🎨 #25 added optional parameter to filter random walks based on an ent…

…ity list
bio-ontology-research-group · Jul 27, 2022 · d732a9a · d732a9a
1 parent 50ac7f2
commit d732a9a
Show file tree

Hide file tree

Showing 5 changed files with 88 additions and 17 deletions.
diff --git a/gateway/src/main/scala/org/mowl/Walking/DeepWalk.scala b/gateway/src/main/scala/org/mowl/Walking/DeepWalk.scala
@@ -4,6 +4,7 @@ import collection.JavaConverters._
 import java.io._
 import java.util.{HashMap, ArrayList}
 import scala.collection.mutable.{MutableList, ListBuffer, Map, ArrayBuffer}
+import scala.collection.immutable.HashSet
 import util.control.Breaks._
 import java.util.concurrent.{ExecutorService, Executors}
 import scala.concurrent.ExecutionContext.Implicits.global
@@ -19,7 +20,9 @@ class DeepWalk (
   var walkLength: Int,
   var alpha: Float,
   var workers: Int,
-  var outfile: String) {
+  var outfile: String,
+  var nodesOfInterest: ArrayList[String]
+) {
 
 
   val edgesSc = edges.asScala.map(x => (x.src, x.rel, x.dst))
@@ -36,6 +39,8 @@ class DeepWalk (
   val rand = scala.util.Random
   val (pathsPerWorker, newWorkers) = numPathsPerWorker()
 
+  val nodesOfInterestIdx =  HashSet() ++ nodesOfInterest.asScala.map(mapEntsIdx(_)).toSet
+
   private[this] val lock = new Object()
 
   val walksFile = new File(outfile)
@@ -145,9 +150,22 @@ class DeepWalk (
     }
 
     val toWrite = walk.filter(_ != -1).map(x => mapIdxEnts(x)).mkString(" ") + "\n"
-    lock.synchronized {
-      bw.write(toWrite)
 
+    if (nodesOfInterest.size > 0){
+      val walkSet = HashSet() ++ walk.toSet
+      val intersection = walkSet & nodesOfInterestIdx
+
+      if (intersection.size > 0){
+
+
+        lock.synchronized {
+          bw.write(toWrite)
+        }
+      }
+    }else{
+      lock.synchronized {
+        bw.write(toWrite)
+      }
     }
 
   }

diff --git a/gateway/src/main/scala/org/mowl/Walking/Node2Vec.scala b/gateway/src/main/scala/org/mowl/Walking/Node2Vec.scala
@@ -4,6 +4,7 @@ import collection.JavaConverters._
 import java.io._
 import java.util.{ArrayList}
 import scala.collection.mutable.{MutableList, ListBuffer, Stack, Map, HashMap, ArrayBuffer}
+import scala.collection.immutable.HashSet
 import util.control.Breaks._
 import java.util.concurrent.{ExecutorService, Executors}
 import scala.concurrent.ExecutionContext.Implicits.global
@@ -20,7 +21,9 @@ class Node2Vec (
   var p: Float,
   var q: Float,
   var workers: Int,
-  var outfile: String) {
+  var outfile: String,
+  var nodesOfInterest: ArrayList[String]
+) {
 
 
   val edgesSc = edges.asScala.map(x => (x.src, x.rel, x.dst, x.weight))
@@ -42,6 +45,8 @@ class Node2Vec (
   var aliasNodes = Map[Int, (Array[Int], Array[Float])]()
   var aliasEdges = Map[(Int, Int), (Array[Int], Array[Float])]()
 
+  val nodesOfInterestIdx =  HashSet() ++ nodesOfInterest.asScala.map(mapEntsIdx(_)).toSet
+
   private[this] val lock = new Object()
 
   val walksFile = new File(outfile)
@@ -196,14 +201,26 @@ class Node2Vec (
     }
 
     val toWrite = walk.filter(_ != -1).map(x => mapIdxEnts(x)).mkString(" ") + "\n"
-    lock.synchronized {
-      bw.write(toWrite)
+
+    if (nodesOfInterest.size > 0){
+      val walkSet = HashSet() ++ walk.toSet
+      val intersection = walkSet & nodesOfInterestIdx
+
+      if (intersection.size > 0){
+
+        lock.synchronized {
+          bw.write(toWrite)
+        }
+      }
+    }else{
+      lock.synchronized {
+        bw.write(toWrite)
+      }
     }
 
   }
 
 
-
   def getAliasEdge(src: Int, dst: Int) = {
     if (graph.contains(dst)){
       val dstNbrs = graph(dst).map(_._2)

diff --git a/mowl/walking/deepwalk/model.py b/mowl/walking/deepwalk/model.py
@@ -7,6 +7,9 @@
 from java.util import HashMap
 from java.util import ArrayList
 from org.mowl import Edge
+from mowl.projection.edge import Edge as PyEdge
+from deprecated.sphinx import versionchanged
+
 logging.basicConfig(level=logging.INFO)
 
 
@@ -36,13 +39,26 @@ def __init__(self,
 
         self.alpha = alpha
 
-    def walk(self, edges):
-
+    @versionchanged(version = "0.1.0", reason = "The method now can accept a list of entities to focus on when generating the random walks.")
+    def walk(self, edges, nodes_of_interest = None):
+        if nodes_of_interest is None:
+            nodes_of_interest = ArrayList()
+        else:
+            all_nodes, _ = PyEdge.getEntitiesAndRelations(edges)
+            all_nodes = set(all_nodes)
+            python_nodes = nodes_of_interest[:]
+            nodes_of_interest = ArrayList()
+            for node in python_nodes:
+                if node in all_nodes:
+                    nodes_of_interest.add(node)
+                else:
+                    logging.info(f"Node {node} does not exist in graph. Ignoring it.")
+
         edgesJ = ArrayList()
         for edge in edges:
             newEdge = Edge(edge.src(), edge.rel(), edge.dst())
             edgesJ.add(newEdge)
 
-        walker = DW(edgesJ, self.num_walks, self.walk_length, self.alpha, self.workers, self.outfile)
+        walker = DW(edgesJ, self.num_walks, self.walk_length, self.alpha, self.workers, self.outfile, nodes_of_interest)
 
         walker.walk()
diff --git a/mowl/walking/node2vec/model.py b/mowl/walking/node2vec/model.py
@@ -4,6 +4,10 @@
 from java.util import ArrayList
 from org.mowl import Edge
 from org.mowl.Walking import Node2Vec as N2V
+from mowl.projection.edge import Edge as PyEdge
+from deprecated.sphinx import versionchanged
+
+logging.basicConfig(level=logging.INFO)
 
 class Node2Vec(WalkingModel):
 
@@ -33,16 +37,26 @@ def __init__(self,
         self.p = p
         self.q = q
 
-    def walk(self, edges):
-
+    def walk(self, edges, nodes_of_interest = None):
+        if nodes_of_interest is None:
+            nodes_of_interest = ArrayList()
+        else:
+            all_nodes, _ = PyEdge.getEntitiesAndRelations(edges)
+            all_nodes = set(all_nodes)
+            python_nodes = nodes_of_interest[:]
+            nodes_of_interest = ArrayList()
+            for node in python_nodes:
+                if node in all_nodes:
+                    nodes_of_interest.add(node)
+                else:
+                    logging.info(f"Node {node} does not exist in graph. Ignoring it.")
+
         edgesJ = ArrayList()
-
         for edge in edges:
             newEdge = Edge(edge.src(), edge.rel(),  edge.dst(), edge.weight())
-
             edgesJ.add(newEdge)
 
-        walker = N2V(edgesJ, self.num_walks, self.walk_length, self.p, self.q, self.workers, self.outfile)
+        walker = N2V(edgesJ, self.num_walks, self.walk_length, self.p, self.q, self.workers, self.outfile, nodes_of_interest)
 
         walker.walk()
 

diff --git a/mowl/walking/walking.py b/mowl/walking/walking.py
@@ -1,3 +1,4 @@
+from deprecated.sphinx import versionchanged
 
 class WalkingModel():
 
@@ -14,14 +15,19 @@ def __init__(self, num_walks, walk_length, outfile, workers=1):
         self.walk_length = walk_length
         self.workers = workers
         self.outfile = outfile
+
+
     # Abstract methods
-    def walk(self, edges):
+    @versionchanged(version = "0.1.0", reason = "The method now can accept a list of entities to focus on when generating the random walks.")
+    def walk(self, edges, nodes_of_interest = None):
 
         '''
-        This method will generate the walks.
+        This method will generate random walks from a graph in the form of edgelist.
 
         :param edges: List of edges
         :type edges: :class:`mowl.graph.edge.Edge`
+        :param nodes_of_interest: List of entity names to filter the generated walks. If a walk contains at least one word of interest, it will be saved into disk, otherwise it will be ignored.  If no list is input, all the nodes will be considered. Defaults to ``None``
+        :type nodes_of_interest: list, optional
         '''
 
         raise NotImplementedError()