[mem_cache][10/N] refactor: drop the redundant _component suffix in unified_cache/components

[alphabetc1]

Motivation

Part of #25371. Mechanical Move.

#32484 moved unified_cache_components/ to unified_cache/components/ but kept the
_component suffix on every file, so the directory now reads
unified_cache/components/full_component.py -- the role is stated twice, once by the
directory and again by every file inside it. The RFC's naming rule is to drop affixes
that do not differentiate (compare models/qwen.py, not models/qwen_model.py); this
finishes the move that was started.

tree_component.py is also misnamed: it is not "a tree component" but the
TreeComponent ABC plus the shared enums and helpers every component imports. base.py
matches how allocator/base.py and pool_host/base.py already name that role.

Modifications

tree_component.py  -> base.py     TreeComponent ABC, ComponentType, ComponentData, ...
full_component.py  -> full.py     FullComponent
mamba_component.py -> mamba.py    MambaComponent
swa_component.py   -> swa.py      SWAComponent

git mv for all four, then the import path updated at every call site. No symbol was
renamed, no re-export shim added.

Diff surface is exactly: 19 import-path lines, 3 filename string literals in
test_session_unified_radix_cache.py (which asserts on file contents by name), 5 lines
in unified_cache/components/README.md, and isort reordering where base now sorts
ahead of its siblings. Nothing else.

Accuracy Test

Mechanical move, so the bar is byte-level equality rather than accuracy:

• base.py is byte-identical to tree_component.py.
• full.py / mamba.py / swa.py differ from their originals in exactly one line
  each -- the tree_component -> base import.

Verified on an H200 devbox (PYTHONPATH shadowing the image's copy):

              PR branch:   1117 passed, 1070 skipped, 52 subtests passed
  pristine main baseline:  1117 passed, 1070 skipped, 52 subtests passed

Identical command over test_unified_radix_cache_unittest.py,
test_tree_core_registry.py, test_mamba_path_state_cap.py,
test_mamba_donated_alloc_ratio.py, test_session_unified_radix_cache.py against both
trees on the same box -- zero delta, so the 1070 skips are pre-existing and none were
introduced here. Package and per-module imports were also smoke-tested.

Benchmark and Profiling Results

Not applicable -- no runtime behavior changes.

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling.
• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://slack.sglang.ai to discuss your PR.

---

Independent of the other in-flight #25371 PRs -- touches no file that #35306 or #35638 touch.

🤖 Generated with Claude Code

---

CI States

Latest PR Test (Base): ❌ Run #34703150536
Latest PR Test (Extra): 🚫 Run #34703150408
Latest PR Test (AMD ROCm 10): ❌ Run #34703150480

[Jialin]

Nice clean up!