PC_Workman_HCK Files

Build Date: March 10, 2026
Status: Alpha Testing
Target: Technical testers & early adopters
Stability: Significantly improved over 1.6.3 Screenshots lower!

---

What's New - The Big Picture

This release is about making PC Workman work for months, not minutes. The app now remembers your system's behavior across sessions, detects patterns, and actually uses that data for intelligent alerts and insights.

Key shift: From "monitoring tool that resets every launch" to "system companion that learns over time."

---

Core Engine Rebuild: HCK_STATS_ENGINE

What Changed

Replaced CSV-only logging with SQLite-backed time-series database. Your data now persists properly across sessions, aggregates intelligently, and doesn't blow up after 3 days of runtime.

Architecture

Raw samples (1s) -> Minute averages (60s) -> Hourly stats -> Daily stats
                                           \/
                                    Process tracking
                                    Event detection
                                    Pattern analysis

New modules:

• db_manager.py - SQLite with WAL mode (concurrent read/write, no locks)
• aggregator.py - Time-series rollup (minute → hour → day)
• process_aggregator.py - Per-process CPU/RAM tracking with hourly/daily aggregation
• query_api.py - Read interface for UI (separate connection, no blocking)
• events.py - Spike detection, anomaly logging, alert generation

Stability guarantees:

• Every DB call wrapped in try/except - scheduler never crashes
• Atomic transactions - crash mid-write = rollback, data stays consistent
• Graceful degradation - SQLite fails = app falls back to CSV mode
• No new dependencies - sqlite3 is Python stdlib

What this fixes:

• Bug A (Data Aggregation): Lifetime uptime now calculated correctly by merging daily/hourly/minute stats without overlap. No more "0 hours total" after relaunch.
• Bug B (Telemetry Noise): System processes like "System Idle Process", "Interrupts", "Memory Compression" are filtered out. No more 800% CPU alerts from kernel threads.

---

Intelligence Layer: InsightsEngine

Context Awareness

hck_GPT now knows what's happening on your system and adjusts its responses accordingly.

Features:

• Time-aware greetings: "Good morning, yesterday you ran for 6.2 hours..."
• Real-time spike detection: CPU/GPU/RAM/temp anomalies trigger notifications
• App recognition: Tracks browsers, games, dev tools by name
• Pattern teasers: "Ready for another round of Battlefield?" (based on 7-day patterns)
• 24h anomaly report: "3 temp spikes detected between 14:00-16:00"
• Dynamic banner status: Changes from "All quiet" to "CPU 87% SPIKE | 2 alerts"

New commands in hck_GPT chat:

• stats / co uzywam - Session & lifetime summary
• alerts / alerty - Recent anomalies report
• insights / co nowego - Current system insights
• teaser - Personalized suggestions based on usage
• health - Quick system health check
• report - Full colored report in chat window

Technical implementation:

• insights.py (~300 LOC) - Pattern detection, greeting logic, spike tracking
• Auto-greeting on panel open (once per session or 30min cooldown)
• Insight ticker: 60s intervals (only shows notable events, not spam)
• Banner ticker: 30s intervals (updates mini-status bar)
• Integrated with process_aggregator for app tracking

---

UI/UX Refinements

Dashboard

• Main loop optimization: 300ms → 1000ms update cycle (-70% UI overhead)
• Widget reuse pattern: TOP 5 process panels no longer destroy/recreate on every update (massive memory improvement)
• Chart engine: Replaced pixel-by-pixel PhotoImage rendering with reusable canvas objects (orders of magnitude faster)
• Cached constants: cpu_count, total_ram cached on startup (no redundant psutil calls)

My PC Page (Work in Progress)

Live now:

• Real-time CPU/RAM/GPU monitoring with temperature simulation
• Hardware detection (actual CPU name, RAM total, GPU model)

Coming soon (buttons visible but not yet functional):

• Health Report - Comprehensive system health analysis
• Cleanup - Temp files, cache cleanup
• Stability Tests - Program diagnostics, log viewer, HCK_STATS_ENGINE status
• Your Account - Details - User settings and preferences

TOP 3 (Planned for 1.7.0):

• Stats & Alerts - Pattern-based alerts for temp spikes, voltage anomalies, suspicious activity (week-scale analysis)
• Optimization & Services - Windows services management, startup optimization
• First Setup & Drivers - Initial configuration wizard, driver checks

Info Section (Dashboard)

• Redesigned compact layout (50px height)
• Purple accent (#a78bfa) styling
• Consolas monospace font for technical readability
• Optimized typewriter animation (70ms typing / 3-char burst deletion)
• Punchy one-liner messages instead of verbose text

Charts

• Default mode: LIVE (real-time, last 30 seconds)
• Historical modes: 1H (3600s), 4H (14400s) with proper range mapping
• Auto-refresh: Historical views refresh every 30s from database
• Fixed bug: Switching between LIVE/1H/4H now loads correct data range

---

Performance Optimizations

Threading Model

Before: Main thread blocked on psutil.process_iter() - caused UI lag spikes
After: Heavy telemetry collection offloaded to background daemon thread

Implementation:

• Telemetry thread runs independently (1s sampling rate)
• read_snapshot() provides instant, non-blocking access to latest data
• Graceful thread lifecycle (starts in startup.py, joins on shutdown)

Result: Zero UI lag, butter-smooth updates

Hardware Polling

• Hardware cards: 2s update interval (balance between real-time and low CPU)
• System tray: 3s update interval (less critical, less overhead)
• Charts: 2s with canvas object reuse (no redraw cost)

Code Cleanup

Removed dead code:

• utils/file_utils.py, utils/net_utils.py, utils/system_info.py (unused)
• ui/expandable_process_list.py (obsolete component)
• settings/ directory (migrated to database)
• _animate_button_shimmer() (60 LOC, unnecessary CPU burn)
• System artifacts (_nul, nul files)
• Obsolete exports (fan_settings_ultimate.json)

main_window.py refactor:

• 1606 lines → 1460 lines (-9%, -146 LOC)
• Extracted magic numbers to named constants
• Unified duplicate process rendering logic (DRY)
• Circuit breaker pattern for error handling (max 10 errors before graceful shutdown)
• Dynamic RAM detection (replaces hardcoded 8192MB)
• Better exception handling with detailed logging

---

Critical

• [A] Lifetime Stats Reset: Session uptime now persists across launches (multi-tier aggregation fix)
• [B] System Process Noise: Kernel threads filtered out (no more 800% CPU alerts)
• [C] Chart Data Loading: Historical modes (1H/4H) now correctly query database instead of showing stale data

Stability

• Scheduler crash prevention: All DB operations wrapped in try/except
• UI thread safety: Separate DB connections for read (UI) and write (scheduler)
• Race condition fix: Widget existence checks before updates (winfo_exists() guards)
• Memory leak fix: Process panels now reuse widgets instead of recreate cycle

UI

• Dashboard-only updates: _update_hardware_cards() and _update_top5_processes() only run when dashboard is visible (eliminates majority of error sources)
• Navigation routing: Fixed sidebar IDs (temperature, voltage, alerts instead of old realtime, processes)

---

For Testers: What to Focus On

Primary Test Areas

1. Long-term stability - Leave running for 24+ hours, check memory usage
2. Session persistence - Close app, reopen, verify lifetime stats continue (not reset to 0)
3. hck_GPT intelligence - Try commands: stats, alerts, insights, report
4. Chart accuracy - Switch between LIVE/1H/4H modes, verify data matches
5. Process tracking - Launch heavy apps (games, browsers), check if detected properly

Known Issues (Work in Progress)

• My PC page: Health Report, Cleanup, Stability Tests buttons are placeholders (visible but not functional yet)
• TOP 3 features: Planned for v1.7.0 (UI visible, logic not implemented)
• Temperature monitoring: Currently simulated (based on load %), real sensor integration coming
• Voltage monitoring: Not yet implemented

Performance Expectations

• CPU usage: ~1-2% idle, ~5-8% when dashboard active
• RAM usage: ~80-120 MB (depends on process history size)
• Database size: ~5-10 MB per month of runtime (tested over 3 months continuous use)

---

Installation & Requirements

No new dependencies. Everything uses Python stdlib.

Python version: 3.9+
Tested on: Windows 10/11

Run from source:

:::bash
python startup.py

Known limitations:

• Windows only (psutil limitations for GPU/temp on other platforms)
• Requires administrator rights for some system process queries

---

Roadmap

v1.6.9 (Optional cleanup release):

• Optimize remaining files (core/monitor.py, hck_gpt modules)
• More constant extraction
• Documentation improvements

v1.7.0 (Feature release, ~3-4 weeks):

• Complete My PC page functionality (Health Report, Cleanup, Stability Tests)
• TOP 3 features (Stats & Alerts, Optimization, First Setup)
• Real temperature sensor integration (OpenHardwareMonitor)
• Voltage monitoring
• Pattern-based alerting system

---

Feedback

This is an alpha build for testers. Expect rough edges.

Report issues:

• GitHub Issues: https://github.com/HuckleR2003/PC_Workman_HCK/issues
• Include logs from data/logs/ folder
• Screenshots appreciated

What helps:

• Long-term stability reports (24h+ runtime)
• Memory usage observations
• Database size after X days
• Any crashes (with logs)

---

Credits

Built by: Marcin "HCK" Firmuga
Part of: HCK_Labs educational AI engineering initiative
License: MIT (educational, non-commercial)

Special thanks to testers who ran 1.6.3 and reported the session reset bug.

---

Screenshots

Version: 1.6.8
Release Date: March 10, 2026
Commit: [your commit hash]
Build: Alpha Testing

---

"From monitoring tool to system companion. This is where PC Workman learns."

[![Download PC_Workman_HCK](https://a.fsdn.com/con/app/sf-download-button)](https://sourceforge.net/projects/pc-workman-hck/files/v_1.6.8/)