added initial

app.py

@@ -0,0 +1,59 @@
+import streamlit as st
+import streamlit.components.v1 as components
+import pyvista as pv
+from pyvista import examples
+import numpy as np
+from dcgan import DCGAN3D_G
+import torch
+import requests
+
+url = "https://raw.githubusercontent.com/LukasMosser/PorousMediaGan/raw/master/checkpoints/berea/berea_generator_epoch_24.pth"
+
+# If repo is private - we need to add a token in header:
+resp = requests.get(url)
+print(resp.status_code)
+
+pv.set_plot_theme("document")
+pl = pv.Plotter(shape=(1, 1),
+                     window_size=(800, 800))
+
+netG = DCGAN3D_G(64, 512, 1, 32, 1)
+netG.load_state_dict(torch.load("./src/berea_generator_epoch_24.pth"))
+z = torch.randn(1, 512, 5, 5, 5)
+with torch.no_grad():
+    X = netG(z)
+print(X.size())
+print(X.min(), X.max())
+st.image((X[0, 0, 32].numpy()+1)/2, output_format="png")
+"""
+data = examples.load_channels()
+channels = data.threshold([0.9, 1.1])
+print(channels)
+bodies = channels.split_bodies()
+# Now remove all bodies with a small volume
+for key in bodies.keys():
+    b = bodies[key]
+    vol = b.volume
+    if vol < 1000.0:
+        del bodies[key]
+        continue
+    # Now lets add a volume array to all blocks
+    b.cell_data["TOTAL VOLUME"] = np.full(b.n_cells, vol)
+
+for i, body in enumerate(bodies):
+    print(f"Body {i:02d} volume: {body.volume:.3f}")
+
+pl.add_mesh(bodies)
+pl.export_html('pyvista.html')
+
+st.header("test html import")
+view_width = 800
+view_height = 800
+HtmlFile = open("pyvista.html", 'r', encoding='utf-8')
+source_code = HtmlFile.read()
+
+components.html(source_code, width=view_width, height=view_height)
+
+#snippet = embed.embed_snippet(views=view(reader.GetOutput()))
+#html = embed.html_template.format(title="", snippet=snippet)
+#components.html(html, width=view_width, height=view_height)"""

dcgan.py

@@ -0,0 +1,50 @@
+import torch.nn as nn
+
+
+class DCGAN3D_G(nn.Module):
+    def __init__(self, isize, nz, nc, ngf, ngpu, n_extra_layers=0):
+        super(DCGAN3D_G, self).__init__()
+        self.ngpu = ngpu
+        assert isize % 16 == 0, "isize has to be a multiple of 16"
+
+        cngf, tisize = ngf // 2, 4
+        while tisize != isize:
+            cngf = cngf * 2
+            tisize = tisize * 2
+
+        main = nn.Sequential(
+            # input is Z, going into a convolution
+            nn.ConvTranspose3d(nz, cngf, 4, 1, 0, bias=False),
+            nn.BatchNorm3d(cngf),
+            nn.ReLU(True),
+        )
+
+        i, csize, cndf = 3, 4, cngf
+        while csize < isize // 2:
+            main.add_module(str(i),
+                            nn.ConvTranspose3d(cngf, cngf // 2, 4, 2, 1, bias=False))
+            main.add_module(str(i + 1),
+                            nn.BatchNorm3d(cngf // 2))
+            main.add_module(str(i + 2),
+                            nn.ReLU(True))
+            i += 3
+            cngf = cngf // 2
+            csize = csize * 2
+
+        # Extra layers
+        for t in range(n_extra_layers):
+            main.add_module(str(i),
+                            nn.Conv3d(cngf, cngf, 3, 1, 1, bias=False))
+            main.add_module(str(i + 1),
+                            nn.BatchNorm3d(cngf))
+            main.add_module(str(i + 2),
+                            nn.ReLU(True))
+            i += 3
+
+        main.add_module(str(i),
+                        nn.ConvTranspose3d(cngf, nc, 4, 2, 1, bias=False))
+        main.add_module(str(i + 1), nn.Tanh())
+        self.main = main
+
+    def forward(self, input):
+        return self.main(input)

requirements.txt

@@ -0,0 +1,8 @@
+pyvista
+streamlit
+pythreejs
+matplotlib
+torch==1.10.1+cu113
+torchvision==0.11.2+cu113
+torchaudio==0.10.1+cu113
+-f https://download.pytorch.org/whl/cu113/torch_stable.html