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

use crate::internal::*;

#[derive(Debug, Default, Clone, new, Hash)]
pub struct Range;



impl Expansion for Range {
    fn name(&self) -> Cow<str> {
        "Range".into()
    }


    fn rules<'r, 'p: 'r, 's: 'r>(
        &'s self,
        s: &mut Solver<'r>,
        inputs: &'p [TensorProxy],
        outputs: &'p [TensorProxy],
    ) -> InferenceResult {
        check_input_arity(inputs, 3)?;
        check_output_arity(outputs, 1)?;
        s.given_3(
            &inputs[0].datum_type,
            &inputs[1].datum_type,
            &inputs[2].datum_type,
            move |s, dt0, dt1, dt2| {
                let dt =
                    DatumType::super_type_for([dt0, dt1, dt2]).context("No supertype found")?;
                s.equals(dt, &outputs[0].datum_type)
            },
        )?;
        s.equals(&inputs[0].rank, 0)?;
        s.equals(&inputs[1].rank, 0)?;
        s.equals(&inputs[2].rank, 0)?;
        s.equals(&outputs[0].rank, 1)?;
        s.given_3(&inputs[0].value, &inputs[1].value, &inputs[2].value, move |s, v0, v1, v2| {
            let v0 = v0.cast_to::<TDim>()?;
            let v1 = v1.cast_to::<TDim>()?;
            let v2 = v2.cast_to::<i64>()?;
            let out = (v1.to_scalar::<TDim>()?.clone() - v0.to_scalar::<TDim>()?)
                .divceil(*v2.to_scalar::<i64>()? as _);
            s.equals(&outputs[0].shape[0], out)
        })?;
        Ok(())
    }

    fn wire(
        &self,
        prefix: &str,
        model: &mut TypedModel,
        inputs: &[OutletId],
    ) -> TractResult<TVec<OutletId>> {
        let start = model
            .outlet_fact(inputs[0])?
            .konst
            .clone()
            .context("Range needs fixed inputs")?
            .into_tensor();
        let end = model
            .outlet_fact(inputs[1])?
            .konst
            .clone()
            .context("Range needs fixed inputs")?
            .into_tensor();
        let step = model
            .outlet_fact(inputs[2])?
            .konst
            .clone()
            .context("Range needs fixed inputs")?
            .into_tensor();
        let dt =
            DatumType::super_type_for([start.datum_type(), end.datum_type(), step.datum_type()])
                .context("No supertype found for range inputs")?;
        let start = start.cast_to_dt(dt)?.into_owned();
        let end = end.cast_to_dt(dt)?.into_owned();
        let step = step.cast_to_dt(dt)?.into_owned();
        model.wire_node(prefix, tract_core::ops::array::Range { start, end, step }, &[])
    }
}