File size: 8,734 Bytes
9f5e507
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
"""
CascadedFlowModel for grn_scalar.

Key changes from grn_svd:
- Latent target: scalar (B, G) or 4-dim (B, G, 4) instead of (B, G, 128)
- Latent encoder: ContinuousValueEncoder (DIM=1, symmetric with expr) or MultiStatsLatentEncoder (DIM=4)
- Latent decoder: ScalarLatentDecoder (symmetric with ExprDecoder, takes backbone+pert concat)
"""

import torch
import torch.nn as nn
from torch import Tensor
from typing import Optional, Tuple

from .layers import MultiStatsLatentEncoder, ScalarLatentDecoder
from .._scdfm_imports import (
    GeneadaLN,
    ContinuousValueEncoder,
    GeneEncoder,
    BatchLabelEncoder,
    TimestepEmbedder,
    ExprDecoder,
    DifferentialTransformerBlock,
    PerceiverBlock,
    DiffPerceiverBlock,
)


class CascadedFlowModel(nn.Module):
    """
    Cascaded Flow Model with scalar latent target.

    Inputs:
        gene_id:         (B, G)       gene token IDs
        cell_1:          (B, G)       source (control) expression
        x_t:             (B, G)       noised target expression (expr flow)
        z_t:             (B, G) or (B, G, 4)  noised scalar latent (latent flow)
        t_expr:          (B,)         expression flow timestep
        t_latent:        (B,)         latent flow timestep
        perturbation_id: (B, 2)       perturbation token IDs

    Outputs:
        pred_v_expr:   (B, G)              predicted expression velocity
        pred_v_latent: (B, G) or (B, G, 4) predicted latent velocity
    """

    def __init__(
        self,
        ntoken: int = 6000,
        d_model: int = 128,
        nhead: int = 8,
        d_hid: int = 512,
        nlayers: int = 4,
        dropout: float = 0.1,
        fusion_method: str = "differential_perceiver",
        perturbation_function: str = "crisper",
        use_perturbation_interaction: bool = True,
        mask_path: str = None,
        latent_dim: int = 1,
    ):
        super().__init__()
        self.d_model = d_model
        self.latent_dim = latent_dim
        self.fusion_method = fusion_method
        self.perturbation_function = perturbation_function

        # === Timestep embedders (separate for expr and latent) ===
        self.t_expr_embedder = TimestepEmbedder(d_model)
        self.t_latent_embedder = TimestepEmbedder(d_model)

        # === Perturbation embedder ===
        self.perturbation_embedder = BatchLabelEncoder(ntoken, d_model)

        # === Expression stream (reused from scDFM) ===
        self.value_encoder_1 = ContinuousValueEncoder(d_model, dropout)
        self.value_encoder_2 = ContinuousValueEncoder(d_model, dropout)
        self.encoder = GeneEncoder(
            ntoken, d_model,
            use_perturbation_interaction=use_perturbation_interaction,
            mask_path=mask_path,
        )
        self.use_perturbation_interaction = use_perturbation_interaction
        self.fusion_layer = nn.Sequential(
            nn.Linear(2 * d_model, d_model),
            nn.GELU(),
            nn.Linear(d_model, d_model),
            nn.LayerNorm(d_model),
        )

        # === Latent stream encoder (symmetric with expression) ===
        if latent_dim == 1:
            # Identical to expression encoder: scalar → d_model MLP
            self.latent_embedder = ContinuousValueEncoder(d_model, dropout)
        else:
            # 4-dim → d_model MLP (mirrors ContinuousValueEncoder structure)
            self.latent_embedder = MultiStatsLatentEncoder(d_model, dropout)

        # === Shared backbone blocks ===
        if fusion_method == "differential_transformer":
            self.blocks = nn.ModuleList([
                DifferentialTransformerBlock(d_model, nhead, i, mlp_ratio=4.0)
                for i in range(nlayers)
            ])
        elif fusion_method == "differential_perceiver":
            self.blocks = nn.ModuleList([
                DiffPerceiverBlock(d_model, nhead, i, mlp_ratio=4.0)
                for i in range(nlayers)
            ])
        elif fusion_method == "perceiver":
            self.blocks = nn.ModuleList([
                PerceiverBlock(d_model, d_model, heads=nhead, mlp_ratio=4.0, dropout=0.1)
                for _ in range(nlayers)
            ])
        else:
            raise ValueError(f"Invalid fusion method: {fusion_method}")

        # === Per-layer gene AdaLN + adapter ===
        self.gene_adaLN = nn.ModuleList([
            GeneadaLN(d_model, dropout) for _ in range(nlayers)
        ])
        self.adapter_layer = nn.ModuleList([
            nn.Sequential(
                nn.Linear(2 * d_model, d_model),
                nn.LeakyReLU(),
                nn.Dropout(dropout),
                nn.Linear(d_model, d_model),
                nn.LeakyReLU(),
            )
            for _ in range(nlayers)
        ])

        # === Expression decoder head (reused from scDFM) ===
        self.final_layer = ExprDecoder(d_model, explicit_zero_prob=False, use_batch_labels=True)

        # === Latent decoder head (symmetric with ExprDecoder) ===
        self.latent_decoder = ScalarLatentDecoder(d_model, latent_dim)

        self.initialize_weights()

    def initialize_weights(self):
        def _basic_init(module):
            if isinstance(module, nn.Linear):
                torch.nn.init.xavier_uniform_(module.weight)
                if module.bias is not None:
                    nn.init.constant_(module.bias, 0)
        self.apply(_basic_init)

    def get_perturbation_emb(
        self,
        perturbation_id: Optional[Tensor] = None,
        perturbation_emb: Optional[Tensor] = None,
        cell_1: Optional[Tensor] = None,
    ) -> Tensor:
        assert perturbation_emb is None or perturbation_id is None
        if perturbation_id is not None:
            if self.perturbation_function == "crisper":
                perturbation_emb = self.encoder(perturbation_id)
            else:
                perturbation_emb = self.perturbation_embedder(perturbation_id)
            perturbation_emb = perturbation_emb.mean(1)
        elif perturbation_emb is not None:
            perturbation_emb = perturbation_emb.to(cell_1.device, dtype=cell_1.dtype)
            if perturbation_emb.dim() == 1:
                perturbation_emb = perturbation_emb.unsqueeze(0)
            if perturbation_emb.size(0) == 1:
                perturbation_emb = perturbation_emb.expand(cell_1.shape[0], -1).contiguous()
            perturbation_emb = self.perturbation_embedder.enc_norm(perturbation_emb)
        return perturbation_emb

    def forward(
        self,
        gene_id: Tensor,        # (B, G)
        cell_1: Tensor,          # (B, G) source expression
        x_t: Tensor,             # (B, G) noised expression
        z_t: Tensor,             # (B, G) or (B, G, 4) noised scalar latent
        t_expr: Tensor,          # (B,)
        t_latent: Tensor,        # (B,)
        perturbation_id: Optional[Tensor] = None,
    ) -> Tuple[Tensor, Tensor]:
        if t_expr.dim() == 0:
            t_expr = t_expr.repeat(cell_1.size(0))
        if t_latent.dim() == 0:
            t_latent = t_latent.repeat(cell_1.size(0))

        # === 1. Expression stream embedding ===
        gene_emb = self.encoder(gene_id)
        val_emb_1 = self.value_encoder_1(x_t)
        val_emb_2 = self.value_encoder_2(cell_1) + gene_emb
        expr_tokens = self.fusion_layer(torch.cat([val_emb_1, val_emb_2], dim=-1)) + gene_emb

        # === 2. Latent stream embedding (symmetric with expression) ===
        # ContinuousValueEncoder handles (B, G) → unsqueeze → (B, G, 1) → MLP → (B, G, d_model)
        # MultiStatsLatentEncoder handles (B, G, 4) → MLP → (B, G, d_model)
        latent_tokens = self.latent_embedder(z_t)

        # === 3. Element-wise addition ===
        x = expr_tokens + latent_tokens

        # === 4. Conditioning vector ===
        t_expr_emb = self.t_expr_embedder(t_expr)
        t_latent_emb = self.t_latent_embedder(t_latent)
        pert_emb = self.get_perturbation_emb(perturbation_id, cell_1=cell_1)
        c = t_expr_emb + t_latent_emb + pert_emb

        # === 5. Shared backbone ===
        for i, block in enumerate(self.blocks):
            x = self.gene_adaLN[i](gene_emb, x)
            pert_exp = pert_emb[:, None, :].expand(-1, x.size(1), -1)
            x = torch.cat([x, pert_exp], dim=-1)
            x = self.adapter_layer[i](x)
            x = block(x, val_emb_2, c)

        # === 6. Decoder heads (both symmetric, both take backbone+pert concat) ===
        x_with_pert = torch.cat([x, pert_emb[:, None, :].expand(-1, x.size(1), -1)], dim=-1)
        pred_v_expr = self.final_layer(x_with_pert)["pred"]   # (B, G)
        pred_v_latent = self.latent_decoder(x_with_pert)       # (B, G) or (B, G, 4)

        return pred_v_expr, pred_v_latent