Optimize inline parsing and rendering hot paths

[abap34]

Summary

• replace regex-heavy inline parsing with string_view-based scanners
• reduce AST traversal and template replacement overhead during render/summary
• keep existing coverage green while cutting the main benchmark hot paths

Benchmark

Compared with codex-test-fixtures-structure using make bench BENCH_ARGS="--iterations 5 --json --output ...":

• ✅ 99.5%% speed up: parse_inline_heavy_document (280.621 ms -> 1.463 ms)
• ✅ 99.1%% speed up: parse_large_document (144.207 ms -> 1.335 ms)
• ✅ 96.5%% speed up: render_large_document (163.139 ms -> 5.678 ms)
• ✅ 88.5%% speed up: render_table_heavy_document (149.835 ms -> 17.299 ms)
• ✅ 97.7%% speed up: summary_inline_heavy_document (298.085 ms -> 6.790 ms)
• ✅ 83.3%% speed up: summary_large_document (36.147 ms -> 6.051 ms)

Validation

• make test
• make bench BENCH_ARGS='--iterations 5 --json --output /tmp/almo-opt-bench.json'

[github-actions]

CI Summary

Check	Status	Details
Tests	Success	16 passed / 0 failed
Benchmarks	Success	21 benchmark(s) captured
Coverage	Success	92.7% (1298 of 1400 lines)

Benchmark Means

Benchmark	Mean (ms)	Min (ms)	Max (ms)	Iterations
parse_small_document	0.260	0.250	0.286	5
parse_large_document	5.005	4.964	5.025	5
parse_realistic_article_document	0.361	0.351	0.381	5
parse_release_notes_document	0.319	0.315	0.328	5
parse_inline_heavy_document	5.787	5.762	5.838	5
parse_table_heavy_document	8.216	8.167	8.298	5
parse_footnote_heavy_document	7.195	7.111	7.272	5
parse_structured_document	18.465	18.328	18.748	5
parse_judge_document	0.028	0.027	0.029	5
render_small_document	0.611	0.592	0.623	5
render_large_document	11.374	11.253	11.574	5
render_realistic_article_document	0.553	0.513	0.644	5
render_release_notes_document	0.488	0.467	0.511	5
render_structured_document	22.699	22.590	22.803	5
render_table_heavy_document	27.711	27.618	27.880	5
render_footnote_heavy_document	10.178	10.086	10.329	5
render_interactive_document	0.330	0.315	0.347	5
render_judge_document	0.679	0.664	0.692	5
summary_inline_heavy_document	11.379	11.351	11.425	5
summary_realistic_article_document	0.533	0.518	0.551	5
summary_large_document	22.130	22.088	22.151	5

Benchmark Diff vs codex-test-fixtures-structure

Benchmark	Base Mean (ms)	PR Mean (ms)	Delta (ms)	Change
parse_small_document	9.712	0.260	-9.452	✅ 97.3% speed up
parse_large_document	201.220	5.005	-196.215	✅ 97.5% speed up
parse_realistic_article_document	22.948	0.361	-22.587	✅ 98.4% speed up
parse_release_notes_document	7.435	0.319	-7.116	✅ 95.7% speed up
parse_inline_heavy_document	370.992	5.787	-365.205	✅ 98.4% speed up
parse_table_heavy_document	157.126	8.216	-148.910	✅ 94.8% speed up
parse_footnote_heavy_document	57.471	7.195	-50.276	✅ 87.5% speed up
parse_structured_document	48.713	18.465	-30.248	✅ 62.1% speed up
parse_judge_document	0.028	0.028	+0.000	➖ 0.0% change
render_small_document	18.749	0.611	-18.138	✅ 96.7% speed up
render_large_document	227.820	11.374	-216.446	✅ 95.0% speed up
render_realistic_article_document	35.104	0.553	-34.551	✅ 98.4% speed up
render_release_notes_document	16.608	0.488	-16.120	✅ 97.1% speed up
render_structured_document	67.179	22.699	-44.480	✅ 66.2% speed up
render_table_heavy_document	217.441	27.711	-189.730	✅ 87.3% speed up
render_footnote_heavy_document	72.725	10.178	-62.547	✅ 86.0% speed up
render_interactive_document	12.898	0.330	-12.568	✅ 97.4% speed up
render_judge_document	13.211	0.679	-12.532	✅ 94.9% speed up
summary_inline_heavy_document	374.750	11.379	-363.371	✅ 97.0% speed up
summary_realistic_article_document	31.850	0.533	-31.317	✅ 98.3% speed up
summary_large_document	66.396	22.130	-44.266	✅ 66.7% speed up

Benchmark dashboard: https://abap34.github.io/almo/dev/bench/?pr=177
PR preview: https://abap34.github.io/almo/dev/previews/pr-177/