Update app.py

app.py

@@ -1,7 +1,7 @@
 import pandas as pd
 import streamlit as st
 from huggingface_hub import HfApi
-from utils import ascending_metrics, metric_ranges
+from utils import ascending_metrics, metric_ranges, LANGUAGES
 import numpy as np
 from st_aggrid import AgGrid, GridOptionsBuilder, JsCode
 from os.path import exists
@@ -155,114 +155,96 @@ dataset_df = dataset_df[dataset_df.split == split]
 
 dataset_df = dataset_df.dropna(axis="columns", how="all")
 
-selectable_datasets = [dataset]
 dataset = st.sidebar.selectbox(
     "Dataset",
-    selectable_datasets,
+    [dataset],
     index=0,
 )
 
-if len(dataset_df) > 0:
-    selectable_configs = list(set(dataset_df["config"]))
-    selectable_configs.sort(key=lambda name: name.lower())
-
-    selectable_configs.remove("-unspecified-")
-
-    if dataset != "-any-":
-        config = st.sidebar.selectbox(
-            "Language",
-            selectable_configs,
-            index=0,
-            help="Filter the results on the current leaderboard by language."
-        )
-        dataset_df = dataset_df[dataset_df.config == config]
-
-        selectable_splits = [split]
-        split = st.sidebar.selectbox(
-            "Split",
-            selectable_splits,
-            index=0,
-        )
-
-    not_selectable_metrics = ["model_id", "dataset", "split", "config", "pipeline_tag", "only_verified"]
-    # also ignore irrelevant ASR metrics
-    not_selectable_metrics.extend(["wer_without_norm", "mer"])
-
-    selectable_metrics = list(filter(lambda column: column not in not_selectable_metrics, dataset_df.columns))
-
-    dataset_df = dataset_df.filter(["model_id"] + (["dataset"] if dataset == "-any-" else []) + selectable_metrics)
-    dataset_df = dataset_df.dropna(thresh=2)  # Want at least two non-na values (one for model_id and one for a metric).
-
-    sorting_metric = st.sidebar.radio(
-        "Sorting Metric",
-        selectable_metrics,
-        index=selectable_metrics.index(default_metric) if default_metric in selectable_metrics else 0,
-        help="Select the metric to sort the leaderboard by. Click on the metric name in the leaderboard to reverse the sorting order."
-    )
-
-    current_query_params.update({"metric": [sorting_metric]})
-
-    st.experimental_set_query_params(**current_query_params)
-
-    st.markdown(
-        "Please click on the model's name to be redirected to its model card."
-    )
-
-    st.markdown(
-        "Want to beat the leaderboard? Don't see your model here? Simply ..."
-    )
-
-    # Make the default metric appear right after model names and dataset names
-    cols = dataset_df.columns.tolist()
-    cols.remove(sorting_metric)
-    sorting_metric_index = 1 if dataset != "-any-" else 2
-    cols = cols[:sorting_metric_index] + [sorting_metric] + cols[sorting_metric_index:]
-    dataset_df = dataset_df[cols]
-
-    # Sort the leaderboard, giving the sorting metric highest priority and then ordering by other metrics in the case of equal values.
-    dataset_df = dataset_df.sort_values(by=cols[sorting_metric_index:], ascending=[metric in ascending_metrics for metric in cols[sorting_metric_index:]])
-    dataset_df = dataset_df.replace(np.nan, '-')
-
-    # If dataset is "-any-", only show the best model for a random sample of 100 datasets.
-    # Otherwise The leaderboard is way too long and doesn't give the users a feel for all of
-    # the datasets available for a task.
-    if dataset == "-any-":
-        filtered_dataset_df_dict = {column: [] for column in dataset_df.columns}
-        seen_datasets = set()
-        for _, row in dataset_df.iterrows():
-            if row["dataset"] not in seen_datasets:
-                for column in dataset_df.columns:
-                    filtered_dataset_df_dict[column].append(row[column])
-                seen_datasets.add(row["dataset"])
-        dataset_df = pd.DataFrame(filtered_dataset_df_dict)
-        dataset_df = dataset_df.sample(min(100, len(dataset_df)))
-
-    # Make the leaderboard
-    gb = GridOptionsBuilder.from_dataframe(dataset_df)
-    gb.configure_default_column(sortable=False)
-    gb.configure_column(
-        "model_id",
-        cellRenderer=JsCode('''function(params) {return '<a target="_blank" href="https://huggingface.co/'+params.value+'">'+params.value+'</a>'}'''),
-    )
-    if dataset == "-any-":
-        gb.configure_column(
-            "dataset",
-            cellRenderer=JsCode('''function(params) {return '<a target="_blank" href="https://huggingface.co/spaces/autoevaluate/leaderboards?dataset='+params.value+'">'+params.value+'</a>'}'''),
-        )
-    for name in selectable_metrics:
-        gb.configure_column(name, type=["numericColumn","numberColumnFilter","customNumericFormat"], precision=4, aggFunc='sum')
-
-    gb.configure_column(
-        sorting_metric,
-        sortable=True,
-        cellStyle=JsCode('''function(params) { return {'backgroundColor': '#FFD21E'}}''')
-    )
-
-    go = gb.build()
-    fit_columns = len(dataset_df.columns) < 10
-    AgGrid(dataset_df, gridOptions=go, height=28*len(dataset_df) + (35 if fit_columns else 41), allow_unsafe_jscode=True, fit_columns_on_grid_load=fit_columns, enable_enterprise_modules=False)
-
-else:
-    st.markdown(
-        "No " + ("verified" if only_verified_results else "unverified") + " results to display. Try toggling the verified results filter."
-    )
+selectable_configs = list(set(dataset_df["config"]))
+selectable_configs.sort(key=lambda name: name.lower())
+selectable_configs.remove("-unspecified-")
+selectable_configs = [config for config in selectable_configs if config in LANGUAGES]
+
+visual_configs = [f"{config}: {LANGUAGES[config]}" for config in selectable_configs]
+
+config = st.sidebar.selectbox(
+    "Language",
+    visual_configs,
+    index=0,
+    help="Filter the results on the current leaderboard by language."
+)
+
+config = config.split(":")[0]
+
+dataset_df = dataset_df[dataset_df.config == config]
+
+split = st.sidebar.selectbox(
+    "Split",
+    [split],
+    index=0,
+)
+
+not_selectable_metrics = ["model_id", "dataset", "split", "config", "pipeline_tag", "only_verified"]
+# also ignore irrelevant ASR metrics
+not_selectable_metrics.extend(["wer_without_norm", "mer"])
+
+selectable_metrics = list(filter(lambda column: column not in not_selectable_metrics, dataset_df.columns))
+
+dataset_df = dataset_df.filter(["model_id"] + (["dataset"] if dataset == "-any-" else []) + selectable_metrics)
+dataset_df = dataset_df.dropna(thresh=2)  # Want at least two non-na values (one for model_id and one for a metric).
+
+sorting_metric = st.sidebar.radio(
+    "Sorting Metric",
+    selectable_metrics,
+    index=selectable_metrics.index(default_metric) if default_metric in selectable_metrics else 0,
+    help="Select the metric to sort the leaderboard by. Click on the metric name in the leaderboard to reverse the sorting order."
+)
+
+current_query_params.update({"metric": [sorting_metric]})
+
+st.experimental_set_query_params(**current_query_params)
+
+st.markdown(
+    f"This is the leaderboard for {LANGUAGES[config]} ({config})."
+)
+
+st.markdown(
+    "Please click on the model's name to be redirected to its model card."
+)
+
+st.markdown(
+    "Want to beat the leaderboard? Don't see your model here? Simply ..."
+)
+
+# Make the default metric appear right after model names and dataset names
+cols = dataset_df.columns.tolist()
+cols.remove(sorting_metric)
+sorting_metric_index = 1 if dataset != "-any-" else 2
+cols = cols[:sorting_metric_index] + [sorting_metric] + cols[sorting_metric_index:]
+dataset_df = dataset_df[cols]
+
+# Sort the leaderboard, giving the sorting metric highest priority and then ordering by other metrics in the case of equal values.
+dataset_df = dataset_df.sort_values(by=cols[sorting_metric_index:], ascending=[metric in ascending_metrics for metric in cols[sorting_metric_index:]])
+dataset_df = dataset_df.replace(np.nan, '-')
+
+# Make the leaderboard
+gb = GridOptionsBuilder.from_dataframe(dataset_df)
+gb.configure_default_column(sortable=False)
+gb.configure_column(
+    "model_id",
+    cellRenderer=JsCode('''function(params) {return '<a target="_blank" href="https://huggingface.co/'+params.value+'">'+params.value+'</a>'}'''),
+)
+
+for name in selectable_metrics:
+    gb.configure_column(name, type=["numericColumn","numberColumnFilter","customNumericFormat"], precision=4, aggFunc='sum')
+
+gb.configure_column(
+    sorting_metric,
+    sortable=True,
+    cellStyle=JsCode('''function(params) { return {'backgroundColor': '#FFD21E'}}''')
+)
+
+go = gb.build()
+fit_columns = len(dataset_df.columns) < 10
+AgGrid(dataset_df, gridOptions=go, height=28*len(dataset_df) + (35 if fit_columns else 41), allow_unsafe_jscode=True, fit_columns_on_grid_load=fit_columns, enable_enterprise_modules=False)