Feature/integrated system by sxhaakee · Pull Request #35 · hacktofuture-sjec/hacktofuture4

sxhaakee · 2026-04-16T12:56:37Z

ML and Firebase

Updated project title and description in README.

Fix/dashboard setup

feat: integrate Stadia Maps into dashboard

…dashboard

Copilot

Pull request overview

This PR integrates an end-to-end “sensor → ML validation/triage → pipeline orchestration → live dashboard + Firebase bridge” workflow, combining multiple ML artifacts, a LangGraph multi-agent pipeline, and a Next.js dashboard for live telemetry visualization.

Changes:

Added training + prediction scripts and committed generated model/data artifacts for seismic/gas/survivor/validator models.
Implemented a LangGraph pipeline (verifier/triage/logistics/reporter) plus a Firebase polling bridge that runs the pipeline and writes results back.
Added a Next.js dashboard (UI + /api/telemetry) and Python tests covering agents/pipeline behavior and artifact presence.

Reviewed changes

Copilot reviewed 42 out of 61 changed files in this pull request and generated 11 comments.

Show a summary per file

File	Description
train/train_validator.py	Trains/saves IsolationForest “validator” model artifact.
train/train_survivor.py	Trains/saves survivor LogisticRegression + scaler artifacts.
train/train_seismic.py	Trains/saves TensorFlow 1D CNN seismic classifier.
train/train_gas.py	Trains/saves RandomForest gas classifier.
tests/test_validation_and_agents.py	Unit tests for verifier/triage helpers and agent outputs.
tests/test_pipeline_and_dashboard.py	Tests pipeline routing, logistics helpers, and expected artifacts/files.
predict/predict_validator.py	Loads validator model and predicts genuineness/anomaly score.
predict/predict_survivor.py	Loads survivor model/scaler and predicts survivor probability + urgency.
predict/predict_seismic.py	Loads seismic CNN model and predicts event type + confidence.
predict/predict_gas.py	Loads gas model and predicts hazard type + severity.
predict/pycache/predict_validator.cpython-313.pyc	Committed Python bytecode artifact (build output).
predict/pycache/predict_survivor.cpython-313.pyc	Committed Python bytecode artifact (build output).
predict/pycache/predict_seismic.cpython-313.pyc	Committed Python bytecode artifact (build output).
predict/pycache/predict_gas.cpython-313.pyc	Committed Python bytecode artifact (build output).
pipeline/state.py	Defines TypedDict state contract for LangGraph pipeline.
pipeline/langgraph_pipeline.py	Standalone “run sub-models + Mistral SITREP” pipeline script.
pipeline/graph.py	LangGraph graph assembly + pipeline runner and routing logic.
pipeline/firebase_live_bridge.py	Polls Firebase RTDB, posts dashboard telemetry, runs pipeline, writes inference back.
models/survivor_scaler.pkl	Committed scaler artifact used for survivor inference.
models/survivor_model.pkl	Committed survivor model artifact used for survivor inference.
data/survivor_data.csv	Committed training data for survivor model.
data/generate_survivor_data.py	Generator script for survivor synthetic dataset.
data/generate_seismic_data.py	Generator script for seismic synthetic dataset.
data/generate_gas_data.py	Generator script for gas synthetic dataset.
dashboard/tsconfig.json	Dashboard TypeScript configuration (strict mode, paths).
dashboard/src/components/SystemHealth.tsx	Top bar health/status component.
dashboard/src/components/SitrepPanel.tsx	UI panel to display SITREP details and checklists.
dashboard/src/components/SimulationDashboard.tsx	Scenario-based simulation UI for pipeline visualization.
dashboard/src/components/MapPanel.tsx	Map placeholder panel + active alert marker rendering.
dashboard/src/components/AlertFeed.tsx	Right-side alert feed UI.
dashboard/src/app/page.tsx	Main dashboard page with polling + panels layout.
dashboard/src/app/layout.tsx	Root layout + fonts.
dashboard/src/app/globals.css	Global Tailwind theme setup.
dashboard/src/app/favicon.ico	Dashboard favicon asset.
dashboard/src/app/api/telemetry/route.ts	Next.js API route for ingesting/serving telemetry snapshots.
dashboard/public/window.svg	Static asset.
dashboard/public/vercel.svg	Static asset.
dashboard/public/next.svg	Static asset.
dashboard/public/globe.svg	Static asset.
dashboard/public/file.svg	Static asset.
dashboard/postcss.config.mjs	Tailwind PostCSS config.
dashboard/package.json	Dashboard dependencies + scripts.
dashboard/next.config.ts	Next.js config (turbopack root).
dashboard/eslint.config.mjs	ESLint config for the dashboard.
dashboard/REALTIME_IOT_SETUP.md	Setup guide for ESP32 → dashboard telemetry ingestion.
dashboard/README.md	Default Next.js README for dashboard subproject.
dashboard/CLAUDE.md	Agent tooling pointer doc.
dashboard/AGENTS.md	Agent rules documentation.
dashboard/.gitignore	Dashboard-specific ignore rules.
agents/verifier_agent.py	Verifier agent: IsolationForest + correlation logic, aborts false alarms.
agents/triage_agent.py	Triage agent: survivability, gas threat, urgency, equipment checklist.
agents/reporter_agent.py	Reporter agent: compact prompt + optional Mistral SITREP generation with fallback.
agents/logistics_agent.py	Logistics agent: debris model + entry points + zones + ETA estimate.
README.md	Project-level README updated to describe NeuroMesh and setup.
.gitignore	Repository-level ignore rules (Python/Node artifacts).

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-16T13:01:34Z

+import { getLiveSnapshot, ingestTelemetry, isValidIncomingTelemetry } from '@/lib/telemetryStore';
+


This API route imports '@/lib/telemetryStore', but there is no corresponding file in the dashboard source tree. This will fail Next.js build/runtime with a module resolution error. Add the missing telemetryStore module (e.g., dashboard/src/lib/telemetryStore.ts) or update the import to the correct existing location.

Suggested change

import { getLiveSnapshot, ingestTelemetry, isValidIncomingTelemetry } from '@/lib/telemetryStore';

type IncomingTelemetry = {

node_id: string;

lat: number;

lng: number;

acceleration_g: number;

gas_raw: number;

motion: boolean;

};

type TelemetrySnapshot = IncomingTelemetry & {

received_at: string;

};

let liveSnapshot: TelemetrySnapshot | null = null;

function isRecord(value: unknown): value is Record<string, unknown> {

return typeof value === 'object' && value !== null;

}

function isValidIncomingTelemetry(payload: unknown): payload is IncomingTelemetry {

if (!isRecord(payload)) {

return false;

}

return (

typeof payload.node_id === 'string' &&

typeof payload.lat === 'number' &&

Number.isFinite(payload.lat) &&

typeof payload.lng === 'number' &&

Number.isFinite(payload.lng) &&

typeof payload.acceleration_g === 'number' &&

Number.isFinite(payload.acceleration_g) &&

typeof payload.gas_raw === 'number' &&

Number.isFinite(payload.gas_raw) &&

typeof payload.motion === 'boolean'

);

}

function ingestTelemetry(payload: IncomingTelemetry): TelemetrySnapshot {

liveSnapshot = {

...payload,

received_at: new Date().toISOString(),

};

return liveSnapshot;

}

function getLiveSnapshot(): TelemetrySnapshot | null {

return liveSnapshot;

}

Copilot · 2026-04-16T13:01:35Z

+        time_window = 60
+        team_size = 6
+    elif survivability_score >= 30:
+        urgency = "moderate"


determine_urgency can return urgency='moderate', but the rest of the codebase/types/documentation (e.g., pipeline/state.py) only mention low/high/immediate/extreme. This creates inconsistent downstream handling. Either map this branch to an existing level (e.g., 'high' or 'low') or update the allowed urgency set everywhere it’s referenced.

Suggested change

urgency = "moderate"

urgency = "high"

Copilot · 2026-04-16T13:01:35Z

+import numpy as np
+import pandas as pd
+from sklearn.ensemble import IsolationForest


Unused import: pandas is not used anywhere in this script. Please remove it to avoid unnecessary dependency/runtime overhead and keep linting clean.

Copilot · 2026-04-16T13:01:36Z

+class TriageOutput(TypedDict):
+    survivability_score: float      # 0-100
+    estimated_persons: str          # "0" | "1-2" | "3-5" | "5+"
+    life_sign_pattern: str          # "active" | "weakening" | "critical" | "none"
+    gas_threat: str                 # "clear" | "warning" | "lethal"
+    entry_protocol: str             # "standard" | "breathing_apparatus" | "hazmat"
+    urgency: str                    # "low" | "high" | "immediate" | "extreme"
+


TriageOutput is missing fields that triage_agent actually emits (time_sensitivity_minutes, recommended_team_size, golden_hour_remaining, equipment_checklist, gas_note), and it documents urgency values that don't include "moderate" (which determine_urgency can return). Please expand/adjust the TypedDict so it matches the real output contract (or change triage_agent to conform).

Copilot · 2026-04-16T13:01:36Z

+class LogisticsOutput(TypedDict):
+    primary_route: dict             # GeoJSON LineString
+    alternate_route: dict           # GeoJSON LineString
+    blocked_roads: List[str]
+    estimated_eta_minutes: float
+    entry_point: dict               # GeoJSON Point
+    assembly_point: dict            # GeoJSON Point
+    debris_risk_zones: List[dict]   # GeoJSON Polygons
+


LogisticsOutput is missing keys that logistics_agent returns (e.g., exclusion_radius_m) and its comments describe GeoJSON types that don't match the actual shape (route dicts with labels/entry_point). Please update the TypedDict to match the emitted structure, or adjust logistics_agent to produce the documented GeoJSON shapes.

Copilot · 2026-04-16T13:01:36Z

+class VerifierOutput(TypedDict):
+    is_genuine: bool
+    confidence: float
+    triggered_nodes: List[str]
+    correlation_type: str   # "single_node" | "cluster" | "mesh_wide"
+    rejection_reason: Optional[str]


TypedDict definitions are out of sync with actual values produced by the agents. For example, VerifierOutput.correlation_type is documented as "single_node"|"cluster"|"mesh_wide", but verifier_agent can return "none" and "pair"; this will break static type checking and can mislead downstream consumers. Update the allowed values (prefer Literal types) or align compute_spatial_correlation outputs to the documented set.

Copilot · 2026-04-16T13:01:37Z

+    def test_model_artifacts_exist(self):
+        expected = [
+            "models/seismic_model.keras",
+            "models/gas_model.pkl",
+            "models/survivor_model.pkl",
+            "models/survivor_scaler.pkl",
+            "models/validator_model.pkl",
+        ]
+        missing = [p for p in expected if not os.path.exists(p)]
+        self.assertEqual(missing, [])


This test asserts model artifacts exist using relative paths (e.g., "models/..."), which makes it dependent on the current working directory the test runner uses. To make it robust, resolve paths relative to the repository root or the test file location (e.g., Path(file).resolve().parents[...] / 'models' / ...).

Copilot · 2026-04-16T13:01:37Z

+from sklearn.model_selection import train_test_split
+from sklearn.preprocessing import LabelEncoder
+import tensorflow as tf
+from tensorflow import keras


Unused imports: LabelEncoder and tensorflow (tf) are not used in this script (keras is used via from tensorflow import keras). Please remove unused imports to keep dependencies clear and avoid failing lint/type checks.

Copilot · 2026-04-16T13:01:37Z

+ACC_THRESHOLD = 0.25
+GAS_THRESHOLD = 2000
+EARTHQUAKE_DURATION_MS = 10_000


GAS_THRESHOLD is defined but never used in this module. If it’s meant to drive quake/gas alert logic, wire it into the relevant checks; otherwise remove it to avoid dead configuration that can drift from reality.

Copilot · 2026-04-16T13:01:38Z

+# Setup Mistral
+client = MistralClient(api_key="H2z0gd6ieaMgUAAByONNnDtnmwLtonuW")
+


Hard-coded Mistral API key is committed in source. This is a credential leak and will also make local/CI behavior depend on a single key. Read the key from an environment variable (e.g., MISTRAL_API_KEY) and fail/disable LLM features when it’s missing; also rotate/revoke the leaked key.

Co-authored-by: sxhaakee <149256691+sxhaakee@users.noreply.github.com>

haashid and others added 7 commits April 15, 2026 19:32

Change project title to 'ABRA CODE ABRA'

8ad0ffc

Updated project title and description in README.

Initialize project and fix Next.js dashboard

2ed7795

docs: Beautify README following HackToFuture template

6f3d762

Merge pull request #1 from haashid/fix/dashboard-setup

61b562f

Fix/dashboard setup

feat: integrate Stadia Maps into dashboard

2c3c167

Merge pull request #2 from haashid/feat/stadia-maps

2ab97cf

feat: integrate Stadia Maps into dashboard

feat: integrate multi-agent system, telemetry pipeline, and realtime …

ccbcd44

…dashboard

Copilot AI review requested due to automatic review settings April 16, 2026 12:56

Copilot AI reviewed Apr 16, 2026

View reviewed changes

Merge origin/main into feature/integrated-system to resolve conflicts

f2aa091

Co-authored-by: sxhaakee <149256691+sxhaakee@users.noreply.github.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feature/integrated system#35

Feature/integrated system#35
sxhaakee wants to merge 8 commits intohacktofuture-sjec:mainfrom
haashid:feature/integrated-system

sxhaakee commented Apr 16, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Copilot AI Apr 16, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		import { getLiveSnapshot, ingestTelemetry, isValidIncomingTelemetry } from '@/lib/telemetryStore';

-import { getLiveSnapshot, ingestTelemetry, isValidIncomingTelemetry } from '@/lib/telemetryStore';
+type IncomingTelemetry = {
+  node_id: string;
+  lat: number;
+  lng: number;
+  acceleration_g: number;
+  gas_raw: number;
+  motion: boolean;
+};
+type TelemetrySnapshot = IncomingTelemetry & {
+  received_at: string;
+};
+let liveSnapshot: TelemetrySnapshot | null = null;
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === 'object' && value !== null;
+}
+function isValidIncomingTelemetry(payload: unknown): payload is IncomingTelemetry {
+  if (!isRecord(payload)) {
+    return false;
+  }
+  return (
+    typeof payload.node_id === 'string' &&
+    typeof payload.lat === 'number' &&
+    Number.isFinite(payload.lat) &&
+    typeof payload.lng === 'number' &&
+    Number.isFinite(payload.lng) &&
+    typeof payload.acceleration_g === 'number' &&
+    Number.isFinite(payload.acceleration_g) &&
+    typeof payload.gas_raw === 'number' &&
+    Number.isFinite(payload.gas_raw) &&
+    typeof payload.motion === 'boolean'
+  );
+}
+function ingestTelemetry(payload: IncomingTelemetry): TelemetrySnapshot {
+  liveSnapshot = {
+    ...payload,
+    received_at: new Date().toISOString(),
+  };
+  return liveSnapshot;
+}
+function getLiveSnapshot(): TelemetrySnapshot | null {
+  return liveSnapshot;
+}

		# Setup Mistral
		client = MistralClient(api_key="H2z0gd6ieaMgUAAByONNnDtnmwLtonuW")

Conversation

sxhaakee commented Apr 16, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 16, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants