CHANGES IN VERSION 2.4.0 ------------------------ NEW FEATURES o add new status codes '4' and '5' to 'statuses' mysql table; change 'status_id' field to '4' for all removed records to date o add getRecordStatus() generic o add package() generic o create 'Hub' VIRTUAL class - add new .Hub() base constructor for all hubs - add getAnnotationHubOption() and setAnnotationHubOption() - promote cache() to generic - add getHub() getter for AnnotationHubResource class - add getUrl(), getCache(), getDate() getters - export as few db helpers as possible o add 'EpigenomeRoadmapNarrowAllPeaks' and 'EpigenomeRoadmapNarrowFDR' classes MODIFICATIONS o distinguish between broad and narrow peak files in EpigenomeRoadmapFileResource dispatch class o don't use cache for AnnotationHub SQLite connection - originally introduced so could be closed if needed, but creates complexity - instead, open / close connection around individual queries (not a performance concern) - expose hub, cache, proxy in AnnotationHub constructor - document dbconn,Hub-method, dbfile,Hub-method, .db_close o snapshotDate now uses timestamp (last date any row was modified) instead of rdatadateadded o .require fails rather than emits warning - unit test on .require() - also, cache(hub[FALSE]) does not create spurious error o work on removed records and biocVersion - .uid0() was reorganized and no longer groups by record_id - metadata is returned for records with biocversion field <= current biocVersion instead of an exact match with the current version - metadata is not returned for removed records BUG FIXES o Work around httr() progress() bug by disabling progress bar CHANGES IN VERSION 2.2.0 ------------------------ SIGNIFICANT USER-VISIBLE CHANGES o seqinfo(GRanges) for all genomes supported by GenomeInfoDb now contain seqlengths. CHANGES IN VERSION 2.1.21 -------------------------- SIGNIFICANT USER-VISIBLE CHANGES o fileName() returns the cache path on the disk for a file that has been cached and NA for files which have not been cached. o The error message (when file is not downloaded from the hub) displays the AnnotationHub name, title, and reason for failure. CHANGES IN VERSION 2.1 ---------------------- NEW FEATURES o as.list() splits AnnotationHub instances into a list of instances, each with a single record. c() concatenates hubs. BUG FIXES o cache<- now behaves as documented, e.g., removing the cached version of the file. CHANGES IN VERSION 2.0.0 ------------------------ NEW FEATURES o AnnotationHub is all new. We basically rewrote the entire thing. o The back end is new (new database, new way of tracking/routing data etc.) o The front end is new (new AnnotationHub object, new methods, new behaviors, new ways of finding and downloading data) o The metadata has also been cleaned up and made more consistent/searchable o The recipes that are used to populate these data have also been cleaned up. o There is also a new vignette to explain how to use the new AnnotationHub in detail IMPROVEMENTS SINCE LAST TIME o The old way of finding data (an enormous tree of paths), was not really scalable to the amount of data we have to provide access to. So we junked it. Now you have a number of better methods to allow you to search for terms instead. o The new hub interface can be searched using a new display method, but it can *also* be searched entirely from the command line. This allows you to use it in examples and scripts in a way that is friendlier for reproducible research. o For users who want to contribute valuable new annotation resources to the AnnotationHub, it is now possible to write a recipe and test that it works for yourself. Then once you are happy with it, you can contact us and we can add data to the AnnotationHub.