python-bigquery-dataframes/third_party/bigframes_vendored/ibis/expr/operations/ai_ops.py at 12f020b402768be895d4559f96759fc3b2604e2b · googleapis/python-bigquery-dataframes · GitHub

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
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
# Contains code from https://github.com/ibis-project/ibis/blob/9.2.0/ibis/expr/operations/maps.py

"""Operations for working with AI operators."""

from __future__ import annotations

from typing import Optional

from bigframes_vendored.ibis.common.annotations import attribute
import bigframes_vendored.ibis.expr.datatypes as dt
from bigframes_vendored.ibis.expr.operations.core import Value
import bigframes_vendored.ibis.expr.rules as rlz
from public import public
import pyarrow as pa

from bigframes.operations import output_schemas


@public
class AIGenerate(Value):
    """Generate content based on the prompt"""

    prompt: Value
    connection_id: Optional[Value[dt.String]]
    endpoint: Optional[Value[dt.String]]
    request_type: Value[dt.String]
    model_params: Optional[Value[dt.String]]
    output_schema: Optional[Value[dt.String]]

    shape = rlz.shape_like("prompt")

    @attribute
    def dtype(self) -> dt.Struct:
        if self.output_schema is None:
            output_pa_fields = (pa.field("result", pa.string()),)
        else:
            output_pa_fields = output_schemas.parse_sql_fields(self.output_schema.value)

        pyarrow_output_type = pa.struct(
            (
                *output_pa_fields,
                pa.field("full_resposne", pa.string()),
                pa.field("status", pa.string()),
            )
        )

        return dt.Struct.from_pyarrow(pyarrow_output_type)


@public
class AIGenerateBool(Value):
    """Generate Bool based on the prompt"""

    prompt: Value
    connection_id: Optional[Value[dt.String]]
    endpoint: Optional[Value[dt.String]]
    request_type: Value[dt.String]
    model_params: Optional[Value[dt.String]]

    shape = rlz.shape_like("prompt")

    @attribute
    def dtype(self) -> dt.Struct:
        return dt.Struct.from_tuples(
            (("result", dt.bool), ("full_resposne", dt.string), ("status", dt.string))
        )


@public
class AIGenerateInt(Value):
    """Generate integers based on the prompt"""

    prompt: Value
    connection_id: Optional[Value[dt.String]]
    endpoint: Optional[Value[dt.String]]
    request_type: Value[dt.String]
    model_params: Optional[Value[dt.String]]

    shape = rlz.shape_like("prompt")

    @attribute
    def dtype(self) -> dt.Struct:
        return dt.Struct.from_tuples(
            (("result", dt.int64), ("full_resposne", dt.string), ("status", dt.string))
        )


@public
class AIGenerateDouble(Value):
    """Generate doubles based on the prompt"""

    prompt: Value
    connection_id: Optional[Value[dt.String]]
    endpoint: Optional[Value[dt.String]]
    request_type: Value[dt.String]
    model_params: Optional[Value[dt.String]]

    shape = rlz.shape_like("prompt")

    @attribute
    def dtype(self) -> dt.Struct:
        return dt.Struct.from_tuples(
            (
                ("result", dt.float64),
                ("full_resposne", dt.string),
                ("status", dt.string),
            )
        )


@public
class AIIf(Value):
    """Generate True/False based on the prompt"""

    prompt: Value
    connection_id: Optional[Value[dt.String]]

    shape = rlz.shape_like("prompt")

    @attribute
    def dtype(self) -> dt.Struct:
        return dt.bool


@public
class AIClassify(Value):
    """Generate True/False based on the prompt"""

    input: Value
    categories: Value[dt.Array[dt.String]]
    connection_id: Optional[Value[dt.String]]

    shape = rlz.shape_like("input")

    @attribute
    def dtype(self) -> dt.Struct:
        return dt.string


@public
class AIScore(Value):
    """Generate doubles based on the prompt"""

    prompt: Value
    connection_id: Optional[Value[dt.String]]

    shape = rlz.shape_like("prompt")

    @attribute
    def dtype(self) -> dt.Struct:
        return dt.float64