File size: 3,119 Bytes
936d43f
 
 
 
 
b4cb4da
936d43f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
from smolagents import Tool


class FlashDataTool(Tool):
    name = 'youtube_download'
    description = 'Download YouTube video, audio or clips to your own S3 or other supported cloud storage. Submit a job once, then read its status and delivery manifest. Requires your own writable storage and permission for your intended use. Downloads are asynchronous and charged by delivered data volume; storage charges are separate. Uses paid FlashData API credits. Never automatically retry an uncertain submission; check Jobs first.'
    inputs = {'job_id': {'description': 'Required for status/result. ID returned by a previous submit.',
                'nullable': True,
                'type': 'string'},
     'operation': {'description': 'submit starts one paid download; status or result reads an existing '
                                  'job without resubmitting.',
                   'type': 'string'},
     'parameters': {'description': 'Submit only: video_id and media options from input-schema.json. '
                                   'Never include storage credentials.',
                    'nullable': True,
                    'type': 'object'}}
    output_type = 'object'

    def forward(self, operation: str, parameters: dict | None = None, job_id: str | None = None) -> dict:
        import os
        api_key = os.environ.get('FLASHDATA_API_KEY', '')
        storage_json = os.environ.get('FLASHDATA_DOWNLOAD_STORAGE', '')

        import json
        from urllib.parse import urlsplit
        from langchain_flashdata import FlashDataYouTubeDownload, FlashDataDownloadJob, FlashDataDownloadResult

        if operation not in ('submit', 'status', 'result'):
            raise ValueError('operation must be submit, status or result.')
        if operation != 'submit':
            if parameters:
                raise ValueError('Status/result reads take job_id only; omit parameters.')
            reader = FlashDataDownloadJob if operation == 'status' else FlashDataDownloadResult
            return reader(api_key=api_key).invoke({'job_id': job_id})
        if job_id:
            raise ValueError('Submit does not accept job_id. To read an existing job, select status or result.')
        if not isinstance(parameters, dict):
            raise ValueError('Submit requires a parameters object with video_id and media options.')
        try:
            storage = json.loads(storage_json)
            if not isinstance(storage, dict):
                raise ValueError()
            endpoint = storage.get('endpoint_url')
            if endpoint:
                url = urlsplit(endpoint)
                if (url.scheme != 'https' or not url.hostname or url.username or url.password
                        or url.query or url.fragment or url.port not in (None, 443)):
                    raise ValueError()
        except (TypeError, ValueError):
            raise ValueError('Configure valid customer storage JSON; endpoints must be HTTPS without credentials, query or fragment.') from None
        return FlashDataYouTubeDownload(api_key=api_key, download_storage=storage_json).invoke(parameters)