add image_matrix

docs/components/README.md

@@ -287,6 +287,76 @@ visualdl --logdir ./log --port 8080
   <img src="https://visualdl.bj.bcebos.com/images/image-eye.gif" width="60%"/>
 </p>
 
+### 添加图片矩阵
+除使用add_image记录一张图片之外，还可以使用add_image_matrix一次添加多张图片并生成一张图片矩阵，接口及参数说明如下：
+add_image_matrix的记录接口如下：
+
+```python
+add_image_matrix(tag, imgs, step, rows=-1, scale=1, walltime=None, dataformats="HWC")
+```
+接口参数说明如下：
+|   参数   |     格式      |                    含义                     |
+| -------- | ------------- | ------------------------------------------- |
+| tag      | string        | 记录指标的标志，如`train/loss`，不能含有`%` |
+| imgs     | numpy.ndarray | 以ndarray格式表示的多张图片，第一维为图片的数量  |
+| step     | int           | 记录的步数                                  |
+| rows     | int           | 生成图片矩阵的行数，默认值为-1，表示尽量把传入的图片组合成行列数相近的形式 |
+| scale    | int           | 图片放大比例，默认为1 |
+| walltime | int           | 记录数据的时间戳，默认为当前时间戳          |
+| dataformats| string      | 传入的图片格式，包括`NCHW`、`HWC`、`HW`，默认为`HWC`|
+
+#### Demo
+下面展示了使用 Image 组件合成并记录多张图片数据的示例，代码文件请见[Image组件](https://github.com/PaddlePaddle/VisualDL/blob/develop/demo/components/image_matrix_test.py)
+```python
+import numpy as np
+from PIL import Image
+from visualdl import LogWriter
+
+
+def random_crop(img):
+    """Get random block of img, which size is 100x100.
+    """
+    img = Image.open(img)
+    w, h = img.size
+    random_w = np.random.randint(0, w - 100)
+    random_h = np.random.randint(0, h - 100)
+    r = img.crop((random_w, random_h, random_w + 100, random_h + 100))
+    return np.asarray(r)
+
+
+if __name__ == '__main__':
+    imgs = []
+    # 获取8张图像
+    for step in range(8):
+        img = random_crop("../../docs/images/dog.jpg")
+        imgs.append(img)
+    imgs = np.array(imgs)
+
+    with LogWriter(logdir='./log/image_matrix_test/train') as writer:
+        # 合成长宽尽量接近的图形矩阵，本例生成3X3的矩阵
+        writer.add_image_matrix(tag='test_images', step=1, imgs=imgs, rows=-1)
+        # 合成长为1的图形矩阵，本例生成1x8的矩阵
+        writer.add_image_matrix(tag='test_images', step=2, imgs=imgs, rows=1)
+        # 合成长为2的图形矩阵，本例生成2X4的矩阵
+        writer.add_image_matrix(tag='test_images', step=3, imgs=imgs, rows=2)
+        # 合成长为3的图形矩阵，本例生成3X3的矩阵
+        writer.add_image_matrix(tag='test_images', step=4, imgs=imgs, rows=3)
+        # 合成长为4的图形矩阵，本例生成4X2的矩阵
+        writer.add_image_matrix(tag='test_images', step=5, imgs=imgs, rows=4)
+```
+运行上述程序后，在命令行执行
+```shell
+visualdl --logdir ./log --port 8080
+```
+
+在浏览器输入`http://127.0.0.1:8080`，即可查看图片数据。
+
+<p align="center">
+  <img src="https://user-images.githubusercontent.com/28444161/104555348-ad63df80-5678-11eb-9d68-04f7f7451eac.png" width="40%"/>
+  <img src="https://user-images.githubusercontent.com/28444161/104556243-2dd71000-567a-11eb-9222-225b0acdf56b.png" width="40%"/>
+</p>
+
+
 ## Audio--音频播放组件
 
 ### 介绍

visualdl/utils/img_util.py

@@ -0,0 +1,89 @@
+# Copyright (c) 2020 VisualDL Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# =======================================================================
+import math
+from functools import reduce
+
+import numpy as np
+
+from visualdl.component.base_component import convert_to_HWC
+
+
+def padding_image(img, height, width):
+    height_old, width_old, _ = img.shape
+    height_before = math.floor((height - height_old) / 2)
+    height_after = height - height_old - height_before
+
+    width_before = math.floor((width - width_old) / 2)
+    width_after = width - width_old - width_before
+
+    return np.pad(img, ((height_before, height_after), (width_before, width_after), (0, 0)))
+
+
+def merge_images(imgs, dataformats, scale=1.0, rows=-1):
+    assert rows <= len(imgs), "rows should not greater than numbers of pictures"
+    channel = imgs[0].shape[2]
+    # convert format of each image to `hwc`
+    for i, img in enumerate(imgs):
+        imgs[i] = convert_to_HWC(img, dataformats)
+
+    height = -1
+    width = -1
+
+    for img in imgs:
+        height = height if height > img.shape[0] else img.shape[0]
+        width = width if width > img.shape[1] else img.shape[1]
+
+    # padding every sub-image with height and width
+    for i, img in enumerate(imgs):
+        imgs[i] = padding_image(img, height, width)
+
+    # get row and col
+    len_imgs = len(imgs)
+    if -1 == rows:
+        rows = cols = math.floor(math.sqrt(len_imgs))
+        while rows*cols < len_imgs:
+            if rows <= cols:
+                rows += 1
+            else:
+                cols += 1
+    else:
+        cols = math.ceil(len_imgs/rows)
+
+    # add white sub-image
+    for i in range(rows*cols-len_imgs):
+        imgs = np.concatenate((imgs, np.zeros((height, width, channel), dtype=np.uint8)[None, :]))
+
+    imgs = reduce(lambda x, y: np.concatenate((x, y)), [
+        reduce(lambda x, y: np.concatenate((x, y), 1),
+               imgs[i * cols: (i + 1) * cols]) for i in range(rows)])
+
+    # choose bigger number of rows and cols
+
+    scale = 1.0/scale * rows if rows > cols else 1.0/scale * cols
+
+    dsize = tuple(map(lambda x: math.floor(x/scale), imgs.shape))[-2::-1]
+
+    try:
+        import cv2
+
+        imgs = cv2.resize(src=imgs, dsize=dsize)
+    except ImportError:
+        from PIL import Image
+
+        imgs = Image.fromarray(imgs)
+        imgs.resize(dsize)
+        imgs = np.array(imgs)
+
+    return imgs

visualdl/writer/writer.py

@@ -17,6 +17,7 @@
 import time
 import numpy as np
 from visualdl.writer.record_writer import RecordFileWriter
+from visualdl.utils.img_util import merge_images
 from visualdl.component.base_component import scalar, image, embedding, audio, histogram, pr_curve, roc_curve, meta_data
 
 
@@ -189,6 +190,36 @@ def add_image(self, tag, img, step, walltime=None, dataformats="HWC"):
             image(tag=tag, image_array=img, step=step, walltime=walltime,
                   dataformats=dataformats))
 
+    def add_image_matrix(self, tag, imgs, step, rows=-1, scale=1.0, walltime=None, dataformats="HWC"):
+        """Add an image to vdl record file.
+
+        Args:
+            tag (string): Data identifier
+            imgs (np.ndarray): Image represented by a numpy.array
+            step (int): Step of image
+            rows (int): Number of rows, -1 means as close as possible to the square
+            scale (float): Image zoom scale
+            walltime (int): Wall time of image
+            dataformats (string): Format of image
+
+        Example:
+            from PIL import Image
+            import numpy as np
+
+            I = Image.open("./test.png")
+            I_array = np.array([I, I, I])
+            writer.add_image_matrix(tag="lll", imgs=I_array, step=0)
+        """
+        if '%' in tag:
+            raise RuntimeError("% can't appear in tag!")
+        walltime = round(time.time() * 1000) if walltime is None else walltime
+        img = merge_images(imgs=imgs, dataformats=dataformats, scale=scale, rows=rows)
+        self.add_image(tag=tag,
+                       img=img,
+                       step=step,
+                       walltime=walltime,
+                       dataformats=dataformats)
+
     def add_embeddings(self, tag, labels, hot_vectors, labels_meta=None, walltime=None):
         """Add embeddings to vdl record file.