Client Versions

import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

from loaders import load_parquet

# Load pre-aggregated data
df_hourly = load_parquet("client_hourly", target_date)
df_versions = load_parquet("client_version_dist", target_date)
df_summary = load_parquet("client_summary", target_date)

# Filter out unknown clients for visualizations
df_hourly_known = df_hourly[df_hourly["client"] != "unknown"]
df_versions_known = df_versions[df_versions["client"] != "unknown"]
df_summary_known = df_summary[df_summary["client"] != "unknown"]

# Compute totals from hourly data
total_connections = df_hourly["connections"].sum()
impl_counts = df_hourly_known.groupby("client")["connections"].sum().reset_index()
impl_counts = impl_counts.sort_values("connections", ascending=False)

print(f"Total connections: {total_connections:,}")
print(f"Unique clients: {df_summary['client'].nunique()}")
print(f"Total unique peers: {df_summary['unique_peers'].sum():,}")

fig = px.pie(
    impl_counts,
    values="connections",
    names="client",
    title="Client Implementation Distribution",
    color_discrete_sequence=px.colors.qualitative.Set2,
)
fig.update_traces(textposition="inside", textinfo="percent+label")
fig.update_layout(height=500)
fig.show()

fig = px.bar(
    impl_counts,
    x="client",
    y="connections",
    title="Connections by Client Implementation",
    labels={"client": "Client", "connections": "Connections"},
    color="client",
    color_discrete_sequence=px.colors.qualitative.Set2,
)
fig.update_layout(
    showlegend=False,
    height=500,
    xaxis_tickangle=-45,
)
fig.show()

# Pivot hourly data for stacked area chart
hourly_pivot = df_hourly_known.pivot(index="hour", columns="client", values="connections").fillna(0)

fig = go.Figure()
for col in hourly_pivot.columns:
    fig.add_trace(go.Scatter(
        x=hourly_pivot.index,
        y=hourly_pivot[col],
        mode="lines",
        stackgroup="one",
        name=col,
    ))

fig.update_layout(
    title="Client Implementation Connections Over Time",
    xaxis_title="Time",
    yaxis_title="Connections",
    height=500,
    legend=dict(orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1),
)
fig.show()

# Get clients sorted by total connections (exclude "Others" and "unknown")
top_clients = impl_counts[~impl_counts["client"].isin(["Others", "unknown"])]["client"].tolist()

for i, client in enumerate(top_clients):
    df_client = df_versions_known[df_versions_known["client"] == client]
    
    if len(df_client) == 0:
        continue
    
    fig = px.bar(
        df_client,
        x="version",
        y="connections",
        title=f"{client.capitalize()} Version Distribution",
        labels={"version": "Version", "connections": "Connections"},
        color_discrete_sequence=[px.colors.qualitative.Set2[i % len(px.colors.qualitative.Set2)]],
    )
    fig.update_layout(
        height=400,
        xaxis_tickangle=-45,
    )
    fig.show()

# Format summary table
summary_display = df_summary_known[["client", "connections", "unique_peers", "version_count", "top_version"]].copy()
summary_display.columns = ["Client", "Connections", "Unique Peers", "Versions", "Top Version"]
summary_display = summary_display.sort_values("Connections", ascending=False)
summary_display["Connections"] = summary_display["Connections"].apply(lambda x: f"{x:,}")
summary_display["Unique Peers"] = summary_display["Unique Peers"].apply(lambda x: f"{x:,}")
summary_display