to emulate
|
|
1011
|
+
*/
|
|
1012
|
+
version?: string;
|
|
1013
|
+
}
|
|
1014
|
+
interface UnionMember0 {
|
|
1015
|
+
creation?: string | null;
|
|
1016
|
+
domain?: string | null;
|
|
1017
|
+
expires?: string;
|
|
1018
|
+
extensions?: Array<string> | null;
|
|
1019
|
+
hostOnly?: boolean | null;
|
|
1020
|
+
httpOnly?: boolean | null;
|
|
1021
|
+
lastAccessed?: string | null;
|
|
1022
|
+
maxAge?: 'Infinity' | '-Infinity' | number | null;
|
|
1023
|
+
name?: string;
|
|
1024
|
+
path?: string | null;
|
|
1025
|
+
pathIsDefault?: boolean | null;
|
|
1026
|
+
sameSite?: 'strict' | 'lax' | 'none';
|
|
1027
|
+
secure?: boolean;
|
|
1028
|
+
value?: string;
|
|
1029
|
+
[k: string]: unknown;
|
|
1030
|
+
}
|
|
1031
|
+
interface NetworkCapture {
|
|
1032
|
+
method?: 'GET' | 'HEAD' | 'POST' | 'PUT' | 'DELETE' | 'CONNECT' | 'OPTIONS' | 'TRACE' | 'PATCH';
|
|
1033
|
+
/**
|
|
1034
|
+
* Resource type for network capture filtering
|
|
1035
|
+
*/
|
|
1036
|
+
resource_type?: string | Array<string>;
|
|
1037
|
+
status_code?: number | Array<number>;
|
|
1038
|
+
url?: NetworkCapture.URL;
|
|
1039
|
+
validation?: boolean;
|
|
1040
|
+
wait_for_requests_count?: number;
|
|
1041
|
+
wait_for_requests_count_timeout?: number;
|
|
1042
|
+
}
|
|
1043
|
+
namespace NetworkCapture {
|
|
1044
|
+
interface URL {
|
|
1045
|
+
value: string;
|
|
1046
|
+
type?: 'exact' | 'contains';
|
|
1047
|
+
}
|
|
1048
|
+
}
|
|
1049
|
+
interface Session {
|
|
1050
|
+
id?: string;
|
|
1051
|
+
prefetch_userbrowser?: boolean;
|
|
1052
|
+
renew_on_blocked?: boolean;
|
|
1053
|
+
retry?: boolean;
|
|
1054
|
+
timeout?: number;
|
|
1055
|
+
}
|
|
1056
|
+
}
|
|
1057
|
+
}
|
|
1058
|
+
export interface ExtractRunParams {
|
|
1059
|
+
/**
|
|
1060
|
+
* Target URL to scrape
|
|
1061
|
+
*/
|
|
1062
|
+
url: string;
|
|
1063
|
+
/**
|
|
1064
|
+
* Custom flow for the optimization engine: maps candidate names to the number of
|
|
1065
|
+
* attempts to spend on each candidate before advancing (0 skips it). Key order
|
|
1066
|
+
* defines the flow order. Providing it opts the request into 'auto' driver
|
|
1067
|
+
* selection.
|
|
1068
|
+
*/
|
|
1069
|
+
auto_driver_configuration?: {
|
|
1070
|
+
[key: string]: number;
|
|
1071
|
+
};
|
|
1072
|
+
/**
|
|
1073
|
+
* Request body for POST, PUT, PATCH methods
|
|
1074
|
+
*/
|
|
1075
|
+
body?: unknown;
|
|
1076
|
+
/**
|
|
1077
|
+
* Browser type to emulate
|
|
1078
|
+
*/
|
|
1079
|
+
browser?: 'chrome' | 'firefox' | ExtractRunParams.UnionMember1;
|
|
1080
|
+
/**
|
|
1081
|
+
* Array of browser automation actions to execute sequentially
|
|
1082
|
+
*/
|
|
1083
|
+
browser_actions?: Array<Shared.AutoScrollAction | Shared.ClickAction | Shared.EvalAction | Shared.FetchAction | Shared.FillAction | Shared.GetCookiesAction | Shared.GotoAction | Shared.PressAction | Shared.ScreenshotAction | Shared.ScrollAction | Shared.WaitAction | Shared.WaitForElementAction | Shared.WaitForNavigationAction>;
|
|
1084
|
+
/**
|
|
1085
|
+
* City for geolocation
|
|
1086
|
+
*/
|
|
1087
|
+
city?: string;
|
|
1088
|
+
/**
|
|
1089
|
+
* Whether to automatically handle cookie consent headers
|
|
1090
|
+
*/
|
|
1091
|
+
consent_header?: boolean;
|
|
1092
|
+
/**
|
|
1093
|
+
* Browser cookies as array of cookie objects
|
|
1094
|
+
*/
|
|
1095
|
+
cookies?: Array<ExtractRunParams.UnionMember0> | string;
|
|
1096
|
+
/**
|
|
1097
|
+
* Country code for geolocation and proxy selection
|
|
1098
|
+
*/
|
|
1099
|
+
country?: 'AD' | 'AE' | 'AF' | 'AG' | 'AI' | 'AL' | 'AM' | 'AO' | 'AQ' | 'AR' | 'AS' | 'AT' | 'AU' | 'AW' | 'AX' | 'AZ' | 'BA' | 'BB' | 'BD' | 'BE' | 'BF' | 'BG' | 'BH' | 'BI' | 'BJ' | 'BL' | 'BM' | 'BN' | 'BO' | 'BQ' | 'BR' | 'BS' | 'BT' | 'BV' | 'BW' | 'BY' | 'BZ' | 'CA' | 'CC' | 'CD' | 'CF' | 'CG' | 'CH' | 'CI' | 'CK' | 'CL' | 'CM' | 'CN' | 'CO' | 'CR' | 'CU' | 'CV' | 'CW' | 'CX' | 'CY' | 'CZ' | 'DE' | 'DJ' | 'DK' | 'DM' | 'DO' | 'DZ' | 'EC' | 'EE' | 'EG' | 'EH' | 'ER' | 'ES' | 'ET' | 'FI' | 'FJ' | 'FK' | 'FM' | 'FO' | 'FR' | 'GA' | 'GB' | 'GD' | 'GE' | 'GF' | 'GG' | 'GH' | 'GI' | 'GL' | 'GM' | 'GN' | 'GP' | 'GQ' | 'GR' | 'GS' | 'GT' | 'GU' | 'GW' | 'GY' | 'HK' | 'HM' | 'HN' | 'HR' | 'HT' | 'HU' | 'ID' | 'IE' | 'IL' | 'IM' | 'IN' | 'IO' | 'IQ' | 'IR' | 'IS' | 'IT' | 'JE' | 'JM' | 'JO' | 'JP' | 'KE' | 'KG' | 'KH' | 'KI' | 'KM' | 'KN' | 'KP' | 'KR' | 'KW' | 'KY' | 'KZ' | 'LA' | 'LB' | 'LC' | 'LI' | 'LK' | 'LR' | 'LS' | 'LT' | 'LU' | 'LV' | 'LY' | 'MA' | 'MC' | 'MD' | 'ME' | 'MF' | 'MG' | 'MH' | 'MK' | 'ML' | 'MM' | 'MN' | 'MO' | 'MP' | 'MQ' | 'MR' | 'MS' | 'MT' | 'MU' | 'MV' | 'MW' | 'MX' | 'MY' | 'MZ' | 'NA' | 'NC' | 'NE' | 'NF' | 'NG' | 'NI' | 'NL' | 'NO' | 'NP' | 'NR' | 'NU' | 'NZ' | 'OM' | 'PA' | 'PE' | 'PF' | 'PG' | 'PH' | 'PK' | 'PL' | 'PM' | 'PN' | 'PR' | 'PS' | 'PT' | 'PW' | 'PY' | 'QA' | 'RE' | 'RO' | 'RS' | 'RU' | 'RW' | 'SA' | 'SB' | 'SC' | 'SD' | 'SE' | 'SG' | 'SH' | 'SI' | 'SJ' | 'SK' | 'SL' | 'SM' | 'SN' | 'SO' | 'SR' | 'SS' | 'ST' | 'SV' | 'SX' | 'SY' | 'SZ' | 'TC' | 'TD' | 'TF' | 'TG' | 'TH' | 'TJ' | 'TK' | 'TL' | 'TM' | 'TN' | 'TO' | 'TR' | 'TT' | 'TV' | 'TW' | 'TZ' | 'UA' | 'UG' | 'UM' | 'US' | 'UY' | 'UZ' | 'VA' | 'VC' | 'VE' | 'VG' | 'VI' | 'VN' | 'VU' | 'WF' | 'WS' | 'XK' | 'YE' | 'YT' | 'ZA' | 'ZM' | 'ZW' | 'ALL';
|
|
1100
|
+
/**
|
|
1101
|
+
* Device type for browser emulation
|
|
1102
|
+
*/
|
|
1103
|
+
device?: 'desktop' | 'mobile' | 'tablet';
|
|
1104
|
+
/**
|
|
1105
|
+
* Browser driver to use
|
|
1106
|
+
*/
|
|
1107
|
+
driver?: 'auto' | 'vx6' | 'vx8' | 'vx8-pro' | 'vx10' | 'vx10-pro' | 'vx12' | 'vx12-pro' | 'media-vx6' | 'fast-vx6';
|
|
1108
|
+
/**
|
|
1109
|
+
* Expected HTTP status codes for successful requests
|
|
1110
|
+
*/
|
|
1111
|
+
expected_status_codes?: Array<number>;
|
|
1112
|
+
/**
|
|
1113
|
+
* List of acceptable response formats in order of preference
|
|
1114
|
+
*/
|
|
1115
|
+
formats?: Array<'html' | 'markdown' | 'screenshot' | 'headers' | 'links'>;
|
|
1116
|
+
/**
|
|
1117
|
+
* Custom HTTP headers to include in the request
|
|
1118
|
+
*/
|
|
1119
|
+
headers?: {
|
|
1120
|
+
[key: string]: string | Array<string> | null;
|
|
1121
|
+
};
|
|
1122
|
+
/**
|
|
1123
|
+
* Whether to use HTTP/2 protocol
|
|
1124
|
+
*/
|
|
1125
|
+
http2?: boolean;
|
|
1126
|
+
/**
|
|
1127
|
+
* Whether to emulate XMLHttpRequest behavior
|
|
1128
|
+
*/
|
|
1129
|
+
is_xhr?: boolean;
|
|
1130
|
+
/**
|
|
1131
|
+
* Locale for browser language and region settings
|
|
1132
|
+
*/
|
|
1133
|
+
locale?: 'aa-DJ' | 'aa-ER' | 'aa-ET' | 'af' | 'af-NA' | 'af-ZA' | 'ak' | 'ak-GH' | 'am' | 'am-ET' | 'an-ES' | 'ar' | 'ar-AE' | 'ar-BH' | 'ar-DZ' | 'ar-EG' | 'ar-IN' | 'ar-IQ' | 'ar-JO' | 'ar-KW' | 'ar-LB' | 'ar-LY' | 'ar-MA' | 'ar-OM' | 'ar-QA' | 'ar-SA' | 'ar-SD' | 'ar-SY' | 'ar-TN' | 'ar-YE' | 'as' | 'as-IN' | 'asa' | 'asa-TZ' | 'ast-ES' | 'az' | 'az-AZ' | 'az-Cyrl' | 'az-Cyrl-AZ' | 'az-Latn' | 'az-Latn-AZ' | 'be' | 'be-BY' | 'bem' | 'bem-ZM' | 'ber-DZ' | 'ber-MA' | 'bez' | 'bez-TZ' | 'bg' | 'bg-BG' | 'bho-IN' | 'bm' | 'bm-ML' | 'bn' | 'bn-BD' | 'bn-IN' | 'bo' | 'bo-CN' | 'bo-IN' | 'br-FR' | 'brx-IN' | 'bs' | 'bs-BA' | 'byn-ER' | 'ca' | 'ca-AD' | 'ca-ES' | 'ca-FR' | 'ca-IT' | 'cgg' | 'cgg-UG' | 'chr' | 'chr-US' | 'crh-UA' | 'cs' | 'cs-CZ' | 'csb-PL' | 'cv-RU' | 'cy' | 'cy-GB' | 'da' | 'da-DK' | 'dav' | 'dav-KE' | 'de' | 'de-AT' | 'de-BE' | 'de-CH' | 'de-DE' | 'de-LI' | 'de-LU' | 'dv-MV' | 'dz-BT' | 'ebu' | 'ebu-KE' | 'ee' | 'ee-GH' | 'ee-TG' | 'el' | 'el-CY' | 'el-GR' | 'en' | 'en-AG' | 'en-AS' | 'en-AU' | 'en-BE' | 'en-BW' | 'en-BZ' | 'en-CA' | 'en-DK' | 'en-GB' | 'en-GU' | 'en-HK' | 'en-IE' | 'en-IN' | 'en-JM' | 'en-MH' | 'en-MP' | 'en-MT' | 'en-MU' | 'en-NA' | 'en-NG' | 'en-NZ' | 'en-PH' | 'en-PK' | 'en-SG' | 'en-TT' | 'en-UM' | 'en-US' | 'en-VI' | 'en-ZA' | 'en-ZM' | 'en-ZW' | 'eo' | 'es' | 'es-419' | 'es-AR' | 'es-BO' | 'es-CL' | 'es-CO' | 'es-CR' | 'es-CU' | 'es-DO' | 'es-EC' | 'es-ES' | 'es-GQ' | 'es-GT' | 'es-HN' | 'es-MX' | 'es-NI' | 'es-PA' | 'es-PE' | 'es-PR' | 'es-PY' | 'es-SV' | 'es-US' | 'es-UY' | 'es-VE' | 'et' | 'et-EE' | 'eu' | 'eu-ES' | 'fa' | 'fa-AF' | 'fa-IR' | 'ff' | 'ff-SN' | 'fi' | 'fi-FI' | 'fil' | 'fil-PH' | 'fo' | 'fo-FO' | 'fr' | 'fr-BE' | 'fr-BF' | 'fr-BI' | 'fr-BJ' | 'fr-BL' | 'fr-CA' | 'fr-CD' | 'fr-CF' | 'fr-CG' | 'fr-CH' | 'fr-CI' | 'fr-CM' | 'fr-DJ' | 'fr-FR' | 'fr-GA' | 'fr-GN' | 'fr-GP' | 'fr-GQ' | 'fr-KM' | 'fr-LU' | 'fr-MC' | 'fr-MF' | 'fr-MG' | 'fr-ML' | 'fr-MQ' | 'fr-NE' | 'fr-RE' | 'fr-RW' | 'fr-SN' | 'fr-TD' | 'fr-TG' | 'fur-IT' | 'fy-DE' | 'fy-NL' | 'ga' | 'ga-IE' | 'gd-GB' | 'gez-ER' | 'gez-ET' | 'gl' | 'gl-ES' | 'gsw' | 'gsw-CH' | 'gu' | 'gu-IN' | 'guz' | 'guz-KE' | 'gv' | 'gv-GB' | 'ha' | 'ha-Latn' | 'ha-Latn-GH' | 'ha-Latn-NE' | 'ha-Latn-NG' | 'ha-NG' | 'haw' | 'haw-US' | 'he' | 'he-IL' | 'hi' | 'hi-IN' | 'hne-IN' | 'hr' | 'hr-HR' | 'hsb-DE' | 'ht-HT' | 'hu' | 'hu-HU' | 'hy' | 'hy-AM' | 'id' | 'id-ID' | 'ig' | 'ig-NG' | 'ii' | 'ii-CN' | 'ik-CA' | 'is' | 'is-IS' | 'it' | 'it-CH' | 'it-IT' | 'iu-CA' | 'iw-IL' | 'ja' | 'ja-JP' | 'jmc' | 'jmc-TZ' | 'ka' | 'ka-GE' | 'kab' | 'kab-DZ' | 'kam' | 'kam-KE' | 'kde' | 'kde-TZ' | 'kea' | 'kea-CV' | 'khq' | 'khq-ML' | 'ki' | 'ki-KE' | 'kk' | 'kk-Cyrl' | 'kk-Cyrl-KZ' | 'kk-KZ' | 'kl' | 'kl-GL' | 'kln' | 'kln-KE' | 'km' | 'km-KH' | 'kn' | 'kn-IN' | 'ko' | 'ko-KR' | 'kok' | 'kok-IN' | 'ks-IN' | 'ku-TR' | 'kw' | 'kw-GB' | 'ky-KG' | 'lag' | 'lag-TZ' | 'lb-LU' | 'lg' | 'lg-UG' | 'li-BE' | 'li-NL' | 'lij-IT' | 'lo-LA' | 'lt' | 'lt-LT' | 'luo' | 'luo-KE' | 'luy' | 'luy-KE' | 'lv' | 'lv-LV' | 'mag-IN' | 'mai-IN' | 'mas' | 'mas-KE' | 'mas-TZ' | 'mer' | 'mer-KE' | 'mfe' | 'mfe-MU' | 'mg' | 'mg-MG' | 'mhr-RU' | 'mi-NZ' | 'mk' | 'mk-MK' | 'ml' | 'ml-IN' | 'mn-MN' | 'mr' | 'mr-IN' | 'ms' | 'ms-BN' | 'ms-MY' | 'mt' | 'mt-MT' | 'my' | 'my-MM' | 'nan-TW' | 'naq' | 'naq-NA' | 'nb' | 'nb-NO' | 'nd' | 'nd-ZW' | 'nds-DE' | 'nds-NL' | 'ne' | 'ne-IN' | 'ne-NP' | 'nl' | 'nl-AW' | 'nl-BE' | 'nl-NL' | 'nn' | 'nn-NO' | 'nr-ZA' | 'nso-ZA' | 'nyn' | 'nyn-UG' | 'oc-FR' | 'om' | 'om-ET' | 'om-KE' | 'or' | 'or-IN' | 'os-RU' | 'pa' | 'pa-Arab' | 'pa-Arab-PK' | 'pa-Guru' | 'pa-Guru-IN' | 'pa-IN' | 'pa-PK' | 'pap-AN' | 'pl' | 'pl-PL' | 'ps' | 'ps-AF' | 'pt' | 'pt-BR' | 'pt-GW' | 'pt-MZ' | 'pt-PT' | 'rm' | 'rm-CH' | 'ro' | 'ro-MD' | 'ro-RO' | 'rof' | 'rof-TZ' | 'ru' | 'ru-MD' | 'ru-RU' | 'ru-UA' | 'rw' | 'rw-RW' | 'rwk' | 'rwk-TZ' | 'sa-IN' | 'saq' | 'saq-KE' | 'sc-IT' | 'sd-IN' | 'se-NO' | 'seh' | 'seh-MZ' | 'ses' | 'ses-ML' | 'sg' | 'sg-CF' | 'shi' | 'shi-Latn' | 'shi-Latn-MA' | 'shi-Tfng' | 'shi-Tfng-MA' | 'shs-CA' | 'si' | 'si-LK' | 'sid-ET' | 'sk' | 'sk-SK' | 'sl' | 'sl-SI' | 'sn' | 'sn-ZW' | 'so' | 'so-DJ' | 'so-ET' | 'so-KE' | 'so-SO' | 'sq' | 'sq-AL' | 'sq-MK' | 'sr' | 'sr-Cyrl' | 'sr-Cyrl-BA' | 'sr-Cyrl-ME' | 'sr-Cyrl-RS' | 'sr-Latn' | 'sr-Latn-BA' | 'sr-Latn-ME' | 'sr-Latn-RS' | 'sr-ME' | 'sr-RS' | 'ss-ZA' | 'st-ZA' | 'sv' | 'sv-FI' | 'sv-SE' | 'sw' | 'sw-KE' | 'sw-TZ' | 'ta' | 'ta-IN' | 'ta-LK' | 'te' | 'te-IN' | 'teo' | 'teo-KE' | 'teo-UG' | 'tg-TJ' | 'th' | 'th-TH' | 'ti' | 'ti-ER' | 'ti-ET' | 'tig-ER' | 'tk-TM' | 'tl-PH' | 'tn-ZA' | 'to' | 'to-TO' | 'tr' | 'tr-CY' | 'tr-TR' | 'ts-ZA' | 'tt-RU' | 'tzm' | 'tzm-Latn' | 'tzm-Latn-MA' | 'ug-CN' | 'uk' | 'uk-UA' | 'unm-US' | 'ur' | 'ur-IN' | 'ur-PK' | 'uz' | 'uz-Arab' | 'uz-Arab-AF' | 'uz-Cyrl' | 'uz-Cyrl-UZ' | 'uz-Latn' | 'uz-Latn-UZ' | 'uz-UZ' | 've-ZA' | 'vi' | 'vi-VN' | 'vun' | 'vun-TZ' | 'wa-BE' | 'wae-CH' | 'wal-ET' | 'wo-SN' | 'xh-ZA' | 'xog' | 'xog-UG' | 'yi-US' | 'yo' | 'yo-NG' | 'yue-HK' | 'zh' | 'zh-CN' | 'zh-HK' | 'zh-Hans' | 'zh-Hans-CN' | 'zh-Hans-HK' | 'zh-Hans-MO' | 'zh-Hans-SG' | 'zh-Hant' | 'zh-Hant-HK' | 'zh-Hant-MO' | 'zh-Hant-TW' | 'zh-SG' | 'zh-TW' | 'zu' | 'zu-ZA' | 'auto';
|
|
1134
|
+
/**
|
|
1135
|
+
* Selects which markdown conversion strategy to use. "full_page" converts the
|
|
1136
|
+
* entire HTML page. "main_content" uses Mozilla Readability to extract the main
|
|
1137
|
+
* article content before converting.
|
|
1138
|
+
*/
|
|
1139
|
+
markdown_backend?: 'full_page' | 'main_content';
|
|
1140
|
+
/**
|
|
1141
|
+
* HTTP method for the request
|
|
1142
|
+
*/
|
|
1143
|
+
method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
|
|
1144
|
+
/**
|
|
1145
|
+
* Filters for capturing network traffic
|
|
1146
|
+
*/
|
|
1147
|
+
network_capture?: Array<ExtractRunParams.NetworkCapture>;
|
|
1148
|
+
/**
|
|
1149
|
+
* Operating system to emulate
|
|
1150
|
+
*/
|
|
1151
|
+
os?: 'windows' | 'mac os' | 'linux' | 'android' | 'ios';
|
|
1152
|
+
/**
|
|
1153
|
+
* Whether to parse the response content
|
|
1154
|
+
*/
|
|
1155
|
+
parse?: boolean;
|
|
1156
|
+
/**
|
|
1157
|
+
* Custom parser configuration as a key-value map
|
|
1158
|
+
*/
|
|
1159
|
+
parser?: {
|
|
1160
|
+
[key: string]: unknown;
|
|
1161
|
+
} | string;
|
|
1162
|
+
/**
|
|
1163
|
+
* Referrer policy for the request
|
|
1164
|
+
*/
|
|
1165
|
+
referrer_type?: 'random' | 'no-referer' | 'same-origin' | 'google' | 'bing' | 'facebook' | 'twitter' | 'instagram';
|
|
1166
|
+
/**
|
|
1167
|
+
* Whether to render JavaScript content using a browser
|
|
1168
|
+
*/
|
|
1169
|
+
render?: boolean | 'auto';
|
|
1170
|
+
/**
|
|
1171
|
+
* Request timeout in milliseconds
|
|
1172
|
+
*/
|
|
1173
|
+
request_timeout?: number;
|
|
1174
|
+
session?: ExtractRunParams.Session;
|
|
1175
|
+
/**
|
|
1176
|
+
* Skills or capabilities required for the request
|
|
1177
|
+
*/
|
|
1178
|
+
skill?: string | Array<string>;
|
|
1179
|
+
/**
|
|
1180
|
+
* US state for geolocation (only valid when country is US)
|
|
1181
|
+
*/
|
|
1182
|
+
state?: 'AL' | 'AK' | 'AS' | 'AZ' | 'AR' | 'CA' | 'CO' | 'CT' | 'DE' | 'DC' | 'FL' | 'GA' | 'GU' | 'HI' | 'ID' | 'IL' | 'IN' | 'IA' | 'KS' | 'KY' | 'LA' | 'ME' | 'MD' | 'MA' | 'MI' | 'MN' | 'MS' | 'MO' | 'MT' | 'NE' | 'NV' | 'NH' | 'NJ' | 'NM' | 'NY' | 'NC' | 'ND' | 'MP' | 'OH' | 'OK' | 'OR' | 'PA' | 'PR' | 'RI' | 'SC' | 'SD' | 'TN' | 'TX' | 'UT' | 'VT' | 'VA' | 'VI' | 'WA' | 'WV' | 'WI' | 'WY';
|
|
1183
|
+
/**
|
|
1184
|
+
* User-defined tag for request identification
|
|
1185
|
+
*/
|
|
1186
|
+
tag?: string;
|
|
1187
|
+
}
|
|
1188
|
+
export declare namespace ExtractRunParams {
|
|
1189
|
+
interface UnionMember1 {
|
|
1190
|
+
name: 'chrome' | 'firefox';
|
|
1191
|
+
/**
|
|
1192
|
+
* Specific browser version to emulate
|
|
1193
|
+
*/
|
|
1194
|
+
version?: string;
|
|
1195
|
+
}
|
|
1196
|
+
interface UnionMember0 {
|
|
1197
|
+
creation?: string | null;
|
|
1198
|
+
domain?: string | null;
|
|
1199
|
+
expires?: string;
|
|
1200
|
+
extensions?: Array<string> | null;
|
|
1201
|
+
hostOnly?: boolean | null;
|
|
1202
|
+
httpOnly?: boolean | null;
|
|
1203
|
+
lastAccessed?: string | null;
|
|
1204
|
+
maxAge?: 'Infinity' | '-Infinity' | number | null;
|
|
1205
|
+
name?: string;
|
|
1206
|
+
path?: string | null;
|
|
1207
|
+
pathIsDefault?: boolean | null;
|
|
1208
|
+
sameSite?: 'strict' | 'lax' | 'none';
|
|
1209
|
+
secure?: boolean;
|
|
1210
|
+
value?: string;
|
|
1211
|
+
[k: string]: unknown;
|
|
1212
|
+
}
|
|
1213
|
+
interface NetworkCapture {
|
|
1214
|
+
method?: 'GET' | 'HEAD' | 'POST' | 'PUT' | 'DELETE' | 'CONNECT' | 'OPTIONS' | 'TRACE' | 'PATCH';
|
|
1215
|
+
/**
|
|
1216
|
+
* Resource type for network capture filtering
|
|
1217
|
+
*/
|
|
1218
|
+
resource_type?: string | Array<string>;
|
|
1219
|
+
status_code?: number | Array<number>;
|
|
1220
|
+
url?: NetworkCapture.URL;
|
|
1221
|
+
validation?: boolean;
|
|
1222
|
+
wait_for_requests_count?: number;
|
|
1223
|
+
wait_for_requests_count_timeout?: number;
|
|
1224
|
+
}
|
|
1225
|
+
namespace NetworkCapture {
|
|
1226
|
+
interface URL {
|
|
1227
|
+
value: string;
|
|
1228
|
+
type?: 'exact' | 'contains';
|
|
1229
|
+
}
|
|
1230
|
+
}
|
|
1231
|
+
interface Session {
|
|
1232
|
+
id?: string;
|
|
1233
|
+
prefetch_userbrowser?: boolean;
|
|
1234
|
+
renew_on_blocked?: boolean;
|
|
1235
|
+
retry?: boolean;
|
|
1236
|
+
timeout?: number;
|
|
1237
|
+
}
|
|
1238
|
+
}
|
|
1239
|
+
export declare namespace Extract {
|
|
1240
|
+
export { type ExtractAsyncResponse as ExtractAsyncResponse, type ExtractBatchResponse as ExtractBatchResponse, type ExtractRunResponse as ExtractRunResponse, type ExtractAsyncParams as ExtractAsyncParams, type ExtractBatchParams as ExtractBatchParams, type ExtractRunParams as ExtractRunParams, };
|
|
1241
|
+
export { Templates as Templates, type TemplateUpdateResponse as TemplateUpdateResponse, type TemplateListResponse as TemplateListResponse, type TemplateAsyncResponse as TemplateAsyncResponse, type TemplateBatchResponse as TemplateBatchResponse, type TemplateGetResponse as TemplateGetResponse, type TemplateRunResponse as TemplateRunResponse, type TemplateUpdateParams as TemplateUpdateParams, type TemplateListParams as TemplateListParams, type TemplateAsyncParams as TemplateAsyncParams, type TemplateBatchParams as TemplateBatchParams, type TemplateRunParams as TemplateRunParams, };
|
|
1242
|
+
}
|
|
1243
|
+
//# sourceMappingURL=extract.d.mts.map
|