SignalForge Studio v1.1.2

SignalForge Studio v1.1.2

Summary

This patch release hardens local file import behavior and improves validation of PCA diagnostic configuration. It preserves the local-first research boundary: the reviewed application workflow imports locally selected CSV/text data, executes diagnostics locally, and exports only to a user-selected destination.

Improvements

Safer local imports

SignalForge now rejects import files larger than 16 MiB before the file is read into memory. This complements the existing sample-count limit and helps keep the desktop workflow responsive when a user selects an unexpectedly large CSV or text file.

Operating-system read errors, including a denied file read, are now normalized into a clear user-facing validation error instead of surfacing an unhandled system exception.

Clearer PCA validation

PCA diagnostics now reject two invalid configurations before they reach third-party library internals:

• A variance-ratio n_components setting requires the full PCA solver.
• An integer n_components setting cannot exceed the smaller of the observation count and asset count.

These checks produce stable DiagnosticValidationError messages that can be handled consistently by the desktop application and callers.

Expanded regression coverage

The regression suite now checks oversized-file rejection, file-access error conversion, PCA solver compatibility, component-dimension limits, and the output contract between sign-aligned PCA loadings and scores.

Validation

The v1.1.2 source line passed the complete local quality gate before tagging:

Gate	Result
Unit and desktop smoke tests	20 passed
Bytecode compilation	Passed
Dependency integrity (pip check)	Passed
Diff whitespace validation	Passed
Local PyInstaller package and offscreen smoke test	Passed

Notes and limitations

SignalForge is a local research and diagnostic application. It does not submit trades, connect to brokers, provide personalized investment advice, make credit decisions, or claim security/regulatory certification. A local-first workflow reduces routine vendor-side custody of data, but the customer remains responsible for endpoint hardening, storage policy, data licensing, and appropriate model governance.

For Windows, verify the downloaded executable using the accompanying SHA-256 file. This release does not claim Authenticode code signing.

SignalForge Studio v1.1.1

Full Changelog: v1.1.0...v1.1.1

SignalForge Studio v1.0.0

SignalForge Studio v1.0.0

SignalForge Studio is the first Windows desktop release of the SignalForge DSP toolkit. It provides a local-first workspace for generating or importing one-dimensional sampled signals, applying optional moving-average smoothing, inspecting waveform and FFT magnitude views, reviewing RMS/peak/mean/dominant-frequency metrics, and exporting CSV or JSON results.

This release includes a portable Windows 10/11 x64 executable, its SHA-256 checksum, and third-party notices. The executable was built on a Windows runner after the full 10-test numerical and desktop smoke suite passed.

Verify the download

In Windows PowerShell, calculate the executable hash and compare it exactly with the contents of the accompanying .sha256 file:

(Get-FileHash .\SignalForge-Studio-v1.0.0-win64.exe -Algorithm SHA256).Hash.ToLower()
Get-Content .\SignalForge-Studio-v1.0.0-win64.exe.sha256

Notes

This product is intended for education and exploratory analysis. It is not certified for medical, safety-critical, metrology, or compliance use. The application is not code-signed in this release, so Windows reputation prompts may occur.