Drag drop Streamlit

pyproject.toml

@@ -66,6 +66,8 @@ test = [
     "geopandas",
 ]
 
+app = ["streamlit>=1.41.1"]
+
 notebooks = ["nbformat", "nbconvert", "jupyter", "plotly", "shapely", "seaborn"]
 
 [project.urls]

web/README.md

@@ -0,0 +1,15 @@
+# FMSkill web application
+
+## Installation
+
+```
+$pip install streamlit
+```
+
+Usage:
+```
+$streamlit run web/app.py
+```
+
+![](screenshot.png)
+

web/app.py

@@ -0,0 +1,55 @@
+import tempfile
+from pathlib import Path
+import streamlit as st
+import mikeio
+import modelskill
+import matplotlib.pyplot as plt
+
+"""
+# ModelSkill Drag and Drop
+"""
+
+tmp = Path(tempfile.mkdtemp())
+
+with st.sidebar:
+    obs = st.file_uploader("Observation", type="dfs0")
+
+    if obs:
+        fn_obs = tmp / "obs.dfs0"
+        fn_obs.write_bytes(obs.getvalue())
+
+        dfs = mikeio.open(fn_obs)
+        obs_item = st.selectbox("Item", options=[i.name for i in dfs.items])
+
+    mod = st.file_uploader("Model", type="dfs0")
+
+    if mod:
+        fn_mod = tmp / "mod.dfs0"
+        fn_mod.write_bytes(mod.getvalue())
+        mdfs = mikeio.open(fn_mod)
+        mod_item = st.selectbox("Item", options=[i.name for i in mdfs.items])
+
+    metrics = st.multiselect(
+        "Metrics",
+        ["bias", "rmse", "mae", "cc", "r2", "si", "kge", "mape", "urmse"],
+        default=["bias", "rmse", "r2"],
+    )
+
+if mod and obs:
+    c = modelskill.match(
+        fn_obs, fn_mod, obs_item=obs_item, mod_item=mod_item, gtype="point"
+    )
+
+    tskill, tts, tsc = st.tabs(["Skill", "Time series", "Scatter"])
+
+    with tskill:
+        df = c.skill(metrics=metrics).to_dataframe()
+        st.dataframe(df)
+
+    with tts:
+        c.plot.timeseries()
+        st.pyplot(plt.gcf())
+
+    with tsc:
+        c.plot.scatter()
+        st.pyplot(plt.gcf())