[Fix] Fix doc and requirements.txt

docs/index.md

@@ -94,7 +94,7 @@
     pip install paddlesci
     ```
 
-=== "[完整安装流程](./zh/install_setup.md)"
+- **完整安装流程**：[安装与使用](./zh/install_setup.md)
 
 --8<--
 ./README.md:feature

docs/requirements.txt

@@ -1,4 +1,4 @@
-Jinja2==3.0.3
+Jinja2~=3.1
 matplotlib
 mkdocs
 mkdocs-autorefs

docs/stylesheets/extra.css

@@ -7,7 +7,7 @@
 
 .md-grid {
   /* readable page width */
-  max-width: 1440px;
+  max-width: 70%;
 }
 
 .md-header__topic > .md-ellipsis {

docs/zh/examples/labelfree_DNN_surrogate.md

@@ -233,7 +233,7 @@ examples/pipe/poiseuille_flow.py:152:164
 
 1. 在 $x=0$ 截面速度 $u(y)$ 随 $y$ 在四种不同的动力粘性系数 ${\nu}$ 采样下的曲线和解析解的对比
 
-2. 当我们选取截断高斯分布的动力粘性系数 ${\nu}$ 采样(均值为 $\hat{\nu} = 10^{−3}$， 方差 $\sigma_{\nu}​=2.67×10^{−4}$)，中心处速度的概率密度函数和解析解对比
+2. 当我们选取截断高斯分布的动力粘性系数 ${\nu}$ 采样(均值为 $\hat{\nu} = 10^{−3}$， 方差 $\sigma_{\nu}​=2.67 \times 10^{−4}$)，中心处速度的概率密度函数和解析解对比
 
 ``` py linenums="166"
 --8<--
@@ -253,7 +253,7 @@ examples/pipe/poiseuille_flow.py
 
 <figure markdown>
   ![laplace 2d]( https://paddle-org.bj.bcebos.com/paddlescience/docs/labelfree_DNN_surrogate/pipe_result.png){ loading=lazy }
-  <figcaption>(左)在 x=0 截面速度 u(y) 随 y 在四种不同的动力粘性系数采样下的曲线和解析解的对比 (右)当我们选取截断高斯分布的动力粘性系数 nu 采样(均值为 nu=0.001， 方差 sigma​=2.67×10e−4)，中心处速度的概率密度函数和解析解对比</figcaption>
+  <figcaption>(左)在 x=0 截面速度 u(y) 随 y 在四种不同的动力粘性系数采样下的曲线和解析解的对比 (右)当我们选取截断高斯分布的动力粘性系数 nu 采样(均值为 nu=0.001， 方差 sigma​=2.67 x 10e−4)，中心处速度的概率密度函数和解析解对比</figcaption>
 </figure>
 
 DNN代理模型的结果如左图所示，和泊肃叶流动的精确解(论文公式13)进行比较：

docs/zh/user_guide.md

@@ -28,7 +28,7 @@ pip install hydra-core
 
 以 bracket 案例为例，其正常运行命令为：`python bracket.py`。若在其运行命令末尾加上  `-c job`，则可以打印出从运行配置文件 `conf/bracket.yaml` 中解析出的配置参数，如下所示。
 
-``` shell title=">>> python bracket.py {++-c job++}"
+``` shell title="$ python bracket.py {++-c job++}"
 mode: train
 seed: 2023
 output_dir: ${hydra:run.dir}
@@ -95,7 +95,7 @@ TRAIN:
 
 执行如下命令即可按顺序自动运行这 4 组实验。
 
-``` shell title=">>> python bracket.py {++-m seed=42,1024 TRAIN.epochs=10,20++}"
+``` shell title="$ python bracket.py {++-m seed=42,1024 TRAIN.epochs=10,20++}"
 [HYDRA] Launching 4 jobs locally
 [HYDRA]        #0 : seed=42 TRAIN.epochs=10
 ....
@@ -109,7 +109,7 @@ TRAIN:
 
 多组实验各自的参数文件、日志文件则保存在以不同参数组合为名称的子文件夹中，如下所示。
 
-``` shell title=">>> tree PaddleScience/examples/bracket/outputs_bracket/"
+``` shell title="$ tree PaddleScience/examples/bracket/outputs_bracket/"
 PaddleScience/examples/bracket/outputs_bracket/
 └──2023-10-14 # (1)
     └── 04-01-52 # (2)

ppsci/utils/misc.py

@@ -345,13 +345,13 @@ def cartesian_product(*arrays: np.ndarray) -> np.ndarray:
     Reference: https://stackoverflow.com/questions/11144513/cartesian-product-of-x-and-y-array-points-into-single-array-of-2d-points
 
     Assume shapes of input arrays are: $(N_1,), (N_2,), (N_3,), ..., (N_M,)$,
-    then the cartesian product result will be shape of $(N_1×N_2×N_3×...×N_M, M)$.
+    then the cartesian product result will be shape of $(N_1xN_2xN_3x...xN_M, M)$.
 
     Args:
         arrays (np.ndarray): Input arrays.
 
     Returns:
-        np.ndarray: Cartesian product result of shape $(N_1×N_2×N_3×...×N_M, M)$.
+        np.ndarray: Cartesian product result of shape $(N_1xN_2xN_3x...xN_M, M)$.
 
     Examples:
         >>> t = np.array([1, 2])

pyproject.toml

@@ -33,26 +33,25 @@ classifiers = [
     "Topic :: Scientific/Engineering :: Mathematics",
 ]
 dependencies = [
-    "numpy>=1.20.0,<=1.23.1",
-    "sympy",
+    "colorlog",
+    "h5py",
+    "hydra-core",
+    "imageio",
     "matplotlib",
-    "vtk",
+    "meshio==5.3.4",
+    "numpy>=1.20.0,<=1.23.1",
     "pyevtk",
-    "wget",
-    "scipy",
-    "visualdl",
     "pyvista==0.37.0",
     "pyyaml",
     "scikit-optimize",
-    "h5py",
-    "meshio==5.3.4",
+    "scipy",
+    "seaborn",
+    "sympy",
     "tqdm",
-    "imageio",
     "typing-extensions",
-    "seaborn",
-    "colorlog",
-    "hydra-core",
-    "opencv-python",
+    "visualdl",
+    "vtk",
+    "wget",
 ]
 
 [project.urls]