2022#

View page source

Release Notes for 2022#

mm-dd#

12-29#

  1. Merge in the node_cov branch. This fixed some bugs in the implementation of the rate_eff_cov_table features.

  2. Change the name of the node_cov table rate_eff_table

12-25#

Merge in the node_cov branch:

  1. Add the node_cov table as a quick fix for this covariate Problem ; see rate_eff_cov_table.py.

  2. Because the node_cov table uses weight_id to identify bilinear function, only weightings used by the data and avgint tables need to be Positive .

  3. The column widths of some of the documentation tables were set to be equal instead of adjusting to best fit the columns. This has been fixed.

12-20#

Change the create_database wish list entry to Empty Tables .

12-19#

Advance the version of at_cascade_version to 2022.12.19. This fixes a problem using older versions of python; e.g., 3.8.10.

12-18#

  1. Advance to cppad_mixed-20220518. This avoids a bug in ipopt which would crash when all the fixed effects had equal lower and upper bounds.

  2. Advance the dock_dismod_at.sh dismod_at_version to 20221218 and at_cascade_version to 2022.12.18. You should build the new mixed, dismod_at and at_cascade image.

12-13#

Finish conversion of dismod_at documentation from omhelp to the sphinx xrst wrapper. The following is a link to the old documentation .

12-02#

  1. There was a bug in the python average_integrand calculator when zero was in the age grid . To be specific, the scipy ode integrator chokes when the integration step size is zero. This has been fixed by using the initial conditions for this case.

  2. Add a Parallel Testing wish list item.

11-27#

  1. There were duplicate copies of dismodat.py and fit_ihme.py in the bin and python/bin directories. This has been fixed and not the only database copy of these files is in python/bin.

  2. There were some documentation (not part of user API) that was not being included in the developer documentation. This has been fixed. It was detected while testing the conversion of the documentation to sphinx.

11-03#

Add the Binomial Distribution wish list item (which has been replaced by binomial) .

10-14#

Advance the dock_dismod_at.sh at_cascade_version to 2022.10.14 You should build the new at_cascade image.

09-16#

  1. Change copyright and license declaration to SPDX format.

  2. Change tabs to 3 spaces.

08-27#

Advance the dock_dismod_at.sh dismod_at_version to 20220826 and at_cascade_version to 2022.8.27. You should build the new dismod_at image before the at_cascade image.

08-26#

  1. Remove rel_tol argument to average_integrand because testing with user_average_integrand.py indicates that it does not work; i.e., there is a problem with scipi.integrate.ode.

  2. Change the average_integrand integrator to use lsoda (which switch automatically switches between stiff and non-stiff).

08-15#

Add the optional rel_tol argument to the average_integrand function.

05-19#

Add build to the list of required python3 packages.

05-18#

Advance to cppad_mixed-20220518. This makes the cppad-20220518 available.

05-02#

Add the optional at_cascade.image to dock_dismod_at.sh.

05-01#

Advance the dock_dismod_at.sh dismod_at_version to 20220421. You should rebuild all the images for this update.

04-21#

Advance to cppad_mixed-20220421. This makes the cppad-20220421 available.

04-19#

  1. Advance to cppad_mixed-20220419 which will include the CppAD example atomic ODE solver. We plan to try using this solver to reduce the dismod_at memory requirements.

  2. Add a derivation of the prevalence only ode .

  3. Add age to error message when prevalence cannot be calculated.

04-05#

The user_trapezoidal.py example was added.

API Change#

The system_command_prc default value for return_stdout was changed from False to True. This provides compatibility with the way system_command_prc used to function before the return_stdout argument was added.

04-04#

Add the trapezoidal method as an option for solving the ODE. This option’s name should be changed from rate_case to ode_method.

03-23#

Advance to cppad_mixed-20220323 which will in turn install cppad-20220323.

03-19#

Balancing Covariates was added to the hold_out command; see the new user_hold_out_2.py example.

03-11#

  1. In plot_curve , find a way around Matplotlib bug when there are lots of figures. This sometimes caused a python crash with an IO error.

  2. Fix setting of C compiler in run_cmake.sh .

03-08#

  1. Advance to cppad_mixed-20220304 (more work on macOS homebrew problem).

  2. Fix linking suitesparse (on some macOS homebrew systems).

03-04#

  1. Advance to cppad_mixed-20220304. This fixes some problems with the install on macOS with homebrew.

  2. Fix a problem with the setting of the C++ compiler using specific_compiler .

  3. Make some improvements to data_density_command documentation.

02-17#

Move the value columns fit_value , start, scale, truth, sam_avg, and sam_std to the left in the variable.csv file.

02-08#

Change from using distutils.core.setup to using pip to install the python part of dismod_at ; see Install .

02-05#

Advance to cppad_mixed-20220205 as part of removing the need to set LD_LIBRARY_PATH for the install to work.

01-23#

The debug version would mistakenly print a warning message about mu + eta being less than or equal zero. This was a mistake when computing log scaled difference priors because the mean mu of the difference does not appear inside the log term; see Difference .

01-22#

Add fact that null is the default value for lower and upper to the documentation for the create_database prior_table .