Precision File Search V1.1.1 - The Regex Fix

This is an important stability and precision update that addresses a key bug in the Classic Search engine's "Use Regex" functionality, thanks to community feedback from issue #18. This fix dramatically improves the accuracy of file name searches when using patterns and wildcards.

---

🐛 Key Bug Fix & Improvement: Precise File Extension Searching

This release resolves a critical issue where the "File Names" search was overly broad and did not correctly handle exact file extension matching.

The Problem:
Previously, when using the "Use Regex" option, the engine would perform a substring match. For example, a search for *.go would incorrectly include files like project.godot in the results. Furthermore, attempts to use the standard regex anchor $ (e.g., *.go$) to force an exact match would fail and return zero results due to incorrect pattern escaping.

The Fix:
The search engine's pattern translator has been re-engineered to be "anchor-aware". It now correctly processes the $ (end-of-string) and ^ (start-of-string) characters, allowing for powerful and precise pattern matching.

This change gives users much more precise control and makes the search behavior more intuitive for those familiar with regular expressions.

Installation

• Download the PFS-SETUP_1.1.1.exe file from the Assets section below.
• Run the installer and follow the on-screen instructions. Administrative privileges may be required.
• For AI & Semantic Search: On the first run, please ensure you have an internet connection to allow the application to download the necessary multilingual AI models. All subsequent runs can be fully offline.

This is a recommended update for all users of the Classic Search feature to ensure the most accurate and predictable search results.

V1.1.0

Precision File Search v1.1.0 - The Multilingual Update

This is a major feature release that introduces powerful multilingual search capabilities, significant performance optimizations, and several key bug fixes. Version 1.1.0 transforms PFS into a truly global tool, allowing users to search their documents in their native language while benefiting from dramatically improved speed and stability across the entire application.

🚀 Key Features & Enhancements in this Release:

🌐 Multilingual AI Search (New Feature)

The AI Search feature is no longer limited to English. The application now supports semantic search for a wide variety of languages.

• Search in Your Language: You can now type your queries directly in languages like Persian, Spanish, French, German, Chinese, and many others.
• Architecture Change: The system has been re-architected to pass your original query directly to the multilingual model, ensuring the highest possible accuracy by comparing vectors in a shared, language-independent space.
• Recommended Model: Any multi-lingual embedding model.

⚡ Performance Overhaul: Up to 5x Faster Searches

The Classic "Keywords in Files" search engine has been significantly optimized, especially for its first run after application startup.

• Proactive Engine Warm-Up: The text extraction engine (unstructured) is now "warmed up" in the background when the application starts. This eliminates the long "cold start" delay, making your first content search dramatically faster.
• Optimized Text Extraction: All text processing now uses a "fast" strategy, which intelligently bypasses slow and unnecessary operations like language detection, resulting in a significant performance boost when searching through complex files like PDFs.
• Cleaner Logs: Verbose and harmless warnings from the pdfminer library have been suppressed, making the application logs much cleaner and easier to read for troubleshooting.

🐛 Bug Fixes & Stability

• Fixed: The Classic Search engine can now correctly find keywords inside Microsoft Office documents (.docx, .xlsx), PDFs, and other complex file formats. This was a critical bug that limited content search to plain-text files.
• Fixed: The application will no longer re-download AI models on every startup if they already exist locally. This resolves long loading times and improves the user experience.
• Fixed (True Offline Mode): The application now works in a fully offline environment. It has been re-architected to download models to a dedicated local folder and load them directly from that path, bypassing all online verification checks and ensuring a seamless experience without an internet connection.
• Fixed (UI): The AI Response panel now correctly aligns its layout, including headers and text, for right-to-left (RTL) languages like Persian, ensuring a polished user experience for international users.

---

This release also incorporates all the usability and security improvements from v1.0.15.

Installation

1. Download the PFS-SETUP_1.1.0.exe file from the Assets section below.
2. Run the installer and follow the on-screen instructions. Administrative privileges may be required.
3. For AI & Semantic Search: On the first run, please ensure you have an internet connection to allow the application to download the necessary multilingual AI models. All subsequent runs can be fully offline.

---

This is a significant feature and performance upgrade. All users are encouraged to update to benefit from multilingual support and a much faster, more stable experience.

V1.0.15

Precision File Search v1.0.15

This is a targeted usability and bug fix release that addresses a limitation in the semantic file indexing system, restoring the ability for users to index folders from any drive on their computer.

Key Changes in this Release:

📂 Usability Fix: Index Folders from Any Drive (Resolves Issue #6)

The application's security validation has been refined to provide the flexibility expected of a desktop application while maintaining robust security.

• Problem Solved: Previously, a strict security check incorrectly prevented users from indexing folders located outside their primary user directory (e.g., on a D:\ or F:\ drive). This has been fixed.
• Enhanced Flexibility: You can now select and index any valid folder on any local drive that your user account has permission to access.
• Security Maintained: The path validation system still provides strong protection against path traversal attacks (e.g., using ../), ensuring that only the explicitly specified directory is ever processed.
• Documentation Updated: All relevant documentation (README.md, security notes) has been updated to accurately reflect this new, more flexible behavior.

---

This release also incorporates all the stability and performance improvements from v1.0.14.

Installation

1. Download the PFS-SETUP_1.0.15.exe file from the Assets section below.
2. Run the installer and follow the on-screen instructions. Administrative privileges will be required.
3. For AI Search features, please remember to configure your LLM and Semantic models in the Settings tab as described in the main README.md.

---

This is a key usability fix that restores expected functionality for a desktop search application. All users are encouraged to upgrade to have full access to their files.

V1.0.14

Precision File Search v1.0.14

This is an important stability and performance update for Precision File Search. This release focuses on resolving critical bugs in the semantic indexing engine, improving the intelligence of the AI, and enhancing the overall robustness of the application for distribution.

Key Changes in this Release:

🚀 Major Fix: Robust PDF Indexing (Resolves Issue #2)

The semantic indexing engine has been significantly overhauled to provide fast, reliable, and self-contained PDF parsing.

• Removed External Dependencies: The dependency on the external Poppler library has been completely eliminated. The application no longer requires users to perform manual installation or PATH configuration to index PDF files.
• New Hybrid Parser: The system now uses the high-performance PyMuPDF library for all PDF documents, ensuring stability and speed. For all other file types (.docx, .txt, .md, etc.), it continues to use the powerful unstructured library to maintain broad format support.

✨ Enhancement: Context-Aware AI Follow-ups

The "Ask AI" feature is now significantly smarter and more intuitive.

• Prevents Hallucination: When you perform a Semantic Search and then ask a follow-up question in the AI Search tab, the AI will now correctly use the on-screen search results as its only context. This provides accurate, grounded answers based on your documents and prevents the AI from answering with irrelevant information from its internal knowledge base.

🔧 Build & Packaging Improvements

• Professional Installer: The Inno Setup script has been refined to create a standard, administrative installer that correctly places all application files and shows the full installation wizard, including the license agreement.
• Self-Contained Executable: The PyInstaller build process is now managed by a robust .spec file, ensuring all necessary dependencies (like unstructured.nlp data and hidden imports) are correctly bundled into the final pfs.exe.

Installation

1. Download the PFS-SETUP_1.0.14.exe file from the Assets section below.
2. Run the installer and follow the on-screen instructions. Administrative privileges will be required.
3. For AI Search features, please remember to configure your LLM and Semantic models in the Settings tab as described in the main README.md.

---

This version closes a major usability issue and significantly improves the reliability of the core semantic features. We highly recommend all users upgrade.

Precision File Search v1.0.0

🎉 First Official Release of Precision File Search!

This is the first stable, production-ready release of Precision File Search (PFS), a private, AI-powered desktop search and knowledge engine.

Key Features in this Release:

• 🚀 High-Performance File Engine: A concurrent, memory-efficient engine for lightning-fast classic file and content searches.
• 🧠 Intelligent Semantic Search: Build a local knowledge base from your files and search by meaning, not just keywords. Features an intelligent re-indexing system that only processes new or changed files.
• ✨ AI-Powered Orchestrator: Use natural language to ask questions and get summarized, synthesized answers directly from your documents.
• 🤖 Trainable ML Document Classifier: Automatically sort your files into logical categories. The model can be re-trained with your own data for custom classification.
• 🔒 Privacy-First Architecture: The entire application runs 100% locally. Your data never leaves your machine.

Installation

• Download the pfs-installer-v1.0.0.exe file from the Assets section below.
• Run the installer and follow the on-screen instructions.
• For AI Search features, please remember to configure your LLM and Semantic models in the Settings tab as described in the main README.md.