power / gui /pages /agentic_linear.py
Deepakresearcher's picture
New build
b555842
Raw History Blame Contribute Delete
5.56 kB
"""
gui/pages/agentic_linear.py — ⇌ Agentic vs Linear
Render function: render(ctx)
ctx keys: ov, runs, tax, lin, age, avg_lin_j, avg_age_j, tax_mult,
plan_ms, exec_ms, synth_ms, plan_pct, exec_pct, synth_pct
"""
import subprocess
import streamlit as st
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from gui.config import PROJECT_ROOT, DB_PATH, LIVE_API, WF_COLORS, PL
from gui.db import q, q_safe, q1
from gui.helpers import fl, _human_energy, _human_water, _human_carbon, _gauge_html, _bar_gauge_html
try:
import requests as _req
_REQUESTS_OK = True
except ImportError:
_REQUESTS_OK = False
class _req:
@staticmethod
def get(*a, **kw): raise RuntimeError("requests not installed")
try:
import yaml as _yaml
_YAML_OK = True
except ImportError:
_YAML_OK = False
def render(ctx: dict):
ov = ctx["ov"]
runs = ctx["runs"]
tax = ctx["tax"]
avg_lin_j = ctx["avg_lin_j"]
avg_age_j = ctx["avg_age_j"]
tax_mult = ctx["tax_mult"]
plan_ms = ctx["plan_ms"]
exec_ms = ctx["exec_ms"]
synth_ms = ctx["synth_ms"]
plan_pct = ctx["plan_pct"]
exec_pct = ctx["exec_pct"]
synth_pct = ctx["synth_pct"]
lin = ctx["lin"]
age = ctx["age"]
st.title("⇌ Agentic vs Linear — Deep Comparison")
st.caption("Head-to-head on every metric that matters for your PhD")
_cmp, _cmp_e = q_safe("""
SELECT e.task_name, e.provider, e.model_name,
r.workflow_type,
COUNT(*) AS runs,
ROUND(AVG(r.total_energy_uj)/1e6, 4) AS avg_energy_j,
ROUND(AVG(r.dynamic_energy_uj)/1e6, 4) AS avg_dynamic_j,
ROUND(AVG(r.duration_ns)/1e9, 3) AS avg_duration_s,
ROUND(AVG(r.total_tokens), 1) AS avg_tokens,
ROUND(AVG(CASE WHEN r.total_tokens>0
THEN r.total_energy_uj/r.total_tokens END)/1e3, 4) AS avg_mj_token,
ROUND(AVG(r.ipc), 3) AS avg_ipc,
ROUND(AVG(r.cache_miss_rate)*100, 2) AS avg_cache_miss,
ROUND(AVG(r.carbon_g)*1000, 4) AS avg_carbon_mg,
ROUND(AVG(r.planning_time_ms), 1) AS avg_plan_ms,
ROUND(AVG(r.execution_time_ms), 1) AS avg_exec_ms,
ROUND(AVG(r.synthesis_time_ms), 1) AS avg_synth_ms,
ROUND(AVG(r.llm_calls), 2) AS avg_llm_calls,
ROUND(AVG(r.tool_calls), 2) AS avg_tool_calls,
ROUND(AVG(r.thermal_delta_c), 2) AS avg_temp_rise
FROM runs r JOIN experiments e ON r.exp_id=e.exp_id
GROUP BY e.task_name, e.provider, e.model_name, r.workflow_type
ORDER BY e.task_name, r.workflow_type
""")
if _cmp_e or _cmp.empty:
st.info("No data yet.")
else:
# Pivot to wide format for direct comparison
_lin = _cmp[_cmp.workflow_type=="linear"].set_index(["task_name","provider"])
_age = _cmp[_cmp.workflow_type=="agentic"].set_index(["task_name","provider"])
_metrics = ["avg_energy_j","avg_dynamic_j","avg_duration_s","avg_tokens",
"avg_mj_token","avg_ipc","avg_cache_miss","avg_carbon_mg",
"avg_plan_ms","avg_temp_rise","avg_llm_calls","avg_tool_calls"]
_tab_heat, _tab_bar, _tab_table = st.tabs(
["🌡 Heatmap", "📊 Bar Charts", "📋 Table"])
with _tab_heat:
st.markdown("**Energy overhead ratio (agentic÷linear) per task × provider**")
_ratio = _age["avg_energy_j"] / _lin["avg_energy_j"].replace(0, float("nan"))
_ratio_df = _ratio.reset_index()
_ratio_df.columns = ["task_name","provider","ratio"]
if not _ratio_df.empty:
_fig_h = px.density_heatmap(
_ratio_df, x="provider", y="task_name", z="ratio",
color_continuous_scale="RdYlGn_r",
labels={"ratio":"Agentic/Linear energy ratio"})
st.plotly_chart(fl(_fig_h), use_container_width=True)
st.caption("Values > 1.0 = agentic costs more. Red = high overhead.")
with _tab_bar:
_sel_m = st.selectbox("Metric", _metrics, key="al_metric")
_fig_al = px.bar(_cmp.dropna(subset=[_sel_m]),
x="task_name", y=_sel_m, color="workflow_type",
barmode="group", facet_col="provider",
color_discrete_map=WF_COLORS,
labels={"task_name":"Task"},
height=400)
_fig_al.update_xaxes(tickangle=35)
st.plotly_chart(fl(_fig_al), use_container_width=True)
with _tab_table:
st.dataframe(_cmp.round(4), use_container_width=True, hide_index=True)
# ══════════════════════════════════════════════════════════════════════════════
# PAGE: RESEARCH INSIGHTS (merged sql_query + guided questions + schema)
# ══════════════════════════════════════════════════════════════════════════════