Repository navigation
Releases: ArturSepp/OptionChainAnalytics
Release list
OptionChainAnalytics 6.0.0
OptionChainAnalytics 6.0.0 requires Python 3.11 or newer. Python 3.10 users must stay on the 5.3.x release line.
- Fix the reported Pillow and Certifi security findings with Pillow 13.0.0 and Certifi 2026.10.19.
- Enforce the Pillow minimum in core package requirements and the Certifi minimum in the Deribit, ThetaData, documentation, and aggregate extras; retain resolver constraints and regression checks.
- Preserve separate stable/latest documentation canonical URLs and show the build version in the documentation title.
- Add repository-only retrospective freshness and parity-attribution examples and improve first-use/contributor guidance.
The library source and public API are unchanged from 5.3.0. See CHANGELOG.md for the complete release notes.
OptionChainAnalytics 5.2.0
OptionChainAnalytics 5.2.0 adds ChainTs.load_price_data for chain-linked underlying series and adopts the shared uv-locked cross-platform CI core. It also replaces the contributor-facing dev extra with PEP 735 test and lint dependency groups.
OptionChainAnalytics 5.0.0
OptionChainAnalytics 5.0.0 is a breaking public-module cleanup.
Highlights:
- Adds explicit public homes for option data, conventions, point-in-time reconstruction, provider adapters, normalized-cache helpers, and forward/discount inference.
- Splits provider loading across focused data modules while preserving the package-root DataSource and ts_data_loader_wrapper entry points.
- Improves ThetaData cache workflows, nearest-valid ATM handling, documentation, and runnable examples.
- Removes the historical config, chain_ts, chain_loader_from_ts, ts_loaders, fitters, and CCXT module paths, together with obsolete research examples.
Supported Python: 3.10-3.14. The full compatibility matrix, warning-free docs build, distribution inspection, clean-wheel test, and SigmaStrats validation passed.
Full changelog: https://github.com/ArturSepp/OptionChainAnalytics/blob/v5.0.0/CHANGELOG.md
OptionChainAnalytics 4.0.0
Added
- Added resumable monthly ThetaData EOD Parquet caches, bounded research reads, point-in-time SOFR discount anchoring, and credential-free/live examples for chain, ATM-volatility, and delta-skew workflows.
- Added exact 08:00 UTC BTC/ETH daily caches from local hourly Tardis archives, including exact-time Deribit index prices and source fingerprints.
- Added a unified local-cache example for SPX, VIX, BTC, and ETH.
Changed
- CBOE SPX and VIX conversion now derives robust parity forwards and discounts plus
vanilla-option-pricersimplied volatilities, delta, and vega under explicit settlement conventions. - Normalized local caches now use schema version 3 and one physical
SliceColumnschema across SPX, VIX, BTC, and ETH. - Call-put parity fitting is now provider-neutral, inverse-spread weighted, Huber-reweighted, discount-bounded, and able to accept an external discount factor.
- ThetaData keeps provider policy in its adapter while all normalized IVs and Greeks continue to use
vanilla-option-pricersexclusively.
Fixed
- Chain reports now use current Matplotlib primitives for ATM annotations and open-interest legends and support filtered single-strike slices.
Removed
- Removed the unused CVXPY quote fitter, the
fittersextra, and CVXPY fromall. - Removed the incomplete Yahoo snapshot adapter, its public lazy export, the
yahooextra, andyfinancefromall.
Full changelog: v3.0.0...v4.0.0
OptionChainAnalytics 3.0.0
Breaking CBOE terminology release. Replaces the legacy local-data API with DataSource.CBOE_LOCAL, load_local_cboe_options_data, map_cboe_options_data, and the [cboe] optional dependency. Normalized local caches now use option_chain_analytics.cboe.normalized metadata and must be rebuilt. Python support remains >=3.10. See CHANGELOG.md for full migration details.
OptionChainAnalytics 2.0.1
Full Changelog: v2.0.0...v2.0.1
OptionChainAnalytics 2.0.0
OptionChainAnalytics 2.0.0
This release establishes OCA as the open point-in-time option-chain data layer for quantitative
research and empirical replication.
Highlights
- deterministic simulated option panels for offline tests, examples, and visualisations;
- point-in-time chain reconstruction and documented no-look-ahead conventions;
- SPX/VIX Vlad fitted-chain mapping into
OptionsDataDFs; - corrected
NearestStrikeOnGrid.BELOW/ABOVEbehavior; src/packaging, optional provider extras, Python 3.14 CI, and audited wheel/sdist contents;- task-oriented documentation, citation metadata, and MIT licensing.
Python 3.14 or newer is required. Historical option_chain_analytics.data.chain_ts, .config,
and .chain_loader_from_ts imports remain available as compatibility modules.
No proprietary or licensed empirical dataset is distributed. The public first-success workflow
uses deterministic synthetic data and requires no credentials or network access.