Added latent-node styling and contract-based edges to cytoscapeNetwork

NEWS.md

@@ -37,8 +37,29 @@ statement-type and entity-type vocabularies, value ranges, `NA` statistics
 on nodes that are not in the input data, and that every edge's source and
 target are nodes. It stops with an error listing every problem found.
 
+* `cytoscapeNetwork()` draws the node status columns. Nodes not in the
+input data (`measured = FALSE`) have no fill and a dashed grey border, and
+the legend lists them as "not in input data"; nodes with
+`included_in_query = FALSE` are faded; nodes with an `NA` `logFC` have no
+fill. `entity_type` sets the node shape: hexagon for metabolites and lipids,
+diamond for drugs, octagon for complexes, barrel for families. The legend
+shows only the node styles, shapes, and edge types present, and covers
+every statement type of the contract.
+
 ## Breaking changes
 
+* `cytoscapeNetwork()` reads the edge `directed` column: an undirected edge
+has no arrow and is drawn once for both directions, even when the reverse
+edge is missing. Edges without the column treat `Complex` and
+`Association` as undirected. Previously a `Complex` edge without its
+reverse was drawn as directed.
+* `cytoscapeNetwork()` colours a protein with PTM `site` rows from its
+row with no `site`. A protein with only site rows is drawn as an uncoloured
+container, and its site nodes keep their own colours. Previously it took
+the colour of its first row, often a site's fold change.
+* `cytoscapeNetwork()` no longer draws an `NA` `logFC` as the grey of a
+zero fold change.
+
 * The edges returned by `getSubnetworkFromIndra()` follow the edge contract.
 Columns are renamed, with no aliases: `stmt_hash` to `statement_id` (now
 always character, taken from INDRA's full-precision `matches_hash`),

R/cytoscapeNetwork.R

@@ -10,10 +10,23 @@
 #'                    columns: \code{logFC} (numeric), \code{entity_name}
 #'                    (character; may be semicolon-joined for multi-grounded
 #'                    rows), \code{entity_id} (character), \code{site}
-#'                    (character, underscore-separated PTM site list).
+#'                    (character, underscore-separated PTM site list), and
+#'                    the node status columns of \code{\link{validate_network}()}:
+#'                    \code{entity_type} sets the node shape (e.g. a hexagon
+#'                    for \code{"metabolite"}); \code{measured = FALSE} draws a
+#'                    node with no fill and a dashed grey border ("not in input
+#'                    data"); \code{included_in_query = FALSE} fades it. A node
+#'                    with an \code{NA} \code{logFC} has no fill. A protein
+#'                    with \code{site} rows takes its colour from its row
+#'                    with no \code{site}; without one, only its sites are
+#'                    coloured. Missing status columns count as \code{TRUE}.
 #' @param edges       Data frame with columns \code{source}, \code{target},
-#'                    \code{interaction}.  Optional: \code{site},
-#'                    \code{evidence_url}, opened when an edge is clicked.
+#'                    \code{interaction}.  Optional: \code{directed} (an
+#'                    edge with \code{FALSE} has no arrow and is drawn once
+#'                    for both directions; without the column, \code{Complex}
+#'                    and \code{Association} edges are undirected),
+#'                    \code{site}, \code{evidence_url}, opened when an edge is
+#'                    clicked.
 #' @param displayLabelType \code{"id"} (default) or \code{"entity_name"}:
 #'                    which column is used as the visible node label.
 #'                    \code{"entityName"} is deprecated and is treated as