archastro.platform.types.extractions

 1# Copyright (c) 2026 ArchAstro Inc. Licensed under the MIT License.
 2# This file is auto-generated by @archastro/sdk-generator. Do not edit.
 3# Content hash: d8459d6fc18e
 4
 5from datetime import datetime
 6from typing import Any, Literal
 7
 8from pydantic import BaseModel, Field
 9
10from .common import StorageFile
11
12
13class ExtractionOutput(BaseModel):
14    """
15    A produced file tracked by an extraction. Type, size, and URL live on the file it points at.
16    """
17
18    created_at: datetime | None = Field(
19        default=None, description="When this output was produced (ISO 8601)."
20    )
21    file: StorageFile | None = Field(
22        default=None,
23        description="The produced file. For a config destination this is the `Storage.File` backing the versioned config row; type, size, and URL live here.",
24    )
25    id: str = Field(..., description="Output ID (`exo_...`).")
26    ordinal: int | None = Field(
27        default=None, description="Position of this output within the extraction's output set."
28    )
29    source_url: str | None = Field(
30        default=None, description="The crawled page path or document path this output came from."
31    )
32    state: Literal["done", "failed"] | None = Field(default=None, description="Output state.")
33
34
35class Extraction(BaseModel):
36    """
37    An extraction job: yields text from a document or website into a destination namespace, without committing knowledge to an agent.
38    """
39
40    agent: str | None = Field(
41        default=None, description="Owning agent (`agt_...`); `null` when not agent-scoped."
42    )
43    byte_count: int | None = Field(
44        default=None, description="Total bytes across produced storage files (a derived aggregate)."
45    )
46    created_at: datetime | None = Field(
47        default=None, description="When the extraction was created (ISO 8601)."
48    )
49    destination: dict[str, Any] | None = Field(
50        default=None, description="Where outputs were written: `{ kind, path_prefix }`."
51    )
52    failure_reason: (
53        Literal[
54            "fetch_failed", "unsupported_content", "extraction_failed", "timeout", "internal_error"
55        ]
56        | None
57    ) = Field(
58        default=None, description="Failure category when `state` is `failed`; `null` otherwise."
59    )
60    file: str | None = Field(
61        default=None,
62        description="Source file (`fil_...`) for document extraction; `null` for link/site.",
63    )
64    id: str = Field(..., description="Extraction ID (`ext_...`).")
65    kind: Literal["document", "link", "site"] | None = Field(
66        default=None, description="What is being extracted."
67    )
68    org: str | None = Field(default=None, description="Owning organization (`org_...`).")
69    output_count: int | None = Field(
70        default=None, description="Number of produced output files (a derived aggregate)."
71    )
72    outputs: list[ExtractionOutput] | None = Field(
73        default=None,
74        description="Produced output files; populated only when the association is preloaded.",
75    )
76    state: Literal["pending", "running", "done", "failed"] | None = Field(
77        default=None, description="Lifecycle state of the extraction job."
78    )
79    updated_at: datetime | None = Field(
80        default=None, description="When the extraction was last updated (ISO 8601)."
81    )
82    url: str | None = Field(
83        default=None, description="Source URL for link/site extraction; `null` for document."
84    )