@tecet/ollm 0.1.4 → 0.1.5

package/docs/Context/Index.md

@@ -1,209 +0,0 @@
-# Context Management Documentation Index
-
-**Quick Reference with Links**
-
-This index provides quick navigation to all Context Management documentation with brief descriptions.
-
----
-
-## 📚 Core Documentation
-
-### [README.md](README.md)
-
-**Overview and Navigation Guide**
-
-Main entry point for Context Management documentation. Provides overview of all features, quick links, learning paths, and common use cases.
-
-**Topics:** Overview, Features, Configuration, Troubleshooting  
-**Audience:** All users
-
----
-
-### [Context Architecture](ContextArchitecture.md)
-
-**System Architecture and Design**
-
-Complete technical documentation of the Context Management system architecture. Covers VRAM monitoring, context sizing, compression service, checkpoint system, and all component interactions.
-
-**Topics:** Architecture, VRAM Detection, Context Tiers, Components, Data Flow  
-**Audience:** Developers, architects
-
-**Key Sections:**
-
-- System overview and architecture
-- VRAM monitoring (NVIDIA, AMD, Apple Silicon)
-- Context sizing and tier selection
-- Compression service architecture
-- Checkpoint system design
-- Component interactions and data flow
-
----
-
-### [Context Management](ContextManagment.md)
-
-**Context Sizing and VRAM Monitoring**
-
-Comprehensive guide to context sizing, VRAM monitoring, and context tier selection. Explains how context size is determined at startup and remains fixed for the session.
-
-**Topics:** Context Tiers, VRAM Detection, Context Sizing, Memory Management  
-**Audience:** All users
-
-**Key Sections:**
-
-- Context tiers (Minimal, Basic, Standard, Premium, Ultra)
-- VRAM detection methods
-- Context size calculation (85% utilization)
-- Fixed context sizing approach
-- Memory safety guards
-- Context usage monitoring
-
----
-
-### [Context Compression](ContextCompression.md)
-
-**Compression Strategies and Implementation**
-
-Detailed guide to the context compression system. Explains how LLM-based summarization works, compression strategies, and when compression triggers.
-
-**Topics:** Compression, Summarization, Strategies, Triggers  
-**Audience:** All users
-
-**Key Sections:**
-
-- LLM-based summarization (LLM does the work)
-- Compression strategies (aggressive, balanced, conservative)
-- Compression triggers and thresholds
-- What gets compressed vs preserved
-- Compression quality and effectiveness
-- Manual compression commands
-
----
-
-### [Checkpoint Flow Diagram](CheckpointFlowDiagram.md)
-
-**Checkpoint System and Rollover Flow**
-
-Visual guide to the checkpoint system with flow diagrams. Explains checkpoint types, creation, restoration, and session rollover process.
-
-**Topics:** Checkpoints, Rollover, Session Management, Recovery  
-**Audience:** All users
-
-**Key Sections:**
-
-- Checkpoint types (manual, automatic, rollover)
-- Checkpoint creation flow
-- Checkpoint restoration process
-- Session rollover workflow
-- Checkpoint storage and management
-- Recovery scenarios
-
----
-
-## 📖 Documentation by Topic
-
-### Context Sizing
-
-- [Context Management](ContextManagment.md) - Context tiers and sizing
-- [Context Architecture](ContextArchitecture.md#context-sizing) - Sizing implementation
-
-### VRAM Monitoring
-
-- [Context Management](ContextManagment.md#vram-monitoring) - VRAM detection
-- [Context Architecture](ContextArchitecture.md#vram-monitoring) - VRAM architecture
-
-### Compression
-
-- [Context Compression](ContextCompression.md) - Complete compression guide
-- [Context Architecture](ContextArchitecture.md#compression-service) - Compression architecture
-
-### Checkpoints
-
-- [Checkpoint Flow](CheckpointFlowDiagram.md) - Checkpoint system
-- [Context Architecture](ContextArchitecture.md#checkpoint-system) - Checkpoint architecture
-
-### Configuration
-
-- [README](README.md#configuration) - Configuration overview
-- [Context Management](ContextManagment.md#configuration) - Detailed settings
-
----
-
-## 📖 Documentation by Audience
-
-### For New Users
-
-1. [README](README.md) - Start here
-2. [Context Management](ContextManagment.md) - Understanding context
-3. [Context Compression](ContextCompression.md) - Managing context
-
-### For Regular Users
-
-1. [Checkpoint Flow](CheckpointFlowDiagram.md) - Using checkpoints
-2. [Context Compression](ContextCompression.md) - Compression strategies
-3. [README](README.md#configuration) - Configuration options
-
-### For Developers
-
-1. [Context Architecture](ContextArchitecture.md) - System design
-2. [Context Management](ContextManagment.md) - Implementation details
-3. [Context Compression](ContextCompression.md) - Compression implementation
-
----
-
-## 🔗 Related Documentation
-
-### Core Systems
-
-- [Model Management](../LLM%20Models/README.md) - Model selection
-- [Prompts System](../Prompts%20System/README.md) - System prompts
-- [User Interface](../UI&Settings/README.md) - UI documentation
-
-### Commands
-
-- [Context Commands](../UI&Settings/Commands.md#context-management) - CLI commands
-- [Snapshot Commands](../UI&Settings/Commands.md#snapshot-management) - Checkpoint commands
-
-### Developer Resources
-
-- Knowledge DB: `dev_ContextManagement.md` - Architecture details
-- Knowledge DB: `dev_ContextCompression.md` - Compression details
-
----
-
-## 📊 Documentation Status
-
-### Completed ✅
-
-| Document                 | Status      |
-| ------------------------ | ----------- |
-| README.md                | ✅ Complete |
-| Index.md                 | ✅ Complete |
-| ContextArchitecture.md   | ✅ Complete |
-| ContextManagment.md      | ✅ Complete |
-| ContextCompression.md    | ✅ Complete |
-| CheckpointFlowDiagram.md | ✅ Complete |
-
-**Overall Progress:** 100% complete (6/6 files)
-
----
-
-## 🎯 Quick Links
-
-### Common Tasks
-
-- Check context usage → [Context Commands](../UI&Settings/Commands.md#context-stats)
-- Compress context → [Context Compression](ContextCompression.md#manual-compression)
-- Create checkpoint → [Checkpoint Flow](CheckpointFlowDiagram.md#creating-checkpoints)
-- Configure context → [README](README.md#configuration)
-
-### Understanding Systems
-
-- How context sizing works → [Context Management](ContextManagment.md)
-- How compression works → [Context Compression](ContextCompression.md)
-- How checkpoints work → [Checkpoint Flow](CheckpointFlowDiagram.md)
-- System architecture → [Context Architecture](ContextArchitecture.md)
-
----
-
-**Last Updated:** January 26, 2026  
-**Version:** 0.1.0

package/docs/Context/README.md

@@ -1,390 +0,0 @@
-# Context Management Documentation
-
-**Last Updated:** January 26, 2026
-
-Welcome to the Context Management documentation for OLLM CLI. This section covers the complete context management system, including context sizing, compression, checkpoints, and VRAM monitoring.
-
----
-
-## 📚 Documentation Overview
-
-### Core Documentation
-
-- **[Context Architecture](ContextArchitecture.md)** - Complete system architecture and design
-- **[Context Management](ContextManagment.md)** - Context sizing, tiers, and VRAM monitoring
-- **[Context Compression](ContextCompression.md)** - Compression strategies and implementation
-- **[Checkpoint Flow](CheckpointFlowDiagram.md)** - Checkpoint system and rollover flow
-
----
-
-## 🎯 What is Context Management?
-
-The **Context Management** system handles the conversation history and system prompts sent to the LLM, ensuring optimal performance within hardware constraints:
-
-### 1. **Fixed Context Sizing**
-
-Context size is determined once at startup based on available VRAM:
-
-- VRAM detection (NVIDIA, AMD, Apple Silicon)
-- Context tier selection (Minimal, Basic, Standard, Premium, Ultra)
-- 85% utilization for optimal performance (pre-calculated in LLM_profiles.json)
-- Context remains FIXED for the entire session
-
-### 2. **Context Compression**
-
-Automatic compression when context approaches limits:
-
-- LLM-based summarization (LLM does the work, not the app)
-- Multiple compression strategies (aggressive, balanced, conservative)
-- Preserves recent messages and critical information
-- Maintains conversation continuity
-
-### 3. **Checkpoint System**
-
-Save and restore conversation state:
-
-- Manual checkpoints for important conversation points
-- Automatic checkpoints before major operations
-- Rollover to new session when context full
-- Checkpoint restoration for session recovery
-
-### 4. **VRAM Monitoring**
-
-Real-time GPU memory tracking:
-
-- Cross-platform support (NVIDIA, AMD, Apple Silicon)
-- Automatic VRAM detection at startup
-- Context tier selection based on available memory
-- Memory safety guards to prevent OOM errors
-
----
-
-## 📖 Documentation Structure
-
-```
-docs/Context/
-├── README.md                    ← You are here
-├── Index.md                     Quick reference with links
-├── ContextArchitecture.md       System architecture
-├── ContextManagment.md          Context sizing and VRAM
-├── ContextCompression.md        Compression system
-└── CheckpointFlowDiagram.md     Checkpoint flow
-```
-
----
-
-## 🎓 Key Concepts
-
-### Context Tiers
-
-Five fixed context tiers based on available VRAM:
-
-| Tier         | Context Size | VRAM Required | Use Case                  |
-| ------------ | ------------ | ------------- | ------------------------- |
-| **Minimal**  | 2K-4K        | <4GB          | Quick tasks, small models |
-| **Basic**    | 4K-8K        | 4-8GB         | Standard conversations    |
-| **Standard** | 8K-16K       | 8-12GB        | Extended conversations    |
-| **Premium**  | 16K-32K      | 12-24GB       | Long conversations        |
-| **Ultra**    | 32K+         | 24GB+         | Maximum context           |
-
-**See:** [Context Management](ContextManagment.md)
-
-### Context Compression
-
-When context approaches the limit, compression automatically triggers:
-
-**Process:**
-
-1. Detect context approaching limit (e.g., 90% full)
-2. Select compression strategy (aggressive, balanced, conservative)
-3. LLM summarizes older messages
-4. Replace old messages with summary
-5. Continue conversation with more space
-
-**See:** [Context Compression](ContextCompression.md)
-
-### Checkpoint System
-
-Save conversation state for recovery or rollover:
-
-**Types:**
-
-- **Manual Checkpoints** - User-created save points
-- **Automatic Checkpoints** - Created before major operations
-- **Rollover Checkpoints** - Created when starting new session
-
-**See:** [Checkpoint Flow](CheckpointFlowDiagram.md)
-
-### VRAM Monitoring
-
-Automatic GPU memory detection:
-
-**Supported Platforms:**
-
-- **NVIDIA** - nvidia-smi (Windows, Linux)
-- **AMD** - rocm-smi (Linux)
-- **Apple Silicon** - Metal API (macOS)
-
-**See:** [Context Architecture](ContextArchitecture.md)
-
----
-
-## 💡 Common Use Cases
-
-### Check Context Usage
-
-```bash
-# View current context stats
-/context stats
-
-# Shows:
-# - Current tokens used
-# - Total context size
-# - Percentage used
-# - Compression status
-```
-
-### Manual Compression
-
-```bash
-# Trigger compression manually
-/compact
-
-# Options:
-/compact aggressive    # Maximum compression
-/compact balanced      # Default compression
-/compact conservative  # Minimal compression
-```
-
-### Create Checkpoint
-
-```bash
-# Save current conversation state
-/snapshot create my-checkpoint
-
-# List checkpoints
-/snapshot list
-
-# Restore checkpoint
-/snapshot restore my-checkpoint
-```
-
-### Start New Session
-
-```bash
-# Start fresh session (creates rollover checkpoint)
-/session new
-
-# Resume previous session
-/session resume <name>
-```
-
-**Learn more:** [Context Commands](../UI&Settings/Commands.md#context-management)
-
----
-
-## 🛠️ Configuration
-
-### Context Settings
-
-```yaml
-context:
-  # Context tier (auto-detected from VRAM)
-  tier: Standard
-
-  # Enable automatic compression
-  autoCompress: true
-
-  # Compression trigger threshold
-  compressionThreshold: 0.9 # 90% full
-
-  # Compression strategy
-  compressionStrategy: balanced # aggressive, balanced, conservative
-```
-
-### VRAM Settings
-
-```yaml
-vram:
-  # Enable VRAM monitoring
-  enabled: true
-
-  # VRAM detection method
-  # auto, nvidia-smi, rocm-smi, metal
-  detectionMethod: auto
-
-  # Manual VRAM override (GB)
-  manualVRAM: null
-```
-
-### Checkpoint Settings
-
-```yaml
-checkpoints:
-  # Enable automatic checkpoints
-  autoCheckpoint: true
-
-  # Checkpoint before major operations
-  checkpointBeforeOperations: true
-
-  # Maximum checkpoints to keep
-  maxCheckpoints: 10
-```
-
-**Learn more:** [Configuration](../UI&Settings/Configuration.md)
-
----
-
-## 🔍 Troubleshooting
-
-### Common Issues
-
-**Context fills up too quickly:**
-
-- Enable automatic compression: `/config set autoCompress true`
-- Use more aggressive compression: `/compact aggressive`
-- Create checkpoint and start new session: `/snapshot create` then `/session new`
-
-**VRAM detection not working:**
-
-- Check GPU drivers installed
-- Verify nvidia-smi/rocm-smi in PATH
-- Manual override: `/config set vram.manualVRAM 8` (for 8GB)
-
-**Compression not triggering:**
-
-- Check compression enabled: `/config get autoCompress`
-- Check threshold: `/config get compressionThreshold`
-- Manually trigger: `/compact`
-
-**Checkpoint restore fails:**
-
-- Verify checkpoint exists: `/snapshot list`
-- Check checkpoint file permissions
-- Try creating new checkpoint: `/snapshot create test`
-
-**See:** [Troubleshooting Guide](../Troubleshooting.md)
-
----
-
-## 📊 Implementation Status
-
-### Current (v0.1.0)
-
-- ✅ VRAM Detection (NVIDIA, AMD, Apple Silicon)
-- ✅ Fixed Context Sizing
-- ✅ Context Tiers (Minimal, Basic, Standard, Premium, Ultra)
-- ✅ LLM-based Compression
-- ✅ Checkpoint System
-- ✅ Context Usage Display
-- ✅ Memory Safety Guards
-
-### Planned (v0.2.0)
-
-- ⏳ Advanced Compression Strategies
-- ⏳ Context Analytics
-- ⏳ Compression Quality Metrics
-
-### Planned (v0.3.0)
-
-- ⏳ Multi-tier Compression
-- ⏳ Semantic Context Pruning
-- ⏳ Context Optimization
-
----
-
-## 🤝 Related Documentation
-
-### Core Systems
-
-- [Model Management](../LLM%20Models/README.md) - Model selection and configuration
-- [Prompts System](../Prompts%20System/README.md) - System prompts and templates
-- [User Interface](../UI&Settings/README.md) - UI documentation
-
-### Developer Resources
-
-- Knowledge DB: `dev_ContextManagement.md` - Context management architecture
-- Knowledge DB: `dev_ContextCompression.md` - Compression implementation
-
----
-
-## 🎯 Quick Start
-
-### For New Users
-
-1. **Check Your Context**
-
-   ```bash
-   /context stats
-   ```
-
-2. **Enable Auto-Compression**
-
-   ```bash
-   /config set autoCompress true
-   ```
-
-3. **Create Your First Checkpoint**
-   ```bash
-   /snapshot create my-first-checkpoint
-   ```
-
-### For Advanced Users
-
-1. **Optimize Compression**
-
-   ```bash
-   /config set compressionStrategy aggressive
-   /config set compressionThreshold 0.85
-   ```
-
-2. **Monitor VRAM**
-
-   ```bash
-   /config get vram
-   ```
-
-3. **Manage Checkpoints**
-   ```bash
-   /snapshot list
-   /snapshot restore <name>
-   /snapshot delete <name>
-   ```
-
----
-
-## 📈 Performance Tips
-
-### Maximize Context Efficiency
-
-1. **Use Appropriate Tier** - Don't force larger context than needed
-2. **Enable Auto-Compression** - Let system manage context automatically
-3. **Create Checkpoints** - Save important conversation points
-4. **Start Fresh Sessions** - Use rollover for very long conversations
-5. **Monitor Usage** - Check `/context stats` regularly
-
-### Optimize for Your Hardware
-
-**Low VRAM (<4GB):**
-
-- Use Minimal tier (2K-4K)
-- Aggressive compression
-- Frequent checkpoints
-
-**Medium VRAM (4-8GB):**
-
-- Use Basic tier (4K-8K)
-- Balanced compression
-- Regular checkpoints
-
-**High VRAM (8GB+):**
-
-- Use Standard/Premium tier (8K-32K)
-- Conservative compression
-- Occasional checkpoints
-
----
-
-**Last Updated:** January 26, 2026  
-**Version:** 0.1.0  
-**Status:** Active Development