Fix README: replace absolute local paths with relative links

The two links to hardware.py and four_two_two.py pointed at
/Users/oho/... which GitHub cannot resolve, causing the 404
Octocat page to render on top. Also updated the project tree
to reflect new files (strategies.py, feedback.py, transfer.py,
rung5.yaml, paper/).
This commit is contained in:
saymrwulf 2026-04-05 12:22:47 +02:00
parent 51d0e5f26b
commit 2cce5af994

View file

@ -16,38 +16,45 @@ The first built-in experiment family targets encoded magic-state preparation in
```text ```text
autoresearch-quantum/ autoresearch-quantum/
├── configs/ ├── configs/rungs/
│ └── rungs/ │ ├── rung1.yaml Baseline: what recipe works?
│ ├── rung1.yaml │ ├── rung2.yaml Stability under noise variation
│ ├── rung2.yaml │ ├── rung3.yaml Transfer across backends
│ ├── rung3.yaml │ ├── rung4.yaml Factory throughput / cost
│ └── rung4.yaml │ └── rung5.yaml Rosenfeld direction
├── src/ ├── src/autoresearch_quantum/
│ └── autoresearch_quantum/ │ ├── cli.py CLI entry point
│ ├── cli.py │ ├── config.py YAML config loader
│ ├── config.py │ ├── models.py All data structures
│ ├── models.py │ ├── codes/
│ ├── codes/ │ │ └── four_two_two.py [[4,2,2]] stabilisers, encoder, seed gates
│ │ └── four_two_two.py │ ├── experiments/
│ ├── experiments/ │ │ └── encoded_magic_state.py Circuit bundle builder
│ │ └── encoded_magic_state.py │ ├── execution/
│ ├── execution/ │ │ ├── analysis.py Postselection, witness, stability
│ │ ├── analysis.py │ │ ├── backends.py Backend resolution
│ │ ├── backends.py │ │ ├── hardware.py IBM hardware executor
│ │ ├── hardware.py │ │ ├── local.py Aer noise simulation executor
│ │ ├── local.py │ │ ├── transfer.py Cross-backend transfer evaluator
│ │ └── transpile.py │ │ └── transpile.py Transpilation utilities
│ ├── lessons/ │ ├── lessons/
│ │ └── extractor.py │ │ ├── extractor.py Human-readable lesson extraction
│ ├── persistence/ │ │ └── feedback.py Machine-readable rules + search narrowing
│ │ └── store.py │ ├── persistence/
│ ├── ratchet/ │ │ └── store.py JSON file store with resumability
│ │ └── runner.py │ ├── ratchet/
│ ├── scoring/ │ │ └── runner.py AutoresearchHarness orchestrator
│ │ └── score.py │ ├── scoring/
│ └── search/ │ │ └── score.py WAC + factory throughput scorers
│ └── challengers.py │ └── search/
│ ├── challengers.py Neighbour generation with dedup
│ └── strategies.py NeighborWalk, RandomCombo, LessonGuided
├── paper/
│ ├── autoresearch_quantum.tex Full technical paper (LaTeX)
│ └── autoresearch_quantum.pdf Compiled PDF (19 pages)
├── tests/ ├── tests/
│ └── test_harness.py 21 tests
├── THE_STORY.md Narrative documentation
├── pyproject.toml ├── pyproject.toml
└── README.md └── README.md
``` ```
@ -104,11 +111,11 @@ Expensive tier:
- IBM Runtime execution through `SamplerV2` - IBM Runtime execution through `SamplerV2`
- only used when enabled and when cheap-tier promotion thresholds are met - only used when enabled and when cheap-tier promotion thresholds are met
- isolated behind [`hardware.py`](/Users/oho/GitClone/CodexProjects/autoresearch-quantum/src/autoresearch_quantum/execution/hardware.py) - isolated behind [`hardware.py`](src/autoresearch_quantum/execution/hardware.py)
## Built-In `[[4,2,2]]` Experiment ## Built-In `[[4,2,2]]` Experiment
The built-in experiment prepares an encoded logical T-state on one logical qubit of the `[[4,2,2]]` code while keeping the spectator logical qubit in `|0⟩`. The code utilities live in [`four_two_two.py`](/Users/oho/GitClone/CodexProjects/autoresearch-quantum/src/autoresearch_quantum/codes/four_two_two.py). The built-in experiment prepares an encoded logical T-state on one logical qubit of the `[[4,2,2]]` code while keeping the spectator logical qubit in `|0⟩`. The code utilities live in [`four_two_two.py`](src/autoresearch_quantum/codes/four_two_two.py).
The harness evaluates: The harness evaluates: