NEWS

dagitty v.0.3-2
==============
Changes:
 * New export options for the causal.effect package and for Python-style edge lists
 * DAGs can now be specified without the enclosing "dag { ... }" if the specification 
   contains at least one directed or bi-directed arrow.

dagitty v0.3 (Release date: 2020-06-11)
==============
Changes:
 * toString: added option 'singular' and 'bnlearn'
 * toMAG: added function to create a MAG from a given DAG
 * impliedConditionalIndependencies: adds option 'all.pairs' for 'type' argument
 * instrumentalVariables: added new algorithm (van der Zander et al., 2015)
 * simulateLogistic: added function to simulate binary data
 * localTests: added option 'cis.chisq' for argument 'type' to test categorical variables
 * isCollider: added function to check for colliders in structure
 * adjustmentSets: added argument 'max.results'
 * simulateSEM: added options 'empirical' and 'verbose'
 * plot.dagitty: show coefficient values for edges.
 * plot.dagitty: fixed bug where bent undirected lines where not drawn thick
   but non-bent undirected lines were

dagitty v0.2-2 (Release date: 2016-08-26)
==============
Changes:

 * Added capability to specify arbitrary path coefficients for simulateSEM
 * Semi-parametric conditional independence testing for additive noise using
   loess smoothing
 * Plot function now sets margins properly and gives better error messages
 * Bugfix in adjustmentSets with type="all" when exposure was defined in graph
 * Bugfix in dconnected and isAdjustmentSet when graphs contain inducing paths
   with 3 or more nodes 
 * Improved parser for larger graphs