Move to modules instead of classes for home page and common interface

data_explorer/app/config.py

@@ -1,12 +1,9 @@
 """General configuration for the app."""
 
 SESSION_STATE_VARIABLES = [
-    "base_path",
-    "pipeline",
     "selected_pipeline_path",
     "run",
     "selected_run_path",
-    "available_runs",
     "component",
     "selected_component_path",
     "partition",

data_explorer/app/interfaces/common.py

@@ -0,0 +1,55 @@
+"""Interface for the data explorer app."""
+
+import os
+
+import streamlit as st
+from interfaces.utils import get_index_from_state
+
+
+def render_common_interface():
+    cols = st.columns(2)
+    with cols[0]:
+        _select_pipeline()
+    with cols[1]:
+        _select_run()
+
+
+def _select_pipeline():
+    """Selects a pipeline from available pipelines."""
+    available_pipelines = [
+        os.path.basename(item)
+        for item in st.session_state.file_system.ls(st.session_state.base_path)
+    ]
+
+    selected_pipeline = st.selectbox(
+        "Pipeline",
+        options=available_pipelines,
+        index=get_index_from_state("pipeline", available_pipelines),
+    )
+    st.session_state.pipeline = selected_pipeline
+    st.session_state.pipeline_path = st.session_state.base_path / selected_pipeline
+
+
+def _select_run():
+    """Selects a run from available runs within the chosen pipeline."""
+    pipeline_path = st.session_state.base_path / st.session_state.pipeline
+
+    def has_manifest_file(path):
+        return any("manifest.json" in files for _, _, files in os.walk(path))
+
+    available_runs = []
+    for run in os.listdir(pipeline_path):
+        run_path = pipeline_path / run
+        if run_path.is_dir() and run != "cache" and has_manifest_file(run_path):
+            available_runs.append(os.path.basename(run))
+
+    available_runs.sort(reverse=True)
+
+    selected_run = st.selectbox(
+        "Run",
+        options=available_runs,
+        index=get_index_from_state("run", available_runs),
+    )
+    st.session_state.run = selected_run
+    st.session_state.run_path = st.session_state.pipeline_path / selected_run
+    st.session_state.available_runs = available_runs

data_explorer/app/interfaces/dataset_interface.py

@@ -6,19 +6,16 @@
 from pathlib import Path
 
 import dask.dataframe as dd
+import fsspec
 import pandas as pd
 import streamlit as st
 from config import DEFAULT_INDEX_NAME, ROWS_TO_RETURN
 from fondant.core.manifest import Manifest
 from fondant.core.schema import Field
-from interfaces.common_interface import MainInterface
-from interfaces.utils import get_default_index
+from interfaces.utils import get_index_from_state
 
 
-class DatasetLoaderApp(MainInterface):
-    def __init__(self):
-        super().__init__()
-
+class DatasetLoaderApp:
     @staticmethod
     def _select_component():
         """Select component from available components."""
@@ -27,13 +24,12 @@ def _select_component():
             os.path.basename(item) for item in os.listdir(selected_run_path)
         ]
 
-        default_index = get_default_index("component", available_components)
         selected_component = st.selectbox(
             "Component",
-            available_components,
-            default_index,
+            options=available_components,
+            index=get_index_from_state("component", available_components),
         )
-        selected_component_path = os.path.join(selected_run_path, selected_component)
+        selected_component_path = selected_run_path / selected_component
 
         st.session_state["component"] = selected_component
         st.session_state["selected_component_path"] = selected_component_path
@@ -74,9 +70,9 @@ def _get_field_location(self, manifest: Manifest, field_name: str) -> str:
         base_path = st.session_state["base_path"]
         field_location = manifest.get_field_location(field_name)
 
-        if (
-            os.path.ismount(base_path) is False
-            and self.fs.__class__.__name__ == "LocalFileSystem"
+        if os.path.ismount(base_path) is False and isinstance(
+            st.session_state.file_system,
+            fsspec.implementations.local.LocalFileSystem,
         ):
             # Used for local development when running the app locally
             field_location = os.path.join(

data_explorer/app/interfaces/sidebar.py

@@ -0,0 +1,22 @@
+import streamlit as st
+from streamlit_extras.app_logo import add_logo
+
+
+def render_sidebar():
+    add_logo("content/fondant_logo.png")
+
+    with st.sidebar:
+        # Increase the width of the sidebar to accommodate logo
+        st.markdown(
+            """
+            <style>
+                section[data-testid="stSidebar"] {
+                    width: 350px !important;
+                }
+            </style>
+            """,
+            unsafe_allow_html=True,
+        )
+
+        with st.expander("## General Configuration"):
+            st.markdown(f"### Base path: \n {st.session_state.base_path}")

data_explorer/app/interfaces/utils.py

@@ -3,9 +3,10 @@
 import streamlit as st
 
 
-def get_default_index(key: str, option_list: list) -> int:
+def get_index_from_state(key: str, option_list: list) -> int:
     """Get the default index for a selectbox based on previous selection from session state."""
-    if st.session_state[key] is None:
+    selected_option = st.session_state.get(key)
+    if selected_option is None or selected_option not in option_list:
         return 0
 
-    return option_list.index(st.session_state[key])
+    return option_list.index(selected_option)