Skip to content

Dev howiemen - #7

Merged
HowieMen merged 4 commits into
mainfrom
dev-howiemen
Aug 28, 2026
Merged

HowieMen merged 4 commits into
mainfrom
dev-howiemen

Conversation

@HowieMen

Copy link
Copy Markdown
Contributor

No description provided.

Replace the two-tier caching description with the new three-tier model
introduced by `HqDataCsvPortal`: per-date dict cache (`_daily_index` /
`_factor_index`), per-symbol cumulative series, and calendar cache.
Document hqdata-side read-once-per-day behaviour, lazy calendar caching
caveat, the avoided hot spots (daily round-trips, redundant Bar
construction, repeated CSV header inference), and updated memory and
benchmark figures (5-stocks MA ~13 s, 50×250 history(20) 3.3 s). Update
test entry-point paths and code references accordingly.
… path

- §5 `data portal` clause now describes reading via `hqdata.api` `csv` source instead of owning CSV layout/validation; column-name checks, missing-file errors, and whole-day boundaries are hqdata's responsibility; portal forbids importing `hqdata.sources` / SDKs or reimplementing column mapping.
- §5 layout description defers to `hqdata.init_source("csv", root=...)`; missing/invalid files surface as `SnapshotFileMissingError` / `InvalidDataError` from hqdata, re-raised by portal.
- Fix stale `src/hqbacktest/data/view.py` references in `docs/isolation.md` and `docs/strategy-api.md` to `src/hqbacktest/data/data_view.py`.
- Append 2026-08-26 changelog row noting the §5 sync and path fix.
@HowieMen
HowieMen merged commit 34f175d into main Aug 28, 2026
0 of 6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant