v0.5.0: Multi-format support, ML dedup, code quality & docs update

Added:
- Multi-format signal loading (CSV, MAT, WAV, NPY, Parquet) via unified load_signal_data()
- python -m predictive_maintenance_mcp entry point (__main__.py)
- Ollama integration guide (docs/OLLAMA_GUIDE.md)
- 3 Copilot Skills (bearing-diagnosis, quick-screening, report-generation)
- MANIFEST.in updated for new data formats

Changed:
- ML training code deduplicated (4 shared helpers, ~163 statements removed)
- PyPDF2 migrated to pypdf
- ISO metadata double-read eliminated
- pd.read_csv calls hardened with engine/dtype/on_bad_lines
- pytest config consolidated into pyproject.toml (pytest.ini removed)
- Version bumped 0.4.1 -> 0.5.0 across all files
- Comprehensive docs update (README, CHANGELOG, QUICKSTART, CONTRIBUTING, INSTALL)

Fixed:
- Envelope analysis band_start/band_end validation
- FFT frequency axis off-by-one
- ISO unit-confirmation flow
- ML predict tool missing model guard
- Anomaly plot data alignment

Tests: 44 passed, 3 skipped, 0 warnings

Author: LGDiMaggio

.env.example

@@ -22,4 +22,4 @@ STATISTICS_WINDOW_SIZE=1024  # For windowed statistics (future feature)
 
 # Server configuration
 SERVER_NAME=Predictive Maintenance
-SERVER_VERSION=0.4.1
+SERVER_VERSION=0.5.0

.gitignore

@@ -43,9 +43,11 @@ venv.bak/
 # Testing
 .pytest_cache/
 .coverage
+coverage.xml
 htmlcov/
 *.cover
 .hypothesis/
+tests/data/
 
 # ML Models (generated during testing)
 models/*.pkl

CHANGELOG.md

@@ -5,6 +5,28 @@ All notable changes to the Predictive Maintenance MCP Server project will be doc
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.5.0] - 2026-02-16
+
+### Added
+- **Multi-format signal loading** — `load_signal_data()` now supports CSV, TXT, NPY, MAT (MATLAB), WAV, and Parquet formats
+- **`__main__.py`** — Server can now be run as `python -m predictive_maintenance_mcp`
+- **Ollama Guide** added to documentation table in README
+
+### Changed
+- **Unified signal loading** — All 16 `pd.read_csv()` call sites refactored to use `load_signal_data()`, enabling all tools to accept any supported format
+- **ML code deduplication** — Extracted 4 helper functions (`_resolve_sampling_rate`, `_segment_and_extract_features`, `_extract_features_from_files`, `_extract_and_transform_validation_features`), reducing ~163 statements in `train_anomaly_model`
+- **ISO metadata consolidation** — `evaluate_iso_20816` now reads metadata file once instead of twice
+- **PyPDF2 → pypdf migration** — Replaced deprecated PyPDF2 with pypdf in `document_reader.py`
+- **Pytest config consolidation** — Merged `pytest.ini` into `pyproject.toml` (`[tool.pytest.ini_options]`)
+- **Logging to stderr** — Server logging now uses `stderr` to avoid polluting MCP stdio transport
+- **Report filenames** — `report_generator.py` uses `Path.stem` for all filename sanitizations
+
+### Fixed
+- **Packaging** — Corrected `pyproject.toml` package-dir mapping (`src/` → `predictive_maintenance_mcp`)
+- **Metadata paths** — `get_metadata_path()` now uses `Path.stem` to work with all signal extensions
+- **Plot output directories** — Report generator creates output directories before writing files
+- **Flaky ML test** — Fixed `test_predict_anomalies` instability with deterministic seed
+
 ## [0.4.1] - 2026-02-15
 
 ### Fixed
@@ -116,14 +138,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## Roadmap
 
-### Planned for v0.5.0
+### Planned for v0.6.0
 - **📦 Docker image** for zero-install setup
-- **📂 Parquet/HDF5 data format support**
 - **📏 Customizable ISO report thresholds**
 - Multi-signal comparison tools
 - Advanced trending and monitoring
 - Additional diagnostic workflows (pumps, motors, gearboxes)
-- Enhanced ML models with hyperparameter tuning
 - Extended dataset with more fault types
 
 ### Future Enhancements

CITATION.cff

@@ -1,8 +1,8 @@
 cff-version: 1.2.0
 message: "If you use this software, please cite it as below."
 title: "Predictive Maintenance MCP Server: An open-source framework for integrating Large Language Models with predictive maintenance and fault diagnosis workflows"
-version: 0.4.1
-date-released: 2026-02-15
+version: 0.5.0
+date-released: 2026-02-16
 authors:
   - family-names: Di Maggio
     given-names: Luigi Gianpio
@@ -52,7 +52,7 @@ preferred-citation:
     - family-names: Di Maggio
       given-names: Luigi Gianpio
   year: 2025
-  version: 0.4.1
+  version: 0.5.0
   repository-code: "https://github.com/LGDiMaggio/predictive-maintenance-mcp"
   license: MIT
   doi: "10.5281/zenodo.17611542"

CONTRIBUTING.md

@@ -232,7 +232,7 @@ Closes #42"
 
 | Task | What You'll Learn | Issue |
 |------|-------------------|-------|
-| **Add Parquet data format support** | Signal loading pipeline, pandas I/O | [Browse issues](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
+| ~~**Add Parquet data format support**~~ | ~~Signal loading pipeline, pandas I/O~~ | ✅ Done in v0.5.0 |
 | **Make ISO thresholds configurable** | Tool parameter design, ISO standard | [Browse issues](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
 | **Add unit conversion tool** | MCP tool pattern, unit systems | [Browse issues](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
 | **Improve error messages** | Error handling best practices | [Browse issues](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
@@ -315,7 +315,7 @@ What actually happens (include full error message)
 **Environment:**
 - OS: Windows 11 / macOS 14 / Ubuntu 22.04
 - Python: 3.11.x / 3.12.x
-- Server version: 0.4.1
+- Server version: 0.5.0
 ```
 
 ---

INSTALL.md

@@ -186,8 +186,8 @@ pip install -e .
    - ✅ `"command": "C:/path/.venv/Scripts/python.exe"` (Windows)
    - ✅ `"command": "/path/.venv/bin/python"` (macOS/Linux)
 
-4. **Don't use `-m` module import**: The package isn't installed as a module:
-   - ❌ `"args": ["-m", "machinery_diagnostics_server"]`
+4. **Module import**: If installed via `pip install -e .`, you can also use:
+   - ✅ `"args": ["-m", "predictive_maintenance_mcp"]`
    - ✅ `"args": ["C:/path/src/machinery_diagnostics_server.py"]`
 
 5. **Restart Claude Desktop completely** after config changes

MANIFEST.in

@@ -7,12 +7,11 @@ include CHANGELOG.md
 include INSTALL.md
 
 # Include configuration files
-include pytest.ini
 include pyproject.toml
 include setup_venv.py
 
 # Include data files
-recursive-include data *.csv *.json *.txt
+recursive-include data *.csv *.json *.txt *.mat *.wav *.npy *.parquet
 recursive-include data/signals/samples README.md
 
 # Include tests

README.md

@@ -92,6 +92,7 @@ This project is built around the **Model Context Protocol (MCP)** — an open st
 - [🚀 Installation](#-installation)
 - [⚙️ Configuration](#️-configuration)
 - [🔧 Available Tools & Resources](#-available-tools--resources)
+- [🤖 Copilot Skills](#copilot-skills-guided-workflows)
 - [📐 Detailed Architecture](#-detailed-architecture)
 - [📊 Sample Dataset](#-sample-dataset)
 - [💡 Usage Examples](#-usage-examples)
@@ -163,14 +164,16 @@ This project serves two audiences. Pick the door that fits you:
   ![ISO Compliance](assets/iso.png)
 - **🔍 Advanced Diagnostics** — FFT spectrum analysis, envelope analysis for bearing faults, time-domain feature extraction
   <details>
+
   <summary><b>Example analysis</b></summary>
 
   ![Envelope analysis 1](assets/envelope_analysis.png)
   ![Envelope analysis 2](assets/envelope_signals.png)
   ![Envelope analysis 3](assets/envelope_list.png)
 
   </details>
-- **🚀 Zero Configuration** — Works out of the box with sample data, auto-detects sampling rates from metadata
+- **� Multi-Format Support** — Load signals from CSV, MAT (MATLAB), WAV, NPY, and Parquet files
+- **�🚀 Zero Configuration** — Works out of the box with sample data, auto-detects sampling rates from metadata
 
 ---
 
@@ -247,6 +250,14 @@ cd predictive-maintenance-mcp
 pip install -e .
 ```
 
+### Run as Module
+
+After installation, you can also run the server directly:
+
+```bash
+python -m predictive_maintenance_mcp
+```
+
 ---
 
 ## ⚙️ Configuration
@@ -307,7 +318,7 @@ Resources provide **direct read access** for Claude to examine data:
 <summary><b>📊 Vibration Signals</b></summary>
 
 - **`signal://list`** — Browse all available signal files with metadata
-- **`signal://read/{filename}`** — Read signal data directly (first 1000 samples preview)
+- **`signal://read/{filename}`** — Read signal data directly (supports CSV, MAT, WAV, NPY, Parquet)
 
 **Usage:** Claude can directly read signals without calling tools first.
 
@@ -390,6 +401,20 @@ Tools perform **computations and generate outputs**:
 
 ---
 
+### Copilot Skills (Guided Workflows)
+
+The `skills/` directory contains pre-built guided workflows that orchestrate multiple MCP tools into structured diagnostic procedures:
+
+| Skill | Steps | Description |
+|-------|:-----:|-------------|
+| [**bearing-diagnosis**](skills/bearing-diagnosis/SKILL.md) | 8 | Complete bearing fault detection: statistics → FFT → envelope → frequency matching → ISO severity → report |
+| [**quick-screening**](skills/quick-screening/SKILL.md) | 5 | Fast health screening with clear Healthy/Suspicious/Critical classification |
+| [**report-generation**](skills/report-generation/SKILL.md) | 6 | Professional HTML report generation with composite multi-report option |
+
+> 💡 Skills are standalone markdown files that any MCP-compatible LLM client can use as system instructions to coordinate multi-step diagnostic workflows.
+
+---
+
 ## 📐 Detailed Architecture
 
 The system follows a **hybrid MCP architecture** combining Resources (direct data access) and Tools (computational processing):
@@ -437,7 +462,7 @@ The system follows a **hybrid MCP architecture** combining Resources (direct dat
 │  SIGNAL ANALYSIS │   │  DOCUMENT READER MODULE          │
 │  MODULE          │   │  ┌────────────┐  ┌────────────┐  │
 │  • FFT Engine    │   │  │ PDF Extract│  │ ISO Formulas│  │
-│  • Envelope      │   │  │ (PyPDF2)   │  │ BPFO/BPFI   │  │
+│  • Envelope      │   │  │ (pypdf)    │  │ BPFO/BPFI   │  │
 │  • Filters       │   │  └────────────┘  └────────────┘  │
 │  • Statistics    │   │  ┌─────────────────────────────┐  │
 │  • ML Models     │   │  │  Bearing Catalog DB         │  │
@@ -467,7 +492,8 @@ The system follows a **hybrid MCP architecture** combining Resources (direct dat
 
 **Key Features:**
 - ✅ **4 MCP Resources** — Direct read access to signals and manuals
-- ✅ **25+ MCP Tools** — Complete diagnostic workflow
+- ✅ **25 MCP Tools** — Complete diagnostic workflow
+- ✅ **4 MCP Prompts** — Guided diagnostic workflows
 - ✅ **Hybrid Architecture** — Resources for reading, Tools for processing
 - ✅ **Local-First** — All data stays on your machine (privacy-preserving)
 
@@ -568,8 +594,10 @@ Generate FFT report for baseline_1.csv
 | [EXAMPLES.md](EXAMPLES.md) | Everyone | Complete diagnostic workflows with step-by-step tutorials |
 | [INSTALL.md](INSTALL.md) | Everyone | Detailed installation and troubleshooting guide |
 | [CONTRIBUTING.md](CONTRIBUTING.md) | Contributors | How to contribute (for every skill level) |
+| [Ollama Guide](docs/OLLAMA_GUIDE.md) | Engineers | Use with local LLMs (fully air-gapped) |
 | [CHANGELOG.md](CHANGELOG.md) | Everyone | Version history |
 | [data/README.md](data/README.md) | Everyone | Dataset documentation |
+| [skills/](skills/) | 🤖 LLM Clients | Copilot Skills — guided diagnostic workflows (bearing, screening, reporting) |
 
 ---
 
@@ -642,21 +670,21 @@ uv run mcp dev src/machinery_diagnostics_server.py
 
 ## 🚀 Roadmap
 
-### ✨ Recent: v0.2.1 — Machine Documentation Reader
+### ✨ Recent: v0.5.0 — Code Quality & Multi-Format Support
 
-AI-powered extraction of machine specifications from equipment manuals:
-- 📄 **Direct PDF Access** via MCP Resources
-- 🔍 **Smart Extraction** — Regex patterns for bearings, RPM, power ratings
-- 🧮 **Auto-Calculation** — Bearing fault frequencies from geometry (ISO 15243:2017)
-- 💾 **Caching System** — Fast repeated queries with JSON caching
+- 📂 **Multi-format signal loading** — CSV, MAT, WAV, NPY, Parquet via unified `load_signal_data()`
+- 🔧 **ML code deduplication** — 4 helper functions reduce ~163 statements
+- 📦 **pypdf migration** — Replaced deprecated PyPDF2 with pypdf
+- ▶️ **`python -m` support** — Run as `python -m predictive_maintenance_mcp`
+- 🧹 **Consolidated metadata reads** — ISO evaluation no longer double-reads metadata files
 
 ### 🔮 Planned Enhancements
 
 Each item below links to an open issue where you can **discuss, contribute, or claim the task**:
 
 | Priority | Enhancement | Status | Get Involved |
 |----------|-------------|--------|--------------|
-| 🔴 High | **Parquet/HDF5 data format support** | Open | [Good First Issue](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
+| ✅ Done | **Parquet/MAT/WAV/NPY data format support** | v0.5.0 | — |
 | 🔴 High | **Customizable ISO report thresholds** | Open | [Good First Issue](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
 | 🔴 High | **Docker image for zero-install setup** | Open | [Help Wanted](https://github.com/LGDiMaggio/predictive-maintenance-mcp/issues) |
 | 🟡 Medium | **Vector search for large documents** (ChromaDB/FAISS) | Planned | [Discuss](https://github.com/LGDiMaggio/predictive-maintenance-mcp/discussions) |
@@ -707,7 +735,7 @@ If you use this server in your research or projects, please cite:
   title = {Predictive Maintenance MCP Server: An open-source framework for integrating Large Language Models with predictive maintenance and fault diagnosis workflows},
   author = {Di Maggio, Luigi Gianpio},
   year = {2025},
-  version = {0.4.1},
+  version = {0.5.0},
   url = {https://github.com/LGDiMaggio/predictive-maintenance-mcp},
   doi = {10.5281/zenodo.17611542}
 }

SECURITY.md

@@ -4,10 +4,10 @@
 
 | Version | Supported          |
 | ------- | ------------------ |
+| 0.5.x   | :white_check_mark: |
 | 0.4.x   | :white_check_mark: |
-| 0.3.x   | :white_check_mark: |
-| 0.2.x   | :x:                |
-| < 0.2   | :x:                |
+| 0.3.x   | :x:                |
+| < 0.3   | :x:                |
 
 ## Reporting a Vulnerability
 

data/README.md

@@ -21,14 +21,14 @@ Perfect for:
 
 ## 📁 Directory Structure
 
-- **`signals/`** - Processed CSV signals ready for analysis (exposed via MCP resources)
+- **`signals/`** - Signal files ready for analysis (CSV, MAT, WAV, NPY, Parquet — exposed via MCP resources)
   - `real_train/` - Training dataset (2 healthy + 12 faulty signals)
   - `real_test/` - Test dataset for validation (1 healthy + 5 faulty signals)
 - **`real_bearings/`** - Source MAT files from MathWorks (archive only, not used by MCP server)
   - `train/` - Original MATLAB .mat files
   - `test/` - Original MATLAB .mat files
 
-> **Note**: The MCP server only uses CSV files in `signals/` directory. The `real_bearings/` folder is kept as source archive.
+> **Note**: The MCP server reads signal files from the `signals/` directory (supports CSV, MAT, WAV, NPY, Parquet). The `real_bearings/` folder is kept as source archive.
 
 ## 📊 Dataset Information