Update app.py

app.py

@@ -1,151 +1,125 @@
-from pathlib import Path
-from typing import List, Dict, Tuple
-import matplotlib.colors as mpl_colors
-
-import pandas as pd
-import seaborn as sns
-import shinyswatch
-
-from shiny import App, Inputs, Outputs, Session, reactive, render, req, ui
-
-sns.set_theme()
-
-www_dir = Path(__file__).parent.resolve() / "www"
-
-df = pd.read_csv(Path(__file__).parent / "penguins.csv", na_values="NA")
-numeric_cols: List[str] = df.select_dtypes(include=["float64"]).columns.tolist()
-species: List[str] = df["Species"].unique().tolist()
-species.sort()
-
-app_ui = ui.page_fillable(
-    shinyswatch.theme.minty(),
-    ui.layout_sidebar(
-        ui.sidebar(
-            # Artwork by @allison_horst
-            ui.input_selectize(
-                "xvar",
-                "X variable",
-                numeric_cols,
-                selected="Bill Length (mm)",
-            ),
-            ui.input_selectize(
-                "yvar",
-                "Y variable",
-                numeric_cols,
-                selected="Bill Depth (mm)",
-            ),
-            ui.input_checkbox_group(
-                "species", "Filter by species", species, selected=species
-            ),
-            ui.hr(),
-            ui.input_switch("by_species", "Show species", value=True),
-            ui.input_switch("show_margins", "Show marginal plots", value=True),
-        ),
-        ui.output_ui("value_boxes"),
-        ui.output_plot("scatter", fill=True),
-        ui.help_text(
-            "Artwork by ",
-            ui.a("@allison_horst", href="https://twitter.com/allison_horst"),
-            class_="text-end",
-        ),
-    ),
+import ipyleaflet as L
+from transformers import SamModel, SamConfig, SamProcessor
+import torch
+from faicons import icon_svg
+from geopy.distance import geodesic, great_circle
+from shiny import reactive
+from shiny.express import input, render, ui
+from shinywidgets import render_widget
+import numpy as np
+import ipywidgets as widgets
+import io
+import base64
+from PIL import Image
+import matplotlib.pyplot as plt
+
+ui.tags.style(
+    "#file1_progress { height: 100%; }",
+    ".bslib-sidebar-layout {--_sidebar-width: 360px !important; }",
+    " img { object-fit: contain; }",
 )
 
+ui.page_opts(title="Segment Anything Model: Sidewalk Masking", fillable=True)
+{"class": "bslib-page-dashboard"}
 
-def server(input: Inputs, output: Outputs, session: Session):
-    @reactive.Calc
-    def filtered_df() -> pd.DataFrame:
-        """Returns a Pandas data frame that includes only the desired rows"""
-
-        # This calculation "req"uires that at least one species is selected
-        req(len(input.species()) > 0)
-
-        # Filter the rows so we only include the desired species
-        return df[df["Species"].isin(input.species())]
-
-    @output
-    @render.plot
-    def scatter():
-        """Generates a plot for Shiny to display to the user"""
-
-        # The plotting function to use depends on whether margins are desired
-        plotfunc = sns.jointplot if input.show_margins() else sns.scatterplot
-
-        plotfunc(
-            data=filtered_df(),
-            x=input.xvar(),
-            y=input.yvar(),
-            palette=palette,
-            hue="Species" if input.by_species() else None,
-            hue_order=species,
-            legend=False,
-        )
-
-    @output
-    @render.ui
-    def value_boxes():
-        df = filtered_df()
-
-        def penguin_value_box(title: str, count: int, bgcol: str, showcase_img: str):
-            return ui.value_box(
-                title,
-                count,
-                {"class_": "pt-1 pb-0"},
-                showcase=ui.fill.as_fill_item(
-                    ui.tags.img(
-                        {"style": "object-fit:contain;"},
-                        src=showcase_img,
-                    )
-                ),
-                theme_color=None,
-                style=f"background-color: {bgcol};",
-            )
-
-        if not input.by_species():
-            return penguin_value_box(
-                "Penguins",
-                len(df.index),
-                bg_palette["default"],
-                # Artwork by @allison_horst
-                showcase_img="penguins.png",
-            )
-
-        value_boxes = [
-            penguin_value_box(
-                name,
-                len(df[df["Species"] == name]),
-                bg_palette[name],
-                # Artwork by @allison_horst
-                showcase_img=f"{name}.png",
-            )
-            for name in species
-            # Only include boxes for _selected_ species
-            if name in input.species()
-        ]
-
-        return ui.layout_column_wrap(*value_boxes, width = 1 / len(value_boxes))
-
-
-# "darkorange", "purple", "cyan4"
-colors = [[255, 140, 0], [160, 32, 240], [0, 139, 139]]
-colors = [(r / 255.0, g / 255.0, b / 255.0) for r, g, b in colors]
-
-palette: Dict[str, Tuple[float, float, float]] = {
-    "Adelie": colors[0],
-    "Chinstrap": colors[1],
-    "Gentoo": colors[2],
-    "default": sns.color_palette()[0],  # type: ignore
-}
-
-bg_palette = {}
-# Use `sns.set_style("whitegrid")` to help find approx alpha value
-for name, col in palette.items():
-    # Adjusted n_colors until `axe` accessibility did not complain about color contrast
-    bg_palette[name] = mpl_colors.to_hex(sns.light_palette(col, n_colors=7)[1])  # type: ignore
-
-
-app = App(
-    app_ui,
-    server,
-    static_assets=str(www_dir),
-)
+with ui.sidebar():
+
+    ui.input_file("file1", "Upload Image", accept=[".jpg", ".png", ".jpeg"], multiple=False),
+    ui.input_dark_mode(mode="dark")
+
+with ui.card():
+
+    ui.card_header("Finalized Segment")
+
+    @render.text
+    def slider_val():
+        if input.file1() is None:
+            return None
+        else:
+            return "Here is the prediction mask:"
+            # return input.file1()[0]['datapath']
+
+    def getSegments():
+        # Load the model configuration
+        model_config = SamConfig.from_pretrained("facebook/sam-vit-base")
+        processor = SamProcessor.from_pretrained("facebook/sam-vit-base")
+
+        # Create an instance of the model architecture with the loaded configuration
+        my_mito_model = SamModel(config=model_config)
+        #Update the model by loading the weights from saved file.
+        my_mito_model.load_state_dict(torch.load("../modelv2.pth"))
+
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        my_mito_model.to(device)
+
+        # Define the size of your array
+        array_size = 256
+
+        # Define the size of your grid
+        grid_size = 10
+
+        # Generate the grid points
+        x = np.linspace(0, array_size-1, grid_size)
+        y = np.linspace(0, array_size-1, grid_size)
+
+        # Generate a grid of coordinates
+        xv, yv = np.meshgrid(x, y)
+
+        # Convert the numpy arrays to lists
+        xv_list = xv.tolist()
+        yv_list = yv.tolist()
+
+        # Combine the x and y coordinates into a list of list of lists
+        input_points = [[[int(x), int(y)] for x, y in zip(x_row, y_row)] for x_row, y_row in zip(xv_list, yv_list)]
+        input_points = torch.tensor(input_points).view(1, 1, grid_size*grid_size, 2)
+
+        inputs = processor(Image.open(input.file1()[0]['datapath']), input_points=input_points, return_tensors="pt")
+
+        inputs = {k: v.to(device) for k, v in inputs.items()}
+        my_mito_model.eval()
+
+        # forward pass
+        with torch.no_grad():
+            outputs = my_mito_model(**inputs, multimask_output=False)
+
+        # apply sigmoid
+        single_patch_prob = torch.sigmoid(outputs.pred_masks.squeeze(1))
+        # convert soft mask to hard mask
+        single_patch_prob = single_patch_prob.cpu().numpy().squeeze()
+        single_patch_prediction = (single_patch_prob > 0.5).astype(np.uint8)
+        return single_patch_prediction
+
+    # @render.image
+    # def render_image():
+    #     # Get the uploaded file
+    #     uploaded_file = input.file1()
+
+    #     # If there is no uploaded file, return None
+    #     if uploaded_file is None:
+    #         return None
+
+    #     # Read the image file
+    #     imagePath = uploaded_file[0]['datapath']
+
+    #     # processImage()
+    #     return {"src": imagePath, "width": "100%"}
+
+    @render.image
+    def render_image():
+        # Get the uploaded file
+        uploaded_file = input.file1()
+
+        # If there is no uploaded file, return None
+        if uploaded_file is None:
+            return None
+
+        # Call getSegments to get the segmented image numpy array
+        segmented_image = np.array(getSegments())
+        segmented_image = segmented_image * 255
+        colorArray = segmented_image.astype(np.uint8)
+        image = Image.fromarray(colorArray)
+        
+        imagePath = "test.jpg"
+        image.save(imagePath)
+
+        return {"src": imagePath, "height": "100%", "class": "contain"}