calgebra 0.4.0

A tiny DSL for calendar interval algebra - compose, filter, and query time ranges using set operations.

calgebra 🗓️

A tiny DSL for merging and searching over calendar-like intervals.

Installation

pip install calgebra

# Or with Google Calendar support
pip install calgebra[google-calendar]

Quick Start

from calgebra import day_of_week, time_of_day, hours, at_tz, HOUR

# Compose time windows from primitives
weekdays = day_of_week(["monday", "tuesday", "wednesday", "thursday", "friday"])
work_hours = time_of_day(start=9*HOUR, duration=8*HOUR, tz="US/Pacific")
business_hours = weekdays & work_hours

# Union: combine busy times (in practice: from Google Calendar, databases, etc.)
monday_meetings = day_of_week("monday") & time_of_day(start=10*HOUR, duration=2*HOUR)
friday_focus = day_of_week("friday") & time_of_day(start=14*HOUR, duration=3*HOUR)
busy = monday_meetings | friday_focus

# Difference: find free time during business hours
free = business_hours - busy

# Filter: only slots >= 2 hours
long_slots = free & (hours >= 2)

# Fetch results with ergonomic date syntax
at = at_tz("US/Pacific")
meeting_options = list(long_slots[at("2025-01-01"):at("2025-01-31")])

Intervals in calgebra are inclusive of both start and end—durations therefore reflect every second covered by an interval. However, timeline slicing uses exclusive end bounds ([start:end)) to match standard Python idioms. Timeline slices accept date strings via at_tz(), timezone-aware datetime objects, or integer timestamps.

Important: Intervals are automatically clipped to your query bounds. When you slice timeline[start:end], any intervals extending beyond those bounds are trimmed to fit. This ensures aggregations like total_duration() and set operations work correctly within your query window. When you subclass Interval, define your subclass as a dataclass (ideally frozen=True) so the algebra can clone and clamp events internally.

Common helpers and aggregates are exposed alongside the core DSL:

Recurring Patterns (RFC 5545 via python-dateutil):

• recurring(freq, ...) generates intervals based on recurrence rules (weekly, bi-weekly, monthly, etc.)
• day_of_week(days, tz) convenience wrapper for filtering by day(s) of week
• time_of_day(start, duration, tz) convenience wrapper for daily time windows
• HOUR, MINUTE, DAY, SECOND constants for readable time specifications
• Compose with & to create complex patterns like business hours, recurring meetings, etc.

Aggregation & Analysis:

• flatten(timeline) converts overlapping/adjacent spans into a coalesced timeline (returns mask Intervals and must be sliced with explicit bounds)
• union(*timelines) / intersection(*timelines) offer functional set operations
• total_duration sums inclusive coverage inside a window
• max_duration / min_duration find the longest or shortest clamped intervals
• count_intervals tallies events over a slice
• coverage_ratio reports utilization as a 0–1 fraction

Transformations:

• buffer(timeline, before, after) adds buffer time around each interval (useful for travel time, setup/teardown)
• merge_within(timeline, gap) coalesces intervals separated by at most gap seconds (useful for grouping related events)

Integrations:

• calgebra.gcsa.calendars() returns ready-to-use Calendar timelines backed by the Google Calendar API (shared auth, per-calendar metadata, inclusive bounds)

→ Read the full tutorial for a complete guide to the DSL
→ API Reference for detailed function signatures and parameters

License

MIT License - see LICENSE file for details.