OHDSI Home | Forums | Wiki | Github

HADES development announcements

Characterization 0.0.5 has been released. This is a new analysis package for executing various characterization studies such as time-to-event, dechallenge/rechallenge and descriptive studies of target populations experiencing an outcome within a user defined time-at-risk and not experiencing an outcome within a time-at-risk.

SqlRender 1.12.0 has been released, including some final changes required for the upcoming DatabaseConnector 6.0.0 (which will be released early next week).

CirceR 1.3.0 has been released, introducing a new function buildConceptSetQuery() to generate queries from concept set expression JSON.

DatabaseConnector 6.0.0 has been released. There are many changes, but these are the most important:

  1. Now supporting dbplyr, which some users seem to prefer over translating OhdsiSql. An important limitation is that field names must be case insensitive (lowercase in R), because by default most database platforms are case-insensitive. DatabaseConnector also includes various date functions for dbplyr, such as dateAdd() and dateDiff() (dbplyr itself doesn’t support date operations).
  2. Added the computeDataHash() function. This computes a hash based on the table structures and row counts in a database schema, so in theory will be stable as long as the data doesn’t change. This should be useful in distributed research network, where for example this function can be used to check whether the data has changed since the last DataQualityDashboard results.
  3. Added the options(LOG_DATABASECONNECTOR_SQL = TRUE) option and extractQueryTimes() function, which together can be used to monitor query performance via the ParallelLogger log.

OhdsiShinyModules v1.0.2 has been released. There are a four changes:

  1. The module servers now require a connection (optionally a pooled one) as input rather than the connection details.
  2. CohortDiagnostic shiny modules were added
  3. Minor bug fixes were made to prediction and cohort method modules
  4. Improvements were made to the characterization module to make it easier to use

ShinyAppBuilder v1.1.1 has been released. This is a new HADES package that lets you build shiny apps that combine multiple modules from OhdsiShinyModules or custom modules that match the module standards.

There is code for creating shiny apps for shiny servers or opening the shiny app locally.

ResultModelManager v0.3.0 has been released.
This release added functionality for uploading results data sets to servers in a robust manner, conforming to a results schema data.frame definition.

For a usage guide on this new functionality, please see this article.

CohortExplorer v0.0.11 has been updated. Just routine clean up so that github actions doesn’t complain :slight_smile: And ofcourse 100% test coverage

ROhdsiWebApi 1.3.2 has been released. This contains some minor changes to allow the released version to pass R Check again.

SqlRender 1.12.1 has been released. This contains two bugfixes for Snowflake.

DataQualityDashboard 2.1.1 has been released. This is the first release of the package which meets all HADES requirements, and hereafter DQD will be a part of HADES.

DataQualityDashboard is an OHDSI mainstay used to evaluate and expose data quality in an OMOP CDM instance. It’s an R package that includes an R shiny application which can be used to visualize and explore the results of its data quality evaluations.

The DQD runs a set of data quality checks against the OMOP CDM (it currently supports CDM versions 5.2, 5.3, and 5.4). These checks are highly configurable; they can be toggled off and on and their failure thresholds can be adjusted according to the user’s needs. The thousands of checks in the DQD are grouped into a smaller set of data quality “check types” which each represent a singular data quality idea. Each check type corresponds to a parameterized SQL query into which OMOP tables, fields, and concepts are substituted to create an individual data quality check.

The threshold files used to configure checks and their failure thresholds can be shared amongst institutions to facilitate consistent use of the DQD for network-level projects. Similarly, the results of a DQD run are portable and can be stored in json, csv, and/or a database table.

For a comprehensive overview of the DataQualityDashboard, please see the package website here. Questions and bug reports can be submitted on our issue tracker here. I’ll be on point as the package maintainer but send most credit to @clairblacketer @Frank and other DQ WG members for building the package and supporting the HADES submission :slight_smile:

BigKnn 1.0.2 has been released. This adds support for the upcoming Andromeda release (using arrow instead of SQLite in the backend for improved speed).

SqlRender 1.13.0 has been released. This has some fixes for BigQuery, SQLite, and DuckDB.

ROhdsiWebApi 1.3.3 has been released. This is again just a minor update to pass R Check again.

Capr 2.0.0 has been released. This is a major update overhauling the user-interface of Capr. The intent of the overhaul is to improve the readability of the Capr code. For full details visit the package website. Below is a brief summary of some updates:

  • new handling of concept sets using cs(). Can build concept sets on the fly (requiring only a name and id) to use in a cohort definition. Improved functionality for handling descendants, mapped and excluded concepts. An example: cs(desecendants(201826L), name = "T2D")
  • new interface for specifying cohort logic via cohort(entry, attrition, exit, era). Stylistically definitions are to be layered as nested functions in order to convey the entire logic within one function. This change improves is intended to improve the readability of the Capr call.
  • maintain support for cohort templating
  • change end point of Capr to only produce the json structure for the cohort definition (either as a character string or saved file). It is suggested that the user deploys the json using the appropriate tools like CohortGenerator or CirceR, which are better suited to generating the cohort definition.

DatabaseConnector 6.1.0 has been released. This adds support for DuckDB, fixes some issues related to Java heap memory, and fixes some Snowflake issues (related to DATETIME).

1 Like

DataQualityDashboard 2.1.2 has been released. This release contains a bug fix for the cdmDataType check SQL.

1 Like

Andromeda 0.6.2 has been released (on github but not CRAN). This adds one function isAndromedaTable() that should work with both SQLite and Arrow andromeda tables.
Reminder, we are moving to a new backend for Andromeda based on Apache Arrow soon. If your package depends on Andromeda please test it using the Andromeda develop branch and report any issues in the Andromeda issue tracker. You may need to make some slight changes to your code such as using nrow instead of tally to get the number of rows in an Andromeda table.

SqlRender 1.13.1 has been released, with a single bugfix for Netezza.

DeepPatientLevelPrediction 1.1 has been released. This adds adds a number of bug fixes and some new features. The new features are:

  • An automatic learning rate finder has been implemented, used by setting learningRate="auto"
  • Training loop parameters are now handled by a new settings object and as a result more of the training loop parameters are customizable by the user
  • Custom metrics can be created by the user to use for early stopping and learning rate schedule.
  • Seeds have been added in various places, making the whole process more reproducible

Thanks to @lhjohn for his contributions to this release.

1 Like