Chroma MCP Server

#!/bin/bash # Wrapper script to run the promote-learning command via hatch # Exit on error set -e # Determine script location and project root SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )" PROJECT_ROOT="$( cd "$SCRIPT_DIR/.." && pwd )" echo "Running promote-learning in $PROJECT_ROOT..." # Documentation on available parameters # --description: Natural language description of the learning (required) # --pattern: Core pattern identified (required) # --code-ref: Code reference illustrating the learning (required) # --tags: Comma-separated tags for categorization (required) # --confidence: Confidence score between 0.0 and 1.0 (required) # --source-chat-id: ID of the source entry in the chat history collection # --collection-name: Target collection for the new learning (default: derived_learnings_v1) # --chat-collection-name: Source chat history collection (default: chat_history_v1) # --include-chat-context: Include rich context from the source chat entry (default: true) # --no-include-chat-context: Don't include rich context from the source chat entry # Change to project root directory cd "${PROJECT_ROOT}" # Execute the chroma-mcp-client promote-learning subcommand using hatch # Pass all arguments received by this script to the subcommand # Use python -m syntax for robustness with hatch if needed hatch run chroma-mcp-client promote-learning "$@" # Alternatively: hatch run python -m chroma_mcp_client.cli promote-learning "$@" exit_code=$? if [ $exit_code -ne 0 ]; then echo "Error: promote-learning failed with exit code $exit_code" exit $exit_code else echo "promote-learning finished successfully." fi exit 0