Upload app.py

app.py

@@ -13,7 +13,12 @@ from quread.circuit_diagram import draw_circuit_svg
 from quread.cost_guard import allow_request
 from quread.export_pdf import md_to_pdf
 from quread.heatmap import make_metric_heatmap, HeatmapConfig
-from quread.metrics import compute_metrics_from_csv, to_metrics_csv
+from quread.metrics import (
+    compute_metrics_from_csv,
+    to_metrics_csv,
+    MetricWeights,
+    MetricThresholds,
+)
 
 # --- Qubit cap (configurable) ---
 DEFAULT_MAX_QUBITS = 16  # safe default for CPU Spaces; change if you want
@@ -180,6 +185,53 @@ def _on_qubit_count_change(n):
     return qc, last_counts, selected_gate, t, c, ct, msg
 
 
+def _metric_controls_to_models(
+    activity_w,
+    gate_error_w,
+    readout_error_w,
+    decoherence_w,
+    fidelity_w,
+    warning_thr,
+    critical_thr,
+):
+    weights = MetricWeights(
+        activity=float(activity_w),
+        gate_error=float(gate_error_w),
+        readout_error=float(readout_error_w),
+        decoherence=float(decoherence_w),
+        fidelity=float(fidelity_w),
+    )
+    thresholds = MetricThresholds(
+        warning=float(warning_thr),
+        critical=float(critical_thr),
+    )
+    return weights, thresholds
+
+
+def _hotspot_rows(metrics, n_qubits, top_k):
+    rows = []
+    n = int(n_qubits)
+    for q in range(n):
+        risk = float(metrics["composite_risk"][q])
+        level = int(metrics["hotspot_level"][q])
+        status = "critical" if level == 2 else ("warning" if level == 1 else "ok")
+        rows.append(
+            [
+                q,
+                status,
+                round(risk, 6),
+                round(float(metrics["activity_count"][q]), 3),
+                round(float(metrics["gate_error"][q]), 6),
+                round(float(metrics["readout_error"][q]), 6),
+                round(float(metrics["decoherence_risk"][q]), 6),
+                round(float(metrics["fidelity"][q]), 6),
+            ]
+        )
+    rows.sort(key=lambda x: x[2], reverse=True)
+    k = max(1, min(int(top_k), len(rows)))
+    return rows[:k]
+
+
 # ---------- Styling ----------
 CSS = """
 #title h1 { font-size: 42px !important; margin-bottom: 6px; }
@@ -319,9 +371,36 @@ with gr.Blocks(theme=theme, css=CSS, title="Quread.ai — State Vector Studio")
                             label="Calibration JSON (optional)",
                             placeholder='{"qubits":{"0":{"gate_error":0.012,"readout_error":0.02,"t1_us":82,"t2_us":61,"fidelity":0.991}}}',
                         )
+                        gr.Markdown("<div class='small-note'>Composite risk weights are normalized automatically.</div>")
+                        with gr.Row():
+                            w_activity = gr.Slider(0.0, 1.0, value=0.25, step=0.01, label="Weight: activity")
+                            w_gate = gr.Slider(0.0, 1.0, value=0.20, step=0.01, label="Weight: gate error")
+                            w_readout = gr.Slider(0.0, 1.0, value=0.15, step=0.01, label="Weight: readout error")
+                        with gr.Row():
+                            w_decoherence = gr.Slider(0.0, 1.0, value=0.25, step=0.01, label="Weight: decoherence")
+                            w_fidelity = gr.Slider(0.0, 1.0, value=0.15, step=0.01, label="Weight: fidelity risk")
+                        with gr.Row():
+                            thr_warning = gr.Slider(0.0, 1.0, value=0.45, step=0.01, label="Threshold: warning")
+                            thr_critical = gr.Slider(0.0, 1.0, value=0.70, step=0.01, label="Threshold: critical")
+                            hotspot_top_k = gr.Slider(1, 64, value=16, step=1, label="Hotspot rows")
                         metrics_csv_dl = gr.DownloadButton("Download metrics CSV")
                         heat_btn = gr.Button("Generate heatmap from CSV", variant="secondary")
                         heat_plot = gr.Plot()
+                        hotspot_status = gr.Markdown()
+                        hotspot_table = gr.Dataframe(
+                            headers=[
+                                "qubit",
+                                "status",
+                                "composite_risk",
+                                "activity_count",
+                                "gate_error",
+                                "readout_error",
+                                "decoherence_risk",
+                                "fidelity",
+                            ],
+                            interactive=False,
+                            label="Hotspot ranking (highest composite risk first)",
+                        )
     
 
             with gr.Group(elem_classes=["card"]):
@@ -445,35 +524,126 @@ with gr.Blocks(theme=theme, css=CSS, title="Quread.ai — State Vector Studio")
         outputs=[llm_out, last_explained_hash, explanation_md],
     )
     
-    def _heat_from_current(qc, n_qubits, rows, cols, metric, calibration_text):
+    def _heat_and_hotspots_from_current(
+        qc,
+        n_qubits,
+        rows,
+        cols,
+        metric,
+        calibration_text,
+        activity_w,
+        gate_error_w,
+        readout_error_w,
+        decoherence_w,
+        fidelity_w,
+        warning_thr,
+        critical_thr,
+        top_k,
+    ):
         csv_text = to_csv(qc.history)  # must exist from Task 2A
         cfg = HeatmapConfig(rows=int(rows), cols=int(cols))
-        return make_metric_heatmap(
+        weights, thresholds = _metric_controls_to_models(
+            activity_w,
+            gate_error_w,
+            readout_error_w,
+            decoherence_w,
+            fidelity_w,
+            warning_thr,
+            critical_thr,
+        )
+        fig = make_metric_heatmap(
             csv_text=csv_text,
             n_qubits=int(n_qubits),
             metric=str(metric),
             cfg=cfg,
             calibration_json=str(calibration_text or ""),
+            weights=weights,
+            thresholds=thresholds,
         )
-
-    def _dl_metrics_csv(qc, n_qubits, calibration_text):
+        metrics, meta = compute_metrics_from_csv(
+            csv_text,
+            int(n_qubits),
+            calibration_json=str(calibration_text or ""),
+            weights=weights,
+            thresholds=thresholds,
+        )
+        hotspot_rows = _hotspot_rows(metrics, int(n_qubits), int(top_k))
+        note = []
+        skipped = int(meta.get("skipped_rows", 0))
+        if skipped:
+            note.append(f"Skipped malformed CSV rows: {skipped}")
+        calibration_note = str(meta.get("calibration_note", "") or "").strip()
+        if calibration_note:
+            note.append(calibration_note)
+        summary = " | ".join(note) if note else "Hotspot ranking updated."
+        return fig, summary, hotspot_rows
+
+    def _dl_metrics_csv(
+        qc,
+        n_qubits,
+        calibration_text,
+        activity_w,
+        gate_error_w,
+        readout_error_w,
+        decoherence_w,
+        fidelity_w,
+        warning_thr,
+        critical_thr,
+    ):
         csv_text = to_csv(qc.history)
+        weights, thresholds = _metric_controls_to_models(
+            activity_w,
+            gate_error_w,
+            readout_error_w,
+            decoherence_w,
+            fidelity_w,
+            warning_thr,
+            critical_thr,
+        )
         metrics, _meta = compute_metrics_from_csv(
             csv_text,
             int(n_qubits),
             calibration_json=str(calibration_text or ""),
+            weights=weights,
+            thresholds=thresholds,
         )
         return _write_tmp("qubit_metrics.csv", to_metrics_csv(metrics))
 
     heat_btn.click(
-        fn=_heat_from_current,
-        inputs=[qc_state, n_qubits, chip_rows, chip_cols, heat_metric, calibration_json],
-        outputs=[heat_plot],
+        fn=_heat_and_hotspots_from_current,
+        inputs=[
+            qc_state,
+            n_qubits,
+            chip_rows,
+            chip_cols,
+            heat_metric,
+            calibration_json,
+            w_activity,
+            w_gate,
+            w_readout,
+            w_decoherence,
+            w_fidelity,
+            thr_warning,
+            thr_critical,
+            hotspot_top_k,
+        ],
+        outputs=[heat_plot, hotspot_status, hotspot_table],
     )
 
     metrics_csv_dl.click(
         fn=_dl_metrics_csv,
-        inputs=[qc_state, n_qubits, calibration_json],
+        inputs=[
+            qc_state,
+            n_qubits,
+            calibration_json,
+            w_activity,
+            w_gate,
+            w_readout,
+            w_decoherence,
+            w_fidelity,
+            thr_warning,
+            thr_critical,
+        ],
         outputs=[metrics_csv_dl],
     )