v0.4.0

What's new

New endpoint

• async_get_killmails() — fetch a character's recent killmails (requires esi-killmails.read_killmails.v1 scope)

Automatic pagination

• async_get_wallet_journal(), async_get_contacts(), and async_get_killmails() now automatically fetch all pages and return a single combined list
• No API changes required — callers always receive the complete dataset

TTL/Expires caching

• The client now respects the ESI Expires header: if a cached response has not yet expired, the HTTP request is skipped entirely and the cached data is returned immediately
• Combined with existing ETag/304 caching, this gives two layers of cache protection

Integration tests

• Integration test suite expanded to 27 tests covering all major endpoints
• New tests: test_corporation_public_live, test_killmails_live, test_etag_caching_round_trip_live, test_etag_cache_populated_after_request_live, test_expires_header_sets_ttl_in_cache_live

Documentation

• Endpoint count corrected to 23
• Authentication scope table completed (all scopes listed)
• New pagination and request caching reference sections in docs/endpoints.md and docs/quickstart.md

Stats

• 172 tests, 98.04% coverage
• pylint 10.00/10, ruff clean, mypy clean

v0.3.0

What's new

ETag/conditional-request caching

• All ESI GET endpoints now send If-None-Match on repeated calls within the same client instance
• HTTP 304 returns cached data without parsing a response body, reducing bandwidth and ESI rate-limit pressure
• Public and authenticated endpoints use separate cache keys
• New clear_etag_cache() public method to reset the in-memory cache

Bug fixes

• response.release() is now called correctly (synchronous in aiohttp ≥3.9) before returning on 304
• Response body is now drained before raising EveOnlineNotFoundError (404) and EveOnlineRateLimitError (420/429), preventing connection pool exhaustion under repeated errors

Stats

• 159 tests, 98.21% coverage
• pylint 10.00/10, ruff clean, mypy clean

v0.2.2

Bug Fixes

• Export EveOnlineNotFoundError from package root - previously only accessible via eveonline.exceptions
• Handle non-numeric Retry-After header - HTTP date values no longer crash with ValueError

Improvements

• Remove redundant .replace("Z", "+00:00") - Python >=3.11 handles Z suffix natively
• Use _parse_datetime consistently for all optional datetime fields
• Change DEFAULT_SCOPES from mutable list to immutable tuple
• Add from __future__ import annotations to all modules
• Add slots=True to all 15 frozen dataclasses for reduced memory usage
• Single-source version via importlib.metadata (no more hardcoded version string)
• Add Attributes: sections to all model docstrings (Google style)
• Document EveOnlineRateLimitError.retry_after parameter

Tests

Expanded from 91 to 115 tests (99.35% coverage):

• Non-numeric and missing Retry-After header handling
• HTTP 400 generic error
• _parse_datetime edge cases (None, Z suffix, +00:00 offset, invalid input)
• HTTP 404 on authenticated endpoint
• EveOnlineNotFoundError top-level export verification
• Parametrized __slots__ verification for all 15 models

v0.2.1

Initial public release of python-eveonline — an async Python client for the Eve Online ESI API.

Features

• Async client built on aiohttp
• Full support for Tranquility server status (public endpoint)
• Authenticated endpoints for character data: online status, location, ship, wallet, skills, skill queue, mail labels, industry jobs, market orders, jump fatigue
• Batch ID-to-name resolution via universe/names endpoint
• Abstract auth class for flexible token management (e.g. Home Assistant OAuth2)
• Typed dataclass models for all API responses
• Comprehensive exception hierarchy (auth, connection, rate limit, not found)
• Fully type-hinted with py.typed marker