error distribution by mech address

app.py

@@ -47,6 +47,7 @@ from tabs.error import (
     plot_error_data_by_market,
     get_error_data_overall_by_market,
     plot_tool_error_data_by_market,
+    plot_weekly_errors_by_mech,
 )
 
 from tabs.about import about_olas_predict, about_this_dashboard
@@ -154,7 +155,15 @@ def load_all_data():
     )
 
     df7 = pd.read_parquet(daily_mech_req_df)
-    return df1, df2, df3, df4, df5, df6, df7
+
+    # errors by mech
+    errors_by_mech = hf_hub_download(
+        repo_id="valory/Olas-predict-dataset",
+        filename="errors_by_mech.parquet",
+        repo_type="dataset",
+    )
+    df8 = pd.read_parquet(errors_by_mech)
+    return df1, df2, df3, df4, df5, df6, df7, df8
 
 
 def prepare_data():
@@ -169,6 +178,7 @@ def prepare_data():
         unknown_trades,
         winning_df,
         daily_mech_requests,
+        errors_by_mech,
     ) = load_all_data()
     print(trades_df.info())
 
@@ -572,6 +582,14 @@ with demo:
             with gr.Row():
                 tool_error_plot
 
+            with gr.Row():
+                gr.Markdown("# Weekly Error distribution per Mech address")
+
+            with gr.Row():
+                errors_by_mech = plot_weekly_errors_by_mech(
+                    errors_by_mech=errors_by_mech
+                )
+
             with gr.Row():
                 gr.Markdown("# Tools distribution of errors per week")
 

tabs/error.py

@@ -126,3 +126,40 @@ def plot_week_error_data_by_market(error_df: pd.DataFrame, week: str) -> gr.Plot
     fig.update_layout(width=WIDTH, height=HEIGHT)
     fig.update_xaxes(tickformat="%b %d\n%Y")
     return gr.Plot(value=fig)
+
+
+def plot_weekly_errors_by_mech(errors_by_mech: pd.DataFrame) -> gr.Plot:
+    """Function to plot the weekly errors by mech address"""
+    unique_mech_addresses = errors_by_mech["mech_address"].unique()
+
+    for mech_address in unique_mech_addresses:
+        # Filter the data for the current mech_address
+        filtered_data = errors_by_mech[errors_by_mech["mech_address"] == mech_address]
+
+        # Create the bar chart for the filtered data
+        fig = px.bar(
+            filtered_data,
+            x="request_month_year_week",
+            y="requests",
+            color="error_cat",
+            barmode="group",
+            color_discrete_sequence=["green", "orange"],
+            title=f"Error distribution for Mech Address: {mech_address}",
+            labels={
+                "requestmonthyearweek": "Month-Year-Week",
+                "requests": "Number of Requests",
+                "error_cat": "Error Category",
+                "errors_percentage": "Percentage",
+            },
+            text="errors_percentage",
+            height=HEIGHT,
+            width=WIDTH,
+        )
+    fig.update_traces(
+        texttemplate="%{y} (%{text:.2f}%)",
+        textposition="outside",
+        textangle=0,
+        cliponaxis=False,
+    )
+    fig.update_xaxes(tickformat="%b %d\n%Y")
+    return gr.Plot(value=fig)