Shedding Hub

Le résumé fourni par la source

The first repository release since v1.0.1 in February, and a large one: 29 new datasets (55 → 84) alongside a statistical modelling layer that did not exist at the last release. Datasets 84 studies, up from 55. The additions span RSV, influenza, rotavirus and SARS-CoV-2, and include the first studies contributed through the agent-assisted extraction workflow. Fitted shedding models — new None of this existed at v1.0.1: fit_shedding_model — censored maximum likelihood over three curve families (exponential, gamma, gamma_shifted), so non-detects contribute what they actually say rather than being dropped or substituted. A shipped catalog of 144 converged fits over 48 studies, with sample size, censoring rate and diagnostics attached, so nothing has to be refitted to be used. simulate_shedding — synthetic cohorts drawn from the fitted population covariance, one row per individual per time point, for agent-based and wastewater models. Cycle-threshold fitting — 78 of the repository's 216 analytes report Ct rather than concentration. They are now fitted directly, with no conversion and no assumed PCR efficiency. Peak time, onset and rise duration compare across scales; decay rate, half-life and peak height do not. Documentation and website A documentation site at shedding-hub.readthedocs.io, including a methods page stating plainly where the estimates should not be trusted. Every dataset page on shedding-hub.github.io now carries a figure per analyte: the fitted curve with the range a simulation would produce, or the measurements alone where no model could be fitted. Package The Python package moved from 0.1.2 at the last release to 0.2.3 on PyPI over this period, which is versioned separately from this repository release. requires-python = ">=3.10" is now declared — earlier releases carried no floor, so pip would install them on interpreters that could not import them. What's Changed Add weekly email report GitHub Action by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/150 Report GitHub traffic API errors instead of silently showing zeros by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/151 Fix pypistats rate limiting with proper User-Agent and retry logic by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/152 feat: persist weekly metrics snapshots to metrics/weekly_metrics.jsonl by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/153 fix: store full GA4 breakdown in weekly metrics snapshot by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/154 Update viz by @haisuzhang in https://github.com/shedding-hub/shedding-hub/pull/155 bump version to 0.1.3 by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/156 fix: persist weekly metrics to unprotected metrics-data branch by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/157 feat: one-off metrics backfill tool + CI workflow by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/159 Promote extraction vocabulary into the schema by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/160 feat: attach weekly trends report to the ops email by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/158 feat: adopt immune_status, coinfection, treatment, nasopharyngeal_aspirate by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/162 data: add 16 agent-extracted datasets + sync 6 existing by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/163 Add shedding simulation from fitted models by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/161 Generated documentation site for the package by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/164 docs: brand the site with the logo, and say what the package is for by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/165 fix: declare the Python 3.10 floor the code has always had by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/166 expand the schema by @hww228 in https://github.com/shedding-hub/shedding-hub/pull/167 add unit gc/ng hDNA in schema by @hww228 in https://github.com/shedding-hub/shedding-hub/pull/169 data: add 13 agent-extracted datasets and rebuild the catalog by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/170 ci: notify the dashboard when data lands on main by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/171 feat: fit cycle-threshold data by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/168 Open a PR from the figures refresh instead of pushing by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/172 Document what happens to a dataset after it is merged by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/173 Correct the review guidance for the figures pull request by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/175 figures: refresh the dataset pages by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/174 Open the figures pull request with the personal access token by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/176 Say that a figures refresh is normally quiet by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/177 Document cycle-threshold fitting, and the range view, in getting started by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/178 Configure Dependabot so its alerts become pull requests by @YWAN446 in https://github.com/shedding-hub/shedding-hub/pull/179 Full Changelog: https://github.com/shedding-hub/shedding-hub/compare/v1.0.1...v1.0.2

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.