refactor!: rename ProviderNotRegisteredError.provider_type to dependency_type, and exception group keywords to exceptions (#614)

[lesnik512]

Closes #614.

Summary

• ProviderNotRegisteredError.provider_type is now .dependency_type, and so is its constructor keyword. The message now reads No provider is registered for B., matching the troubleshooting page.
• FinalizerError(finalizer_errors=) and ValidationFailedError(errors=) now take exceptions=, the name of the attribute that holds them.
• modern_di.types.P is deleted. Nothing used it.
• Migration guide: new entries for the three renames and the new message, the creator= that ArgumentResolutionError now requires, and a "Only four modules are public" section covering the deleted cache_registry / context_registry modules, the helpers removed from modern_di.scope, and types.P.
• Stale text: "context registry" and "cache registry" wording in docstrings and docs, the CacheRegistry.fetch_cache_item reference in the performance history, the InvalidChildScopeError note about a constructor parent, the internal register named on the lifecycle page, the container page H1 (now "Container", as in the nav), its explicit-injection example (the parameter is no longer annotated Container), and the factories.md finalizer lambda that failed pyright (now a def).

Design decisions

• The type in the message is rendered by a new type_name helper in exceptions/rendering.py: a plain class, NewType or type alias shows its bare name; a parameterized type such as list[int] shows its repr, because its __name__ would print list.
• DuplicateProviderTypeError.provider_type keeps its name. It holds a bound type, which is what the issue contrasts it with.
• CacheSettings.clear_cache keeps its name, as decided in the issue.
• The migration entry for the stricter validate() from validate() bugs: Container under a custom scope enum, reference-only providers, Alias scope #612 is left to the validate() bugs: Container under a custom scope enum, reference-only providers, Alias scope #612 PR.
• The keyword rename touches one line each in cache.py and providers_registry.py; the other changes in those files belong to parallel PRs.

Test plan

• Tests updated first and seen failing: new attribute and keyword names, the new message text, and a full-string assert for the typo suggestion (the old one avoided the <locals> repr)
• New test: the message names list[int] and a NewType correctly
• just lint, just lint-ci
• just test-ci (100% coverage)
• mkdocs build --strict
• The edited container and factories doc samples run and pass pyright

[github-actions]

Benchmark

Details

Benchmark suite	Current: 6e24334	Previous: f6e343d	Ratio
benchmarks/test_guard_by_type.py::test_g16_resolve_by_type	2876876.7521229764 iter/sec (stddev: 1.1696033830753136e-8)	2921718.256656569 iter/sec (stddev: 1.675866103278825e-8)	1.02
benchmarks/test_guard_by_type.py::test_g17_resolve_by_type_large_registry	2873344.0020654784 iter/sec (stddev: 1.0792850543847373e-8)	2939203.314248642 iter/sec (stddev: 1.204015671986658e-8)	1.02
benchmarks/test_guard_cold.py::test_g8_cold_first_resolve	21337.417788597268 iter/sec (stddev: 0.00004319748084508497)	21762.967229655445 iter/sec (stddev: 0.00004094890871011781)	1.02
benchmarks/test_guard_cold.py::test_g8b_cold_first_resolve_cached	16602.147921189095 iter/sec (stddev: 0.00011596111660614653)	16433.6760366631 iter/sec (stddev: 0.00010812780919654217)	0.99
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[1]	339.74360636823667 iter/sec (stddev: 0.00004409823041302261)	461.45135894055323 iter/sec (stddev: 0.00003990784243928985)	1.36
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[2]	340.14769253060797 iter/sec (stddev: 0.0000848787600443331)	451.0026376153234 iter/sec (stddev: 0.000054903965874733085)	1.33
benchmarks/test_guard_concurrency.py::test_g14_concurrent_cached_hit[4]	357.1023656420849 iter/sec (stddev: 0.00014472413870029096)	425.7675844707922 iter/sec (stddev: 0.00008308038620630768)	1.19
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[1]	12905.503003303023 iter/sec (stddev: 0.000007207691697000248)	12751.290749215299 iter/sec (stddev: 0.00000819216963265991)	0.99
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[2]	6409.680155344794 iter/sec (stddev: 0.000024502105748011574)	6401.705755814795 iter/sec (stddev: 0.000022867724842553285)	1.00
benchmarks/test_guard_concurrency.py::test_g15_concurrent_first_resolve[4]	3344.018121458962 iter/sec (stddev: 0.00004400377315312043)	3423.3858792657397 iter/sec (stddev: 0.000040939280354384506)	1.02
benchmarks/test_guard_concurrency.py::test_g15b_concurrent_first_resolve_sibling_children[1]	8239.983476082682 iter/sec (stddev: 0.000010742401355025445)	7612.8302325882405 iter/sec (stddev: 0.000010256134066683693)	0.92
benchmarks/test_guard_concurrency.py::test_g15b_concurrent_first_resolve_sibling_children[2]	3527.312495050392 iter/sec (stddev: 0.000020281655325898572)	3618.370685087365 iter/sec (stddev: 0.00002517012025414019)	1.03
benchmarks/test_guard_concurrency.py::test_g15b_concurrent_first_resolve_sibling_children[4]	1715.014227829269 iter/sec (stddev: 0.00004653093307089315)	1719.2318706952121 iter/sec (stddev: 0.00004854649644947838)	1.00
benchmarks/test_guard_concurrency.py::test_g15c_worker_pool_floor_control[1]	40502.52837066591 iter/sec (stddev: 0.000005362763750588661)	32996.42153964083 iter/sec (stddev: 0.000006880551983878211)	0.81
benchmarks/test_guard_concurrency.py::test_g15c_worker_pool_floor_control[2]	13958.854415514363 iter/sec (stddev: 0.000014119539297385203)	14659.535942872064 iter/sec (stddev: 0.000012750288363749525)	1.05
benchmarks/test_guard_concurrency.py::test_g15c_worker_pool_floor_control[4]	5919.772578139759 iter/sec (stddev: 0.00003551057028411714)	5849.76259472311 iter/sec (stddev: 0.0000335570564798276)	0.99
benchmarks/test_guard_lifecycle.py::test_g6_build_child_container	1628864.1238215764 iter/sec (stddev: 3.169392048789073e-8)	1585121.4791400717 iter/sec (stddev: 3.581802492697611e-8)	0.97
benchmarks/test_guard_lifecycle.py::test_g6b_build_child_container_auto_scope	1577316.983020222 iter/sec (stddev: 2.4889369271199436e-8)	1634928.8822412132 iter/sec (stddev: 2.2739949474734155e-8)	1.04
benchmarks/test_guard_lifecycle.py::test_g7_request_lifecycle_batch	2805.286465082796 iter/sec (stddev: 0.000013069684558158616)	2851.9403996074443 iter/sec (stddev: 0.00001246850076439463)	1.02
benchmarks/test_guard_lifecycle.py::test_g7c_event_loop_floor_control	57610.72882387156 iter/sec (stddev: 0.000002558145117166013)	58007.21870000694 iter/sec (stddev: 0.00000208313451377107)	1.01
benchmarks/test_guard_lifecycle.py::test_g7b_request_cycle_sync	344700.5046398797 iter/sec (stddev: 1.4150320288726064e-7)	332497.14734137605 iter/sec (stddev: 7.44393106694809e-8)	0.96
benchmarks/test_guard_lifecycle.py::test_g13_teardown_at_scale	48879.97919771222 iter/sec (stddev: 0.0000026061143853176474)	48889.69298560324 iter/sec (stddev: 0.000001820471012608131)	1.00
benchmarks/test_guard_lifecycle.py::test_g13b_teardown_at_scale_async_no_finalizers	460.741650330517 iter/sec (stddev: 0.001109702313569377)	472.1940049135932 iter/sec (stddev: 0.001026790425194055)	1.02
benchmarks/test_guard_resolve.py::test_g1_transient_resolve	1985595.8917150293 iter/sec (stddev: 4.91556318072766e-8)	2178750.936263875 iter/sec (stddev: 2.6353327029762944e-8)	1.10
benchmarks/test_guard_resolve.py::test_g2_cached_resolve	2772472.805092418 iter/sec (stddev: 1.0591957006213357e-8)	3311376.0403177035 iter/sec (stddev: 2.072399608841274e-8)	1.19
benchmarks/test_guard_resolve.py::test_g3_deep_chain	707343.6541965557 iter/sec (stddev: 4.23623889972603e-8)	678836.554249412 iter/sec (stddev: 4.6008684988825406e-8)	0.96
benchmarks/test_guard_resolve.py::test_g4_wide_resolve	444648.4096104095 iter/sec (stddev: 4.038849480355304e-7)	436782.78901172994 iter/sec (stddev: 3.7702362683787223e-7)	0.98
benchmarks/test_guard_resolve.py::test_g5_cross_scope	1723795.846715483 iter/sec (stddev: 2.2071017689722112e-8)	1876457.2449674914 iter/sec (stddev: 3.702446327477425e-8)	1.09
benchmarks/test_guard_resolve.py::test_g9_context_resolve	1309741.741194973 iter/sec (stddev: 1.555277413482454e-7)	1310469.1843290261 iter/sec (stddev: 1.5759486009534166e-7)	1.00
benchmarks/test_guard_resolve.py::test_g12_override_active_resolve	730129.231779183 iter/sec (stddev: 4.7694322862630576e-8)	715658.0140887419 iter/sec (stddev: 4.497059959957813e-8)	0.98
benchmarks/test_guard_resolve.py::test_g18_alias_hop	2749724.4776143027 iter/sec (stddev: 1.0229654417429118e-8)	2903717.0685449257 iter/sec (stddev: 8.479702107212049e-9)	1.06
benchmarks/test_guard_validate.py::test_g10_validate_deep_chain	27449.318470454247 iter/sec (stddev: 0.000021319509854083832)	27290.39780542258 iter/sec (stddev: 0.00002305556618279706)	0.99
benchmarks/test_guard_validate.py::test_g11_validate_wide	16462.904203780417 iter/sec (stddev: 0.000025955323870805172)	16384.134733354367 iter/sec (stddev: 0.000027709419502202343)	1.00

This comment was automatically generated by workflow using github-action-benchmark.