v1.4.6

Restores MCP Python SDK v1.28+ compatibility while retaining v2 and future-version coverage. The scheduled compatibility workflow now tests the latest v1.x and latest stable SDK.

v1.4.5

• Upgrade to MCP Python SDK v2.\n- Add a twice-monthly compatibility check against the latest stable MCP SDK, opening a deduplicated issue mentioning @dl2080 on failure.

v1.4.4

Changed

• direct CLI now unwraps embedded resources into real stdout/file payloads instead of printing Python reprs
• added --output support for historical exports and plots in direct CLI mode

v1.4.3

Fixed

• historical multi-device exports and plots now skip devices that return no history instead of failing the whole request

v1.4.2

Changed

• historical export and plot requests now combine all matching devices into one response instead of emitting one artifact per device
• plot guidance now prefers png unless another output format is explicitly requested
• dependency floor now requires airq-mcp-timeseries>=0.1.7

Fixed

• historical exports keep sensor units like ppm for CO2

v1.4.1

• Applied ruff formatting to src/mcp_airq_cloud/tools/read.py so the CI format check passes again.
• Updated CLI tests to match the current tool argument mapping.
• No functional tool behavior changes.

v1.4.0

Full Changelog: v1.3.1...v1.4.0

v1.3.1

Highlights

• Formatting-only follow-up release for the new CLI work.
• Applies ruff format to the new direct CLI files so GitHub Actions passes again.
• No functional changes compared to v1.3.0.

v1.3.0

Highlights

• mcp-airq-cloud now works both as an MCP server and as a direct CLI.
• All text outputs are pipeline-friendly: data goes to stdout, runtime errors go to stderr, and failures return exit code 1.
• Added --compact-json, --json, and --yaml for easier shell integration with tools like jq and yq.
• plot-air-quality-history can now stream output directly to stdout via --output -, enabling shell redirects like > co2.png.
• Added mcp-airq-cloud serve to force stdio MCP mode from an interactive terminal.

v1.2.0

What's changed

New features

• plot_air_quality_history — new tool that fetches sensor history from the cloud and renders it as a chart
  • Default output: PNG via matplotlib (no browser required), returned as vision tokens
  • Optional: HTML for interactive Plotly charts with dark/light toggle
  • LLM can set title, x_axis_title, y_axis_title in any language at call time
  • Parameters: sensor, device, last_hours, from_datetime, to_datetime, title, x_axis_title, y_axis_title, chart_type, dark, output_format, max_points

Dependencies

• Added airq-mcp-timeseries>=0.1.3 (replaces Kaleido with matplotlib for static images)