What's New in 0.5.0

Scatter Plot Support

• New --scatter flag renders scatter plots in the terminal
• New --marker CHAR option to customize the scatter marker character (default: •)
• --plot-type scatter also accepted as the long form
• Scatter follows the same mutually exclusive group as --line and --hist

Quality Fixes

• Fixed Pandas4Warning deprecation: JSON output now uses date_format="iso" instead of the deprecated "epoch" default
• Updated README with scatter documentation, usage examples, and corrected install command

Test Coverage

• Added unit test for --scatter flag acceptance
• Added TestPlotTypeOptions integration test class covering --scatter and --line with both --json and --csv output modes
• Test suite: 52 tests, all passing with zero warnings

Install

uv tool install git+https://github.com/glentner/plot-cli

Usage

plot data.csv --scatter -x x -y y
plot data.csv --scatter --marker x -x x -y y

Full Changelog: 0.4.0...0.5.0

What's New in 0.4.0

This release is a major overhaul of the internals, replacing the old data loading pipeline with a DuckDB-based query engine and significantly expanding time-series capabilities.

Build System

• Migrated from Poetry to uv/hatchling (PEP 621 [project] table)
• Added duckdb>=1.0.0 as a core dependency
• Added pyarrow>=15.0.0 for Parquet support
• Replaced poetry.lock with uv.lock
• Install via uv tool install plot-cli

DuckDB Query Engine

The old provider.py / data.py pipeline has been replaced with a new query.py module built around QueryBuilder, which uses DuckDB for all data loading and transformation:

• Load CSV, JSON, and Parquet files (auto-detected or explicit with --format)
• Read from stdin in CSV or JSON format
• --where EXPR — SQL WHERE clause filtering
• --after DATETIME / --before DATETIME — datetime range filters on the x-column
• -S/--scale OFFSET — datetime offset conversion
• Column names are safely quoted in all generated SQL

Time-Series Aggregation (-B/--bucket)

• -B/--bucket INTERVAL — group data into time buckets (e.g. 15min, 1h, 1d)
• Aggregation methods: --mean, --sum, --min, --max, --first, --last
• --count — count rows per bucket; works without a -y column, eliminating the need for manual awk injection
• -F/--resample is now deprecated in favor of -B/--bucket

Multi-Series Grouping (--by)

• --by COLUMN — group data by a categorical column and plot each group as a separate series
• Implemented via a DuckDB PIVOT query (long → wide format)
• NaN values in line plots are handled gracefully
• Works with or without time bucketing

Smart Time-Series Axis Labels

• Time-series plots (-T) now show formatted HH:MM tick labels
• Adaptive tick generation based on detected time granularity
• Secondary date context labels below the time axis
• UTC used consistently for timezone handling

Output Modes

• --csv — output the processed/aggregated data as CSV instead of plotting
• --json — output the processed/aggregated data as JSON instead of plotting

Tests

• New tests/test_query.py with comprehensive QueryBuilder unit tests
• New tests/test_cli.py with CLI integration tests covering all major options

Timeseries data with tz info was broken. Uses proper to_datetime method now.