REI SignalLab 2.1

Extended Typed DSP Node Library, Kahn Execution Engine, Numerical Golden Precision Suite & REI Vibration Analysis Workbench

Engineered by Batuhan Ayribas at RootCastle — Engineering Beyond Boundaries

Live Demo | 16-Page PDF User Manual | GitHub Wiki | Overview | Security | Testing | Screenshots | Vibration Workbench | Canonical Node Library | Numerical Golden Precision | API Spec | Quick Start

---

🏢 About

REI SignalLab 2.1 is developed and maintained by Batuhan Ayribas, founder of RootCastle.

Developer	Batuhan Ayribas
Company	RootCastle — Engineering Beyond Boundaries
Live Application	signallab.site
Firebase Mirrors	signallab-3305b.web.app · signallab-3305b.firebaseapp.com
Source Repository	github.com/rootcastleco/rei-signallab
GitHub Wiki	github.com/rootcastleco/rei-signallab/wiki
Version	2.1.0
License	MIT

---

Overview

REI SignalLab 2.1 is an instrument-grade digital signal processing, spectral analysis, visual node graph flow studio, and industrial vibration analysis workbench. Developed by Batuhan Ayribas at RootCastle, it provides:

1. Canonical Typed Node Library (.rei-signal 2.1): Versioned node registry (GET /api/nodes) exposing canonical DSP and Vibration nodes with strictly enforced port data types (Signal<Real64>, Signal<Complex128>, SpectrumFrame, Scalar<Real64>, PatternEvent). Includes a multi-column responsive grid palette and full 📚 Canonical Node Library Explorer Modal.
2. Kahn Topological Execution Engine (POST /api/graph/execute): Deterministic graph validation and execution engine with 10-point port compatibility checks, cycle detection, and Kahn's topological scheduler.
3. 🤖 OpenRouter AI Copilot & LLM Assistant: Real-time scientific AI telemetry reasoning and DSP analysis supporting 5 free LLM models (google/gemini-2.0-flash-lite-preview-02-05:free, meta-llama/llama-3.3-70b-instruct:free, deepseek/deepseek-r1:free, qwen/qwen-2.5-coder-32b-instruct:free, mistralai/mistral-7b-instruct:free) with ⚙️ AI Settings modal, obfuscated system key, and Smart Local In-Browser Fallback Engine.
4. 👤 Phase 4 Firebase Authentication & Cloud Project Sync: Multi-provider user authentication (Email/Password, Google SSO with redirect fallback, Guest Mode) and real-time Cloud Firestore project persistence (users/{userId}/projects) with LocalStorage offline synchronization.
5. REI Vibration Analysis Workbench: End-to-end industrial machinery condition monitoring, single-plane complex vector rotor balancing, kinematic bearing defect frequency tracking (BPFO, BPFI, BSF, FTF), Hilbert envelope demodulation, 1X-10X harmonic order spectrum bar view, and rule-based fault classification.
6. Numerical Golden Verification Suite: Comprehensive Pytest suite enforcing IEEE double-precision float64 error $\le 10^{-12}$, FFT/IFFT reconstruction RMS error $\le 10^{-9}$, DCT Type II error $\le 10^{-12}$, and Haar wavelet error $\le 10^{-12}$.
7. Layered Sandbox & Resource Governance: Static AST policy, out-of-process script execution under OS resource limits, enforced graph size ceilings, and per-caller rate limiting on every route — see Security & Hardening.

---

🔒 Security & Hardening

Every endpoint is unauthenticated, so all input is treated as hostile. The full threat model lives in SECURITY.md.

Python scripting sandbox

POST /api/python/execute executes caller-supplied Python and is the highest-risk surface in the project. It is disabled by default whenever APP_ENV=production and must be opted into explicitly with ENABLE_PYTHON_SANDBOX=true.

Defence is layered — no submitted code ever runs inside the API worker:

Layer	Implementation
Static AST policy	backend/app/sandbox/guard.py — import allowlist, rejection of all dunder access (the entry point for ().__class__.__base__.__subclasses__() traversal), and of the reflection / code-loading builtins (getattr, eval, compile, open, globals, …)
Process isolation	backend/app/sandbox/runner.py — child interpreter under RLIMIT_AS and RLIMIT_CPU, parent-enforced wall-clock timeout, socket creation disabled, restricted builtins map
Kill switch	ENABLE_PYTHON_SANDBOX, off by default in production

Host data is injected into the script namespace as values, never formatted into the source, so a large signal vector cannot inflate the script past the size cap or influence how it parses.

28 known escape classes are pinned as regression tests in backend/tests/test_security_golden.py.

This is a hardened interpreter, not an OS-level jail. Before enabling it on a public deployment, isolate it as described under Python Scripting Sandbox in docs/DEPLOYMENT.md.

Resource governance

Control	Default	Enforced in
MAX_UPLOAD_BYTES	25 MB	Request middleware + upload routes
MAX_SIGNAL_SAMPLES	2,000,000	/api/process, /api/upload/signal
MAX_GRAPH_NODES	200	Graph validator, before any per-node work
MAX_GRAPH_CONNECTIONS	500	Graph validator
MAX_SANDBOX_NODES_PER_GRAPH	4	Graph validator — caps interpreter fan-out from a single request
PYTHON_SANDBOX_TIMEOUT_SECONDS	8	Parent process kill deadline
PYTHON_SANDBOX_MEMORY_MB	512	RLIMIT_AS on the sandbox child

Rate limiting

A dependency-free fixed-window limiter (backend/app/ratelimit.py) keyed on client IP, returning HTTP 429 with a Retry-After header:

Bucket	Default	Scope
global	240 / min	Every request
compute	60 / min	DSP, graph execution, Lisp, WAV export, uploads
sandbox	10 / min	/api/python/execute

Limits are enforced per Cloud Run instance. This protects a single instance from CPU saturation; it is not a distributed quota.

---

🧪 Testing & Quality Gates

Suite	Count	Scope
Numerical golden	—	Float64 / FFT / DCT / Haar precision against NumPy & SciPy references
Security golden	—	Sandbox escape classes, graph limits, rate limiter, request validation
Route smoke	36	Drives every endpoint over HTTP, plus a guard that fails when a registered route has no coverage
Docs contract	—	Asserts documented endpoints exist and node-count claims match the registry
Backend total	154	pytest
Frontend	17	vitest — API error envelope, non-JSON rejection, timeout mapping, backend handshake states

The route smoke suite exists because the golden suites exercise engines directly: a route calling an engine method that had lost its def line returned HTTP 500 in production while the entire suite passed.

CI gates (.github/workflows/ci.yml) run on every pull request:

• Backend pytest with the sandbox enabled, so the boundary is exercised rather than skipped
• Frontend ESLint and Vitest, then the Vite production build
• pip-audit on production Python dependencies and npm audit --omit=dev
• Docker container build and readiness-probe startup test
• Dependabot (.github/dependabot.yml) tracks pip, npm and GitHub Actions updates

# Backend
pip install -r backend/requirements-dev.txt
$env:PYTHONPATH="backend"; py -m pytest backend/tests

# Frontend
cd frontend && npm run lint && npm test

---

🚀 Production Deployment & Cloud Run Integration

Firebase Hosting serves the static frontend and proxies all /api/** calls to the Cloud Run FastAPI container. Deployment is fully automated by .github/workflows/deploy.yml, which builds, pushes, deploys, smoke-tests the live service, and auto-rolls-back to the previous revision if any check fails.

For architecture, health probes, and rollback procedures, see docs/DEPLOYMENT.md.

First-time GCP setup

The pipeline authenticates via Workload Identity Federation — no long-lived service account keys. scripts/setup-github-wif.sh provisions the deployer service account, its roles, and a WIF pool and OIDC provider pinned to this repository, then prints the three required GitHub secrets:

gcloud auth login
./scripts/setup-github-wif.sh

The script is idempotent and safe to re-run.

Manual deployment

# 1. Authenticate with Google Cloud
gcloud auth login
gcloud config set project signallab-3305b

# 2. Execute Cloud Run Deployment Script
GCP_PROJECT_ID=signallab-3305b GCP_REGION=europe-west1 ./scripts/deploy-cloud-run.sh

# 3. Deploy Frontend Assets & API Rewrites to Firebase Hosting
npx firebase deploy --only hosting

Firebase Hosting must be deployed after the Cloud Run service exists — a run rewrite cannot resolve to a service that is absent.

Health Verification Commands

curl -fsS https://signallab.site/api/health/live
curl -fsS https://signallab.site/api/health/ready
curl -fsS https://signallab.site/api/version

---

Application Screenshots

1. REI SignalLab 2.1 Workspace & GPS SDR Simulator

2. Dual CRT Oscilloscope & Spectrum Analyzer Workspace

3. Signal Flow Studio Visual Canvas & Typed Node Catalog

---

🛰️ GPS SDR Signal Simulator Workbench (gps-sdr-sim)

The GPS L1 C/A SDR Signal Simulator Workbench is modeled after Takuji Ebinuma's open-source gps-sdr-sim. It provides full constellation orbit propagation, baseband I/Q synthesis, and downloadable binary signal streams for software-defined radios.

• 1023-Chip C/A Gold Code Generator: Generates exact 1023-chip PRN 1–32 Gold Code sequences at $1.023\text{ MHz}$ using G1 ($x^{10} + x^3 + 1$) and G2 LFSR polynomials.
• Orbit Kinematics & Coordinate Transformations: Computes WGS84 Geodetic $(Lat, Lon, Alt) \to$ ECEF $(X, Y, Z)$ and ENU vector coordinates for 32 MEO satellites.
• Doppler Shift & C/N0 Estimation: Calculates line-of-sight range, propagation delay, Doppler frequency shift ($\pm 5\text{ kHz}$), and Carrier-to-Noise Ratio ($C/N_0 = 35\text{--}50\text{ dB-Hz}$).
• Dilution of Precision (DOP): Calculates GDOP, PDOP, HDOP, and VDOP from the satellite observation geometry matrix $G$.
• Polar Skyplot & Spectrum Visualizations: Real-time canvas rendering of polar satellite skyplots ($0^\circ\text{--}360^\circ$ azimuth, $0^\circ\text{--}90^\circ$ elevation), baseband spectrum, I/Q scatter plot, and Gold Code auto-correlation peaks.
• Multi-SDR Binary Export: Export raw binary baseband signal streams (.bin) formatted for HackRF, LimeSDR (8-bit signed int8), USRP, BladeRF (16-bit signed int16), and RTL-SDR (8-bit unsigned uint8).

---

☀️ SRW Synchrotron & Undulator Radiation Workbench

Modeled after Oleg Chubar's SRW (Synchrotron Radiation Workshop), this module provides relativistic electron beam kinematics and undulator radiation wave optics calculations backed by FastAPI endpoints (/api/srw/*):

• Relativistic Electron Beam Dynamics: Calculates Lorentz factor $\gamma = E_e / (m_e c^2)$ and electron energy spread for 0.1–15 GeV storage rings.
• Undulator Deflection Parameter ($K$): Computes $K = \frac{e B_0 \lambda_u}{2\pi m_e c} \approx 0.9337 \cdot B_0 [\text{T}] \cdot \lambda_u [\text{cm}]$.
• Harmonic Photon Energies ($E_n$): Computes fundamental $E_1 = \frac{949.63 E_e^2}{\lambda_u (1 + K^2/2)}$ and odd harmonic energies $E_3, E_5, E_7$ (eV).
• Total Radiated Power ($P_{\text{rad}}$): Computes undulator total power $P_{\text{rad}} [\text{kW}] = 0.6331 E_e^2 B_0^2 L_u I_e$.
• Spectral Flux & Wavefront Intensity: Visualizes spectral flux distribution $F(E)$, 2D transverse wavefront intensity heatmap $I(x,y)$ on observation screens, and angular power density $d^2P/d\Omega$.

---

⚙️ REI Vibration Analysis Workbench & Condition Monitoring Suite

The REI Vibration Analysis Workbench is an end-to-end industrial condition monitoring suite incorporating comprehensive diagnostics tools:

• 13 Standard Industrial Vibration Units: Supports simultaneous conversion across Acceleration ($g$ RMS/Peak, $\text{in/s}^2$, $\text{mm/s}^2$), Velocity ($\text{mm/s}$, $\text{in/s}$), and Displacement ($\text{mils}$ pk-pk, $\text{mm}$ pk-pk, $\mu\text{m}$ pk-pk).
• 3,570+ Kinematic Bearing Catalog: Integrated database of over 3,570 rolling element bearings across SKF, NTN, Cooper, and Dodge with an instant 🔍 Search button. Computes exact BPFO, BPFI, BSF, and FTF fault frequencies.
• Interactive Shaft Orbit Plot Simulator:
  • Superimposes up to 3 frequency components (F1 1X, F2 2X/Misalignment, F3 Sub-Synchronous Whirl).
  • Configurable X/Y probe amplitudes, phase angles, and 5 probe pair orientations (e.g., $45^\circ\text{R} \ & \ 45^\circ\text{L}$).
  • Keyphasor TDC timing mark placement once per 1X revolution and Clockwise / Counter-Clockwise (CW/CCW) rotation direction.
• Visual Signal Tone Generator: Dual sine wave acoustic tone generator demonstrating beating effects ($\Delta f = |F_1 - F_2|$) with real-time waveform and discrete peak spectrum visualization.
• Rotor Balancing & Fault Classification: Single-plane complex vector balancing ($\vec{V}0, \vec{V}1, \vec{W}{\text{trial}} \to \vec{W}{\text{correction}}$) and rule-based diagnostic fault classifier.

---

📘 Digital Signal Processing Lab (DSP Lab)

This module provides interactive educational and analytical tools backed by SciPy API endpoints (/api/dsp-lab/*):

• Sampling Theorem & Aliasing Fold-Over: Visualizes continuous signal $s(t) = \sin(2\pi f t)$ vs discrete samples $s[n]$, highlights Nyquist frequency $f_{\text{nyquist}} = f_s / 2$, and calculates folded aliased frequency $f_{\text{aliased}} = |f_{\text{signal}} - k \cdot f_{\text{sample}}|$.
• Parks-McClellan Optimal Equiripple FIR Filter Design: Uses scipy.signal.remez to compute optimal Type I FIR impulse response $h[n]$ and magnitude response $|H(e^{j\omega})|$ dB with exact passband ripple and stopband attenuation metrics.
• Autocorrelation & Pitch Detector ($R_{xx}[\tau]$): Computes normalized autocorrelation sequence $R_{xx}[\tau] = \frac{\sum x[n]x[n+\tau]}{\sum x^2[n]}$ and extracts dominant fundamental pitch frequency $f_0$.
• Least Mean Squares (LMS) Adaptive Noise Canceller: Implements LMS adaptive filter iteration ($y[n] = \mathbf{w}^T \mathbf{x}[n]$, $e[n] = d[n] - y[n]$, $\mathbf{w}[n+1] = \mathbf{w}[n] + 2\mu e[n] \mathbf{x}[n]$) demonstrating real-time interference rejection and SNR improvement.
• Continuous Wavelet Transform (CWT) Time-Frequency Scalogram: Computes 2D wavelet scalogram matrix $|W(a,b)|$ using Morlet/Ricker wavelets for chirp time-frequency analysis.

---

⚡ Electrical & 📡 Antenna RF Workbenches

• ⚡ Electrical Workbench: 3-phase symmetrical component decomposition ($V_0, V_1, V_2$) via Fortescue matrix transform, real/reactive/apparent power ($P, Q, S$), power factor ($\cos\phi$), and THD.
• 📡 Antenna & RF Workbench: VSWR & Return Loss ($S_{11}$), Friis Free Space Path Loss (FSPL), and Rectangular Waveguide Cutoff Frequency ($TE_{10}$).

---

📊 MATLAB Add-On & Client Toolbox Integration

REI SignalLab 2.1 includes native MATLAB (R2018b – R2026a+) integration via an object-oriented client class and 1-click installer:

• REISignalLab.m Client Class: Connects MATLAB directly to https://signallab.site or local API engines using MATLAB's webwrite and webread.
• install_rei_signallab.m 1-Click Installer: Adds the toolbox to the MATLAB search path (addpath) and runs connection self-tests.
• matlab/examples/matlab_demo.m: Complete demonstration script plotting oscilloscope waveforms, FFT magnitude spectrum, single-plane rotor balancing, SKF/NTN/Cooper/Dodge bearing queries, 3-phase Fortescue power quality, and OpenRouter AI queries directly inside MATLAB figure windows.
• Node Graph to .m Exporter (POST /api/matlab/export-m): Converts visual node graph projects into executable MATLAB .m scripts.

% Quick Start in MATLAB
install_rei_signallab

lab = REISignalLab('https://signallab.site');
sig = lab.processSignal('sine', 440, 1.0, 44100, 0.1, 'lowpass', 1000);
plot(sig.time, sig.signal); title('REI SignalLab Signal in MATLAB');

---

🎛️ Canonical DSP Node Library

The registry holds 69 canonical nodes. The table below is a representative selection — query GET /api/nodes for the authoritative, versioned list including port signatures and parameter schemas.

Category	Canonical Type	Legacy Alias	Description
Analysis	analysis.noise_stats	SLNoiseStats	Coherent gain corrected THD, SNR, SINAD, SFDR, ENOB
Analysis	analysis.pattern_detector	SLPatternDetector	Normalized cross-correlation pattern detector
Analysis	analysis.min_max	SLMinMax	Min, max, and peak-to-peak amplitude calculator
Arithmetic	arithmetic.add	SLAdd	Sample-by-sample signal addition ($x + y$)
Arithmetic	arithmetic.subtract	SLSubtract	Sample-by-sample signal subtraction ($x - y$)
Arithmetic	arithmetic.multiply	SLMultiply	Sample-by-sample signal multiplication ($x \cdot y$)
Arithmetic	arithmetic.divide	SLDivide	Zero-division clamped signal division ($x / y$)
Arithmetic	arithmetic.apply_window	SLApplyWindow	Hann, Hamming, Blackman windowing function
Converters	converter.complex_to_real	SLComplexToReal	Complex signal splitter (Re, Im, Mag, Phase)
Converters	converter.real_to_complex	SLRealToComplex	Real Re/Im to complex signal converter
Converters	transform.hilbert	SLHilbert	Hilbert transform envelope follower
Converters	transform.power_spectrum	SLPowerSpectrum	Power Spectral Density (PSD / Welch)
Filters	filter.lowpass	SLLowPass	LowPass IIR/FIR filter
Filters	filter.highpass	SLHighPass	HighPass Butterworth IIR filter
Filters	filter.biquad_iir	SLBiQuadIir	Pole-stability checked Biquad SOS filter
Filters	filter.median	SLMedian	Spike impulse noise removal median filter
Generic	generic.real_value_filter	SLGenericRealValue	AST-sandboxed custom math expression filter
Generators	generator.signal	SLSignalGen	Sine, Square, Triangle, Sawtooth, Noise, ECG synthesizer
Generators	generator.gaussian_noise	SLGaussGen	Deterministic seeded Gaussian white noise generator
Meters	meter.rms	SLRMSMeter	True RMS voltage meter & sliding envelope
Transforms	transform.fft	SLFourier	Fast Fourier Transform (FFT)
Transforms	transform.inverse_real_fft	SLInverseFourier	Inverse Real FFT time signal reconstructor
Transforms	transform.dct	SLDct	Discrete Cosine Transform (DCT Type II)
Transforms	transform.haar	SLHaar	Haar Discrete Wavelet Transform (DWT)
Vibration	vibration.sensor_calibration	SLVibCalibrate	IEPE & MEMS sensor calibration ($\text{mV/g}$)
Vibration	vibration.acceleration_to_velocity	SLAccToVel	High-pass regularized velocity integration ($\text{mm/s}$)
Vibration	vibration.velocity_to_displacement	SLVelToDisp	Regularized displacement integration ($\mu\text{m}$)
Vibration	vibration.bearing_frequencies	SLBearingFreqs	Kinematic BPFO, BPFI, BSF, FTF bearing frequency calculator
Vibration	vibration.envelope_analysis	SLEnvelopeAnalysis	Bandpass + Hilbert envelope spectrum for bearing faults
Vibration	vibration.balance.single_plane	SLSinglePlaneBalance	Single-plane complex vector rotor balancing
Vibration	vibration.fault_classifier	SLFaultClassifier	Rule-based machine fault diagnostic classifier

---

🎯 Numerical Golden Precision Suite

All mathematical nodes and algorithms are verified against NumPy/SciPy reference standards in backend/tests/test_golden.py and backend/tests/test_vibration_golden.py:

• Arithmetic Float64 Error: $|x_{\text{computed}} - x_{\text{reference}}| \le 1.0 \times 10^{-12}$ (Pass)
• FFT / IFFT Reconstruction RMS Error: $RMS \le 1.0 \times 10^{-9}$ (Pass)
• DCT Type II Reconstruction Error: $RMS \le 1.0 \times 10^{-12}$ (Pass)
• Haar Wavelet Reconstruction Error: $RMS \le 1.0 \times 10^{-12}$ (Pass)
• Sensor Calibration Precision: Relative error $\le 0.1%$ (Pass)
• Velocity Integration Precision: Relative error $\le 1.0%$ (Pass)
• Bearing Fault Frequency Precision: Relative error $\le 0.1%$ (Pass)
• Single-Plane Rotor Balancing Precision: Mass error $\le 1.0%$, Angle error $\le 1.0^\circ$ (Pass)

Domain-specific golden suites cover the Antenna, Electrical, GPS, SRW, DSP Lab and node-registry surfaces, alongside the security, route-smoke and docs-contract suites described in Testing & Quality Gates.

---

API Endpoints Summary

The interactive OpenAPI schema is served at /docs (Swagger UI) and /redoc.

Core & health

Method	Endpoint	Description
GET	/	Service manifest & registered node count
GET	/api/health/live	Liveness probe
GET	/api/health/ready	Readiness probe — validates registry and engine modules
GET	/api/version	Build manifest (version, API version, commit SHA)

Node registry & graph engine

Method	Endpoint	Description
GET	/api/nodes	Canonical node spec registry array
GET	/api/nodes/{node_type}	Specific canonical node spec
POST	/api/graph/validate	10-point Graph Validation (ValidationResult)
POST	/api/graph/execute	Kahn Topological Graph Execution Engine (2.1.0)

Signal processing

Method	Endpoint	Description
POST	/api/process	Processing pipeline (Time, FFT, Metrics, Spectrogram)
POST	/api/upload/signal	Upload .wav, .csv, .txt, .json signal files
POST	/api/lisp/process	S-Expression DSP DSL kernel execution
POST	/api/export/wav	Downloadable 16-bit PCM WAV audio generator
POST	/api/python/execute	Python DSP sandbox — returns 403 unless ENABLE_PYTHON_SANDBOX=true

Vibration analysis

Method	Endpoint	Description
POST	/api/vibration/analyze	Full condition-monitoring analysis (metrics, FFT, envelope, diagnostics)
POST	/api/vibration/upload	Analyse an uploaded vibration signal file
POST	/api/vibration/balance	Single-plane complex vector rotor balancing
POST	/api/vibration/balance/two-plane	Two-plane influence coefficient balancing
POST	/api/vibration/balance/four-run-nophase	Phase-less four-run balancing
POST	/api/vibration/balance/static-couple	Static / couple unbalance decomposition
POST	/api/vibration/balance/split-weight	Correction weight split across fixed holes
POST	/api/vibration/belt-calculator	Belt passing frequency calculator
POST	/api/vibration/alignment-calculator	Face & rim shaft alignment shim calculator
POST	/api/vibration/unit-converter	13-unit vibration amplitude converter
POST	/api/vibration/sdof-simulator	Mass-spring-damper SDOF free response
GET	/api/vibration/bearing-database	Kinematic bearing catalog
GET	/api/vibration/bearing-search	Bearing catalog search by model or brand

Domain workbenches

Method	Endpoint	Description
POST	/api/electrical/analyze	3-phase symmetrical components, power, THD
POST	/api/electrical/upload	Analyse an uploaded electrical signal file
POST	/api/antenna/analyze	VSWR, return loss, FSPL, waveguide cutoff
POST	/api/gps/simulate	GPS L1 C/A constellation & baseband simulation
GET	/api/gps/gold-code/{prn}	1023-chip C/A Gold Code for a PRN
POST	/api/gps/export-iq-bin	Binary I/Q baseband stream for SDR hardware
POST	/api/srw/simulate	Synchrotron & undulator radiation optics
POST	/api/dsp-lab/sampling-aliasing	Sampling theorem & aliasing fold-over
POST	/api/dsp-lab/fir-parks-mcclellan	Parks-McClellan equiripple FIR design
POST	/api/dsp-lab/autocorrelation	Normalized autocorrelation & pitch detection
POST	/api/dsp-lab/lms-adaptive	LMS adaptive noise canceller
POST	/api/dsp-lab/cwt-scalogram	Continuous Wavelet Transform scalogram

MATLAB Add-On & Code Export

Method	Endpoint	Description
GET	/api/matlab/client	Download REISignalLab.m object-oriented MATLAB client class
GET	/api/matlab/installer	Download install_rei_signallab.m 1-click MATLAB installer script
POST	/api/matlab/export-m	Export node graph project into a standalone executable .m script

Error envelope

Every error response shares one structured shape, with X-Request-ID echoed on the response header:

{
  "code": "RATE_LIMIT_EXCEEDED",
  "message": "Rate limit exceeded for 'sandbox': 10 requests per 60s.",
  "details": { "bucket": "sandbox", "limit": 10, "retryAfterSeconds": 42 },
  "requestId": "0f4c1a9e-..."
}

---

Quick Start

1. Clone & Run Backend

git clone https://github.com/rootcastleco/rei-signallab.git
cd rei-signallab

# Create virtual environment
py -m venv .venv
.venv\Scripts\activate

# Install dependencies (requirements-dev.txt adds pytest + httpx on top of
# requirements.txt, which holds runtime dependencies only)
pip install -r backend/requirements-dev.txt

# Run Pytest suite
$env:PYTHONPATH="backend"; py -m pytest backend/tests

# Start FastAPI server
$env:PYTHONPATH="backend"; py -m uvicorn app.main:app --reload --port 8000

Interactive API docs are then available at http://localhost:8000/docs.

To exercise the Python scripting sandbox locally, start the server with ENABLE_PYTHON_SANDBOX=true — it is off by default under APP_ENV=production.

2. Run Frontend Studio

cd frontend
npm install
npm run dev

Open http://localhost:3000 in your browser. The Vite dev server proxies /api to http://localhost:8000.

npm run lint          # ESLint
npm test              # Vitest
npm run test:coverage # Vitest with coverage report
npm run build         # Production bundle

---

License

Distributed under the MIT License. See LICENSE for more information.

---

Developed by Batuhan Ayribas

RootCastle — Engineering Beyond Boundaries

REI SignalLab 2.1 • Version 2.1.0 • August 2026