From fbea96f6d76b082a2587d0fc36e30dcaad15b6ce Mon Sep 17 00:00:00 2001
From: Ray Wong <wangbudui@foxmail.com>
Date: Thu, 24 Sep 2020 16:50:53 +0800
Subject: [PATCH] add new dataset type

---
 data/dataset.py | 48 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/data/dataset.py b/data/dataset.py
index 2941889..9e53028 100644
--- a/data/dataset.py
+++ b/data/dataset.py
@@ -1,6 +1,7 @@
 import os
 import pickle
 from collections import defaultdict
+from itertools import permutations, combinations
 from pathlib import Path
 
 import lmdb
@@ -237,3 +238,50 @@ class GenerationUnpairedDatasetWithEdge(Dataset):
 
     def __repr__(self):
         return f"<GenerationUnpairedDatasetWithEdge:\n\tA: {self.A}\n\tB: {self.B}>\nPipeline:\n{self.A.pipeline}"
+
+
+@DATASET.register_module()
+class PoseFacesWithSingleAnime(Dataset):
+    def __init__(self, root_face, root_anime, landmark_path, num_face, face_pipeline, anime_pipeline, img_size,
+                 with_order=True):
+        self.num_face = num_face
+        self.landmark_path = Path(landmark_path)
+        self.with_order = with_order
+        self.root_face = Path(root_face)
+        self.root_anime = Path(root_anime)
+        self.img_size = img_size
+        self.face_samples = self.iter_folders()
+        self.face_pipeline = transform_pipeline(face_pipeline)
+        self.B = SingleFolderDataset(root_anime, anime_pipeline, with_path=True)
+
+    def iter_folders(self):
+        pics_per_person = defaultdict(list)
+        for p in self.root_face.glob("*.jpg"):
+            pics_per_person[p.stem[:7]].append(p.stem)
+        data = []
+        for p in pics_per_person:
+            if len(pics_per_person[p]) >= self.num_face:
+                if self.with_order:
+                    data.extend(list(combinations(pics_per_person[p], self.num_face)))
+                else:
+                    data.extend(list(permutations(pics_per_person[p], self.num_face)))
+        return data
+
+    def read_pose(self, pose_txt):
+        key_points, part_labels, part_edge = dlib_landmark.read_keypoints(pose_txt, size=self.img_size)
+        dist_tensor = normalize_tensor(torch.from_numpy(dlib_landmark.dist_tensor(key_points, size=self.img_size)))
+        part_labels = normalize_tensor(torch.from_numpy(part_labels))
+        part_edge = torch.from_numpy(part_edge).unsqueeze(0).float()
+        return torch.cat([part_labels, part_edge, dist_tensor])
+
+    def __len__(self):
+        return len(self.face_samples)
+
+    def __getitem__(self, idx):
+        output = dict()
+        output["anime_img"], output["anime_path"] = self.B[torch.randint(len(self.B), (1,)).item()]
+        for i, f in enumerate(self.face_samples[idx]):
+            output[f"face_{i}"] = self.face_pipeline(self.root_face / f"{f}.jpg")
+            output[f"pose_{i}"] = self.read_pose(self.landmark_path / self.root_face.name / f"{f}.txt")
+            output[f"stem_{i}"] = f
+        return output