uv support and rag, fastmode toggle

Author: Shashikant86

CHANGELOG.md

@@ -9,7 +9,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 - Planned:
-  - **Plan / Code modes** in interactive CLI (explicit “planning” vs “coding” flows for complex tasks).
+  - **Plan / Code modes** in interactive CLI (explicit "planning" vs "coding" flows for complex tasks).
   - First‑class support for **open‑source models via third‑party providers** (e.g. OpenRouter, Groq and similar gateways), alongside existing Ollama + cloud integrations.
 
 ### Changed
@@ -20,87 +20,40 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ---
 
-## [0.1.0] - 2025-11-26
+## [0.1.1] - 2025-11-27
+
+### Added
+- **UV Support**: Full support for `uv` as an alternative to `python -m venv` for creating virtual environments. Documentation updated to recommend `uv` as the primary method.
+- **Performance Toggles**: New `/fast-mode [on|off]`, `/disable-rag`, and `/enable-rag` commands for controlling RAG indexing and response speed. Performance settings now visible in welcome screen and `/status` command.
+- **Venv Detection**: Automatic detection of virtual environment in project root with startup warnings if missing.
+
+### Changed
+- Welcome screen now displays RAG Mode and Fast Mode status with context-aware tips.
+- Code execution prefers Python from project's `.venv/bin/python` when available.
+- Documentation updated to recommend `uv` as the primary installation method.
+
+---
 
-### Overview
-- **First public release** of DSPy Code: an AI-powered, interactive development and optimization assistant for DSPy (think "Claude Code for DSPy").
+## [0.1.0] - 2025-11-26
 
 ### Added
-- **Interactive CLI & Workflows**
-  - Rich TUI with animated thinking indicators, status panels, and history-aware prompts.
-  - Fully conversational flow: describe what you want in natural language, get DSPy code, ask follow‑ups.
-  - Two core workflows:
-    - **Development**: `/init` → describe task → generate → `/validate` → `/run` → iterate.
-    - **Optimization**: `/data` → `/optimize` → `/eval` → `/export`.
-- **Model Connection & Providers**
-  - Support for local **Ollama** models.
-  - Cloud providers: **OpenAI**, **Anthropic (Claude)**, **Google Gemini**.
-  - New interactive `/model` command:
-    - Auto-detect and list Ollama models, pick by number.
-    - Cloud flow to pick provider, then type model name (e.g. `gpt-5-nano`, `claude-sonnet-4.5`, `gemini-2.5-flash`).
-  - Direct `/connect <provider> <model>` command for advanced users.
-  - `/models`, `/status`, `/disconnect` for model management.
-- **LLM Integration & SDK Support**
-  - OpenAI integration compatible with `openai>=2.x` chat completions API.
-  - Anthropic integration via the current `anthropic` Python SDK.
-  - Gemini integration via the official `google-genai` SDK, with fallback to `google-generativeai` when present.
-  - Local Ollama integration with configurable HTTP timeouts for large models (`OLLAMA_HTTP_TIMEOUT`, `OLLAMA_TEST_TIMEOUT`).
-  - Optional extras in `pyproject.toml`:
-    - `dspy-code[openai]`, `dspy-code[anthropic]`, `dspy-code[gemini]`, `dspy-code[llm-all]`.
-- **DSPy-Aware Code Generation**
-  - Natural language → DSPy **Signatures**, **Modules**, and full **Programs**.
-  - Support for major DSPy patterns: predictors, ChainOfThought, ReAct, RAG, etc.
-  - Templates and examples for:
-    - RAG systems
-    - Question answering
-    - Classification (e.g. sentiment analyzer)
-    - Optimization/evaluation workflows
-- **Validation & Execution**
-  - `/validate` to check generated code against DSPy best practices and structure.
-  - `/run` and `/test` to execute and test generated programs within a sandboxed engine.
-  - Validation support for signatures, modules, predictors, adapters, metrics, and anti‑patterns.
-- **GEPA Optimization**
-  - End‑to‑end optimization workflows:
-    - `/optimize` for one‑shot optimization scripts.
-    - `/optimize-start`, `/optimize-status`, `/optimize-resume`, `/optimize-cancel` for long‑running GEPA jobs.
-  - Integration with evaluation metrics (Accuracy, F1, ROUGE, BLEU, etc.).
-  - Documentation and warnings about cloud costs and recommended hardware (32 GB RAM for heavy local optimization).
-- **MCP (Model Context Protocol) Integration**
-  - Built‑in MCP client with commands:
-    - `/mcp-connect`, `/mcp-disconnect`, `/mcp-servers`, `/mcp-tools`, `/mcp-call`, `/mcp-resources`, `/mcp-read`, `/mcp-prompts`, `/mcp-prompt`.
-  - Enables connecting DSPy Code to external tools and data sources.
-- **Project & Session Management**
-  - `/init` and `/project` for initializing and inspecting DSPy projects.
-  - Codebase indexing and RAG support for answering questions about your own code.
-  - Session management commands: `/sessions`, `/session`, `/history`, `/clear`, `/save-data`, export/import.
-  - Export/import of sessions and packages for deployment via `/export` and `/import`.
-- **Documentation & Examples**
-  - Full docs site (MkDocs Material) with:
-    - Getting Started (installation, quick start, first program, understanding DSPy Code).
-    - Guides (model connection, interactive mode, natural language commands, optimization, project management, validation, slash commands).
-    - Tutorials (RAG system, question answering, sentiment analyzer, GEPA optimization).
-    - Reference (commands, configuration, templates, troubleshooting, FAQ, security).
-  - Homepage and README aligned around DSPy Code as:
-    - **Development assistant** (build DSPy apps quickly).
-    - **Optimization engine** (real GEPA).
-    - **Learning environment** for DSPy concepts.
+- **Interactive CLI**: Rich TUI with natural language interface for generating DSPy Signatures, Modules, and Programs. Core workflows: development (`/init` → generate → `/validate` → `/run`) and optimization (`/data` → `/optimize` → `/eval`).
+- **Model Support**: Local Ollama models and cloud providers (OpenAI, Anthropic, Gemini) with interactive `/model` command for easy connection. SDK support via optional extras: `dspy-code[openai]`, `dspy-code[anthropic]`, `dspy-code[gemini]`, `dspy-code[llm-all]`.
+- **Code Generation**: Natural language to DSPy code with support for major patterns (ChainOfThought, ReAct, RAG, etc.) and templates for common use cases.
+- **Validation & Execution**: `/validate` for code checks, `/run` and `/test` for sandboxed execution.
+- **GEPA Optimization**: End-to-end optimization workflows with `/optimize` commands and evaluation metrics integration.
+- **MCP Integration**: Built-in MCP client for connecting to external tools and data sources.
+- **Project Management**: `/init`, codebase indexing, RAG support, session management, and export/import functionality.
+- **Documentation**: Complete docs site (MkDocs Material) with getting started guides, tutorials, and reference documentation.
 
 ### Changed
-- Default Ollama generation timeout increased to 120 seconds to better support large models.
-- Examples across README and docs updated to use modern models (e.g. `gpt-5-nano`, `claude-sonnet-4.5`, `gemini-2.5-flash`, `gpt-oss:120b`) and to recommend `/model` as the primary way to connect.
-- Quick Start and model‑connection docs now make model connection mandatory and show clear virtual‑env + provider‑SDK installation flows using `dspy-code[...]` extras and `uv`/`pip`.
-- Interactive UI improved with modern Rich versions and a `DSPY_CODE_SIMPLE_UI` mode for environments with limited emoji/spinner support.
-- Natural language intent routing in interactive mode refined to:
-  - Prefer natural‑language answers for questions.
-  - Avoid double code generation and incorrect `/explain` follow‑ups.
-- MkDocs navigation configuration tuned (tabs, sections) to keep the left nav stable and highlight the active page correctly.
+- Default Ollama timeout increased to 120 seconds for large models.
+- Examples updated to use modern models (`gpt-5-nano`, `claude-sonnet-4.5`, `gemini-2.5-flash`).
+- Interactive UI improved with Rich library and `DSPY_CODE_SIMPLE_UI` mode for limited emoji support.
+- Natural language routing refined to prefer answers for questions and avoid duplicate code generation.
 
 ### Fixed
-- OpenAI deprecation issues (`APIRemovedInV1`) by migrating from `ChatCompletion` to the new client API, and removing unsupported `max_tokens`/`temperature` parameters for models like `gpt-5-nano`.
-- Interactive mode errors:
-  - `name 'explanations' is not defined` during `/explain`.
-  - Syntax errors in `nl_command_router` debug logging.
-- Ollama timeout handling for large models, with clearer error messages on connection/generation failures.
-- Documentation glitches:
-  - Stray `\n` in callouts.
-  - Navigation behavior that caused pages to disappear or not highlight correctly.
+- OpenAI SDK migration to new client API, removed unsupported parameters for newer models.
+- Interactive mode errors (`name 'explanations' is not defined`, syntax errors).
+- Ollama timeout handling and error messages.
+- Documentation formatting and navigation issues.

docs/getting-started/installation.md

@@ -31,18 +31,38 @@ cd my-dspy-project
 
 ### Step 2: Create Virtual Environment IN This Directory
 
-```bash
-# Create .venv INSIDE your project directory (not elsewhere!)
-python -m venv .venv
+=== "uv (Recommended)"
 
-# Activate it
-# For bash/zsh (macOS/Linux):
-source .venv/bin/activate
-# For fish shell:
-source .venv/bin/activate.fish
-# On Windows:
-.venv\Scripts\activate
-```
+    ```bash
+    # Create .venv INSIDE your project directory (not elsewhere!)
+    uv venv
+
+    # Activate it
+    # For bash/zsh (macOS/Linux):
+    source .venv/bin/activate
+    # For fish shell:
+    source .venv/bin/activate.fish
+    # On Windows:
+    .venv\Scripts\activate
+    ```
+
+=== "python -m venv"
+
+    ```bash
+    # Create .venv INSIDE your project directory (not elsewhere!)
+    python -m venv .venv
+
+    # Activate it
+    # For bash/zsh (macOS/Linux):
+    source .venv/bin/activate
+    # For fish shell:
+    source .venv/bin/activate.fish
+    # On Windows:
+    .venv\Scripts\activate
+    ```
+
+!!! tip "Why uv?"
+    `uv` is a fast Python package manager written in Rust. It's 10-100x faster than pip and provides better dependency resolution. [Learn more about uv](https://docs.astral.sh/uv/)
 
 !!! success "Why .venv in the Project?"
     When you create the virtual environment inside your project:
@@ -55,39 +75,42 @@ source .venv/bin/activate.fish
 
 ### Step 3: Install DSPy Code
 
-=== "pip"
+=== "uv (Recommended)"
 
     ```bash
     # This installs into .venv/ in your project
-    pip install --upgrade dspy-code
-    ```
-
-=== "uv"
-
-    ```bash
-    # If you use uv, you can install dspy-code like this
     uv pip install --upgrade dspy-code
 
     # Or add it to your project dependencies (pyproject.toml) in one step
     uv add dspy-code
     ```
 
+=== "pip"
+
+    ```bash
+    # This installs into .venv/ in your project
+    pip install --upgrade dspy-code
+    ```
+
 That's it! DSPy Code is now installed in your project.
 
 ### Step 4: Install DSPy (Optional)
 
 DSPy Code will install DSPy automatically if needed, but you can install/upgrade it explicitly:
 
-=== "pip"
+!!! tip "Use the same tool you used for venv"
+    If you created your venv with `uv venv`, use `uv pip install` for consistency. If you used `python -m venv`, use `pip install`.
+
+=== "uv (Recommended)"
 
     ```bash
-    pip install --upgrade dspy
+    uv pip install --upgrade dspy
     ```
 
-=== "uv"
+=== "pip"
 
     ```bash
-    uv pip install --upgrade dspy
+    pip install --upgrade dspy
     ```
 
 !!! info "DSPy Version"
@@ -156,38 +179,41 @@ DSPy Code has optional dependencies for different features. Install only what yo
 
 ### Cloud Model Providers (via dspy-code extras)
 
-Use extras so versions stay aligned with dspy-code’s tested matrix.
+Use extras so versions stay aligned with dspy-code's tested matrix.
 
-=== "pip"
+!!! tip "Use the same tool you used for venv"
+    If you created your venv with `uv venv`, use `uv pip install` for consistency. If you used `python -m venv`, use `pip install`.
+
+=== "uv (Recommended)"
 
     ```bash
     # OpenAI support
-    pip install "dspy-code[openai]"
+    uv pip install "dspy-code[openai]"
 
     # Google Gemini support
-    pip install "dspy-code[gemini]"
+    uv pip install "dspy-code[gemini]"
 
     # Anthropic (paid key required)
-    pip install "dspy-code[anthropic]"
+    uv pip install "dspy-code[anthropic]"
 
     # Or install all cloud providers at once
-    pip install "dspy-code[llm-all]"
+    uv pip install "dspy-code[llm-all]"
     ```
 
-=== "uv"
+=== "pip"
 
     ```bash
     # OpenAI support
-    uv pip install "dspy-code[openai]"
+    pip install "dspy-code[openai]"
 
     # Google Gemini support
-    uv pip install "dspy-code[gemini]"
+    pip install "dspy-code[gemini]"
 
     # Anthropic (paid key required)
-    uv pip install "dspy-code[anthropic]"
+    pip install "dspy-code[anthropic]"
 
     # Or install all cloud providers at once
-    uv pip install "dspy-code[llm-all]"
+    pip install "dspy-code[llm-all]"
     ```
 
 > **Note:** Anthropic has discontinued free API keys. DSPy Code fully supports Claude **if you already have a paid API key**, but Anthropic integration will simply not work without one.

docs/getting-started/quick-start.md

@@ -12,7 +12,11 @@ mkdir my-dspy-project
 cd my-dspy-project
 
 # 2. Create a virtual environment INSIDE this directory
-python -m venv .venv
+# Recommended: Use uv (faster and more reliable)
+uv venv
+
+# Alternative: Use standard Python venv
+# python -m venv .venv
 
 # 3. Activate the virtual environment
 # For bash/zsh (macOS/Linux):
@@ -23,23 +27,27 @@ source .venv/bin/activate.fish
 .venv\Scripts\activate
 
 # 4. Install dspy-code (always upgrade for latest features)
-pip install --upgrade dspy-code
+# Recommended: Use uv
+uv pip install --upgrade dspy-code
+
+# Alternative: Use pip
+# pip install --upgrade dspy-code
 
 # 5. (Optional but recommended) Install provider SDKs via dspy-code extras
 
 # If you ONLY want local models (Ollama), you can skip this step.
 
-# OpenAI support
-pip install "dspy-code[openai]"
-
-# Google Gemini support
-pip install "dspy-code[gemini]"
-
-# Anthropic (paid key required)
-pip install "dspy-code[anthropic]"
-
-# Or install all cloud providers at once
-pip install "dspy-code[llm-all]"
+# If using uv (recommended):
+uv pip install "dspy-code[openai]"      # OpenAI support
+uv pip install "dspy-code[gemini]"     # Google Gemini support
+uv pip install "dspy-code[anthropic]"  # Anthropic (paid key required)
+uv pip install "dspy-code[llm-all]"   # Or install all cloud providers at once
+
+# If using pip:
+# pip install "dspy-code[openai]"
+# pip install "dspy-code[gemini]"
+# pip install "dspy-code[anthropic]"
+# pip install "dspy-code[llm-all]"
 ```
 
 For more details, see the [Installation Guide](installation.md).