Skip to content

RELEASE_REPORT.md

Latest commit

 

History

History
82 lines (65 loc) · 2.92 KB

RELEASE_REPORT.md

File metadata and controls

82 lines (65 loc) · 2.92 KB
Raw

QECTOR Decoder Workbench v3.5.0 — Production Release Report

Date: 2026-07-13 Version: 3.5.0 (WORKBENCH_VERSION) Backend: qector_decoder_v3 0.6.6 (auto-upgrades to PyPI latest)

1. Test Suite Results

Suite Result
pytest All tests passing (including streaming & security depth)
MCP Server (29 tools) 29/29 passed
Module imports (22 modules) 22/22 passed

2. Module Inventory (22 modules)

Core (9 modules)

Module Purpose
backend.py QEC wrapper — 5 decoders, 6 families, layout, decoder info, streaming
app.py Main GUI app — wired tabs + console tab + status bar
state.py AppState for shared configuration
main.py PyInstaller entry point
version.py Central version info (v3.5.0, MCP_TOOLS=29)
logger.py File + stdout logging
console.py Console buffer for UI redirection
utils.py String & math utils
theme.py UI COLORS + Fonts

GUI Tabs (6 modules)

Tab Module Purpose
code_explorer_tab.py Qubit/check configuration and spring layout
decoder_lab_tab.py Single-shot decode error/syndrome visualization
benchmark_tab.py Timing throughput & latency statistics per decoder
batch_streaming_tab.py CPU/CUDA/OpenCL batch decode and streaming
hardware_tab.py GPU availability and recommended decoder mapping
documentation_tab.py Multi-format generator interface with preview

Auxiliary & MCP (7 modules)

Module Purpose
auto_updater.py Updates package on boot (with pre-release parsing)
threading_utils.py Threading utilities
results_tracker.py Saves benchmark runs to JSON
hardware_routing.py Reimplemented recommendation logic
mcp_server.py stdio JSON-RPC 2.0 transport + 29 tools
mcp_resources.py Allocates resources for tools
dialogs.py CustomTkinter message dialog wrappers

3. Production Packaging

PyInstaller Configuration

  • spec file: QectorWorkbench.spec
  • Hidden imports explicitly collect all 22 custom QECTOR modules plus standard dependencies
  • Excludes PyTorch, pandas, TensorFlow, etc.

4. Validation Checklist

  • All pytest tests pass (including previously skipped streaming tests)
  • All 29 MCP tools functional (verified by test_mcp_all.py)
  • All 22 modules import cleanly (even headless)
  • Auto-updater detects PyPI version and parses rc/pre-release versions properly
  • Hardware detection (CUDA/OpenCL/CPU) works
  • All 5 decoders functional across all 6 code families
  • Batch decode with correct shape/dtype and selectable backends
  • Spring layout for all Tanner graph families
  • Multi-format doc export (Markdown, JSON, HTML, LaTeX, PDF, SVG)
  • PyInstaller build completes with 0 errors

Release Approved. Status: PRODUCTION READY — FLAWLESS