ums.utils.schema

 1# Agenten Plattform
 2#
 3# (c) 2024 Magnus Bender
 4# 	Institute of Humanities-Centered Artificial Intelligence (CHAI)
 5# 	Universitaet Hamburg
 6# 	https://www.chai.uni-hamburg.de/~bender
 7#  
 8# source code released under the terms of GNU Public License Version 3
 9# https://www.gnu.org/licenses/gpl-3.0.txt
10
11"""
12	This represents the basic types used for representing extracted information from the data.
13	The types are implemented using [pydantic](https://docs.pydantic.dev/).
14	It provides validation, allow JSON serialization and works well with [FastAPI](https://fastapi.tiangolo.com/) which is used internally for the http request between the agents and the management.
15
16	**This is work in progress!**
17"""
18
19from typing import List, Any, Dict
20
21from pydantic import BaseModel
22
23class ExtractionSchema(BaseModel):
24	"""
25		This is the basic class used as superclass for all extracted information from data items.
26
27		For all the `ExtractionSchema` is is required that the data can be serialized to json. 
28		Thus, mostly only default data types like `int, str, bool, list, dict, tuple` also including `ExtractionSchema` and `RiddleInformation` can be used here!
29	"""
30
31class ExtractedContent(ExtractionSchema):
32	"""
33		An extracted content item.
34	"""
35
36	type : str 
37	"""
38		The type, as a string, the actual string will depend on the extraction agent.
39	"""
40
41	content : str | Any
42	"""
43		The extracted content
44	"""
45
46class ExtractedPositions(ExtractionSchema):
47	"""
48		A position (like time, coordinates, ...) where something was extracted (each position should belong to a content item).
49	"""
50
51	type : str 
52	"""
53		The type, as a string, the actual string will depend on the extraction agent.
54	"""
55
56	position : str | int | Any
57	"""
58		The position, will also depend on the extraction agent.
59	"""
60
61	description : str | Any = None
62	"""
63		An optional description for more details.
64	"""
65
66class ExtractedData(ExtractionSchema):
67	"""
68		Contains the extracted items from a data file.
69	"""
70
71	contents : List[ExtractedContent] = []
72	"""
73		The extracted contents (i.e., transcriptions etc.), each item here should belong a position item at the same index.
74	"""
75
76	positions : List[ExtractedPositions] = []
77	"""
78		The positions of extracted contents, each item here should belong a content item at the same index.
79	"""
80
81	other : Dict[str, Any] = {}
82	"""
83		Possibly more data. Use a keywords (depending on agent) and store the data there.
84	"""