pval fix, and --include-all mode for scoring

reality · Feb 12, 2022 · 6baacc1 · 6baacc1
1 parent 3de679c
commit 6baacc1
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 11 deletions.
diff --git a/klarigi/src/main/groovy/klarigi/App.groovy b/klarigi/src/main/groovy/klarigi/App.groovy
@@ -47,6 +47,7 @@ class App {
       _ longOpt: 'min-inclusion', 'Explanations with inclusion below this level will not be considered for explanations.', args: 1
       _ longOpt: 'min-power', 'Explanations with power below this level will not be considered for explanations.', args: 1
       _ longOpt: 'min-ic', 'Explanations with IC below this level will not be considered for explanations.', args: 1
+      _ longOpt: 'include-all', 'Ignore all min scores', type: Boolean
 
       _ longOpt: 'max-exclusion', 'Variables with exclusion higher than this will not count to total overall inclusion in the stepdown algorithm. They will, however, still appear in explanations.', args: 1
       _ longOpt: 'max-inclusion', 'Variables with inclusion higher than this will not count to total overall inclusion in the stepdown algorithm. They will, however, still appear in explanations.', args: 1
@@ -120,11 +121,11 @@ class App {
           System.exit(1)
         }
 
-        allExplanations = k.explainClusters(groups, excludeClasses, o['scores-only'], o['output-scores'], o['output-type'], o['power'], threads, o['debug'])
+        allExplanations = k.explainClusters(groups, excludeClasses, o['scores-only'], o['output-scores'], o['output-type'], o['power'], threads, o['debug'], o['include-all'])
       } else if(o['group'] && o['group'] != '*') {
-        allExplanations = k.explainClusters([o['group']], excludeClasses, o['scores-only'], o['output-scores'], o['output-type'], o['power'], threads, o['debug'])
+        allExplanations = k.explainClusters([o['group']], excludeClasses, o['scores-only'], o['output-scores'], o['output-type'], o['power'], threads, o['debug'], o['include-all'])
       } else {
-        allExplanations = k.explainAllClusters(o['output-scores'], excludeClasses, o['scores-only'], o['output-type'], o['power'], threads, o['debug'])
+        allExplanations = k.explainAllClusters(o['output-scores'], excludeClasses, o['scores-only'], o['output-type'], o['power'], threads, o['debug'], o['include-all'])
       }
 
       if(o['scores-only']) {

diff --git a/klarigi/src/main/groovy/klarigi/Classifier.groovy b/klarigi/src/main/groovy/klarigi/Classifier.groovy
@@ -38,7 +38,8 @@ public class Classifier {
           def score = 0
           if(subeq.any { codes.containsKey(it) }) {
             //score = e.nPower * e.nIc
-            score = e.nExclusion * e.nIc
+            //score = e.nExclusion * e.nIc
+            score = e.nExclusion
           }
 
           return score

diff --git a/klarigi/src/main/groovy/klarigi/Klarigi.groovy b/klarigi/src/main/groovy/klarigi/Klarigi.groovy
@@ -254,13 +254,23 @@ public class Klarigi {
       }
 
       ae.each { g, gp ->
-        def candidates = scorer.scoreClasses(g, excludeClasses, threads, ae[g].keySet().toList(), true)
+        def cSet = ae[g].keySet().toList()
+        def candidates = scorer.scoreClasses(g, excludeClasses, threads, cSet, true)
 
         candidates.each { v ->
           def k = v.iri
           ae[g][k].incVals << v.nInclusion
           ae[g][k].excVals << v.nExclusion
           ae[g][k].powVals << v.nPower
+
+          cSet -= k
+        }
+
+        // Add zeros for remaining bois
+        cSet.each { k ->
+          ae[g][k].incVals << 0
+          ae[g][k].excVals << 0
+          ae[g][k].powVals << 0
         }
       }
     }
@@ -269,6 +279,7 @@ public class Klarigi {
     ae.each { c, terms ->
       ps[c] = [:]
       terms.each { iri, cv ->
+      if(iri == "http://purl.obolibrary.org/obo/HP_0009763") { println cv.incVals }
         ps[c][iri] = [
           incP: new BigDecimal(cv.incVals.findAll { it >= cv.nInclusion }.size() / cv.incVals.size()).round(new MathContext(3)),
           excP: new BigDecimal(cv.excVals.findAll { it >= cv.nExclusion }.size() / cv.excVals.size()).round(new MathContext(3)),
@@ -286,9 +297,9 @@ public class Klarigi {
     ps
   }
 
-  def explainCluster(cid, powerMode, excludeClasses, scoreOnly, outputScores, outputType, threads, debug) {
+  def explainCluster(cid, powerMode, excludeClasses, scoreOnly, outputScores, outputType, threads, debug, includeAll) {
     def scorer = new Scorer(ontoHelper, coefficients, data)
-    def candidates = scorer.scoreAllClasses(cid, excludeClasses, threads)
+    def candidates = scorer.scoreAllClasses(cid, excludeClasses, threads, includeAll)
 
     println "$cid: Scoring completed. Candidates: ${candidates.size()}"
 
@@ -315,15 +326,15 @@ public class Klarigi {
     return res
   }
 
-  def explainClusters(groups, excludeClasses, scoreOnly, outputScores, outputType, powerMode, threads, debug) {
+  def explainClusters(groups, excludeClasses, scoreOnly, outputScores, outputType, powerMode, threads, debug, includeAll) {
     data.groupings.findAll { g, v -> groups.contains(g) }.collect { g, v ->
-      [ cluster: g, results: explainCluster(g, powerMode, excludeClasses, scoreOnly, outputScores, outputType, threads, debug) ]
+      [ cluster: g, results: explainCluster(g, powerMode, excludeClasses, scoreOnly, outputScores, outputType, threads, debug, includeAll) ]
     }
   }
 
-  def explainAllClusters(outputScores, excludeClasses, scoreOnly, outputType, powerMode, threads, debug) {
+  def explainAllClusters(outputScores, excludeClasses, scoreOnly, outputType, powerMode, threads, debug, includeAll) {
     data.groupings.collect { g, v ->
-      [ cluster: g, results: explainCluster(g, powerMode, excludeClasses, scoreOnly, outputScores, outputType, threads, debug) ]
+      [ cluster: g, results: explainCluster(g, powerMode, excludeClasses, scoreOnly, outputScores, outputType, threads, debug, includeAll) ]
     }
   }