Datasets:
Formats:
json
Languages:
English
Size:
< 1K
Tags:
mathematics
information-theory
experimental-design
memory-compression
linear-algebra
quantum-information
File size: 2,947 Bytes
9552b49 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 | # OQAIVAEI Hugging Face packaging edition 1
Prepared 20 October 2026 for the target dataset repository `PureOne/OQAIVAEI`.
## Preserved research
The 83 research files are copied unchanged under `research/`. The original complete ZIP is retained byte-for-byte under `releases/`. The separately supplied verification report is retained under `verification/`. The manuscript, mathematical claims, proofs, bibliography, figures, and computational programs are unchanged.
The original manifest inside `research/MANIFEST.sha256` describes its 82 other research files. Run its checks relative to `research/`. The top-level `SHA256SUMS.txt` describes the full Hub payload apart from itself.
## Added for the Hub
- A dataset card with direct manuscript, code, evidence, and archive links.
- Four JSONL configurations and a deterministic transformation script.
- Citation metadata, dataset schemas, provenance, and explicit license-status documentation.
- An upload kit with an explicit file manifest, local integrity validation, and Windows/POSIX launchers. Upload tooling is outside `repo/`; only manifest-listed repository files are submitted.
- A packaging QA report distinct from the original scientific execution record.
## Historical environment notes
The source README and requirements comments mention CPython 3.12.14 from an earlier run. The preserved renamed-edition execution report records CPython 3.13.5 with NumPy 2.3.5, SciPy 1.17.0, SymPy 1.14.0, and Matplotlib 3.10.8. The Hub card uses the latter record. Those preserved files are historical evidence and were not silently rewritten.
Historical execution logs contain container paths such as `/mnt/data/oqaivaei_work` and `/opt/pyvenv`. Reproduction scripts locate their files relative to their own locations; users do not need those historical paths.
The original comparison of the two PDF editions and visual QA are author-provided records. This packaging step checks the delivered PDF's bytes, page count, links to source, data transformations, and release integrity. It does not claim to repeat the original full visual review or provide an external proof review.
## Publication status
The complete computational pipeline was also rerun for this packaging edition in a separate copy on CPython 3.12.14 with the pinned dependencies. All eight stages passed, with no failures or skipped stages. The fresh record is `verification/packaging_replay.json`; `verification/content_audit.json` summarizes the source-hash and content checks. This was an internal replay, not external replication.
This is an upload-ready package. Preparation does not create a Hugging Face repository. The citation URL and examples identify the intended destination and become usable after upload. No DOI or arXiv identifier is assigned or implied.
No license grant was present in the supplied source. This edition adds none. No new scientific priority claim or external validation is introduced.
|