5 files changed, 263 insertions, 65 deletions

diff --git a/run_tests.bash b/run_tests.bash
new file mode 100755
index 0000000..17cbf01
--- /dev/null
+++ b/run_tests.bash
@@ -0,0 +1,203 @@
+#!/usr/bin/env bash
+
+NC='\033[0m'
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[0;33m'
+
+msg_info() {
+    echo -e "${GREEN}$1${NC}"
+}
+
+msg_warn() {
+    echo -e "${YELLOW}$1${NC}"
+}
+
+msg_error() {
+    echo -e "${RED}$1${NC}"
+}
+
+print_help() {
+    echo
+    echo "testRSA - a quick script to run tests all night long"
+    echo
+    echo "USAGE:"
+    echo "    testRSA OPTION [...]"
+    echo
+    echo "OPTIONs are:"
+    echo "    -o | --ontology <path>:"
+    echo "        path to ontology."
+    echo "    -d | --data <path>:"
+    echo "        path to a folder containing data for the ontology."
+    echo "    -q | --queries <path>:"
+    echo "        path to a folder containing SPARQL query files to be"
+    echo "        executed against the ontology and data."
+    echo "    -p | --prefix <path>:"
+    echo "        provides a folder to prefix to the output files."
+    echo "        Defaults to './results'."
+    echo "    -h | -? | --help:"
+    echo "        print this help"
+    echo
+}
+
+ONTOLOGY=""
+DATA=""
+QUERIES=""
+PREFIX="./results"
+
+while [[ $# -gt 0 ]]
+do
+    case $1 in
+        -o|--ontology)
+            shift
+            ONTOLOGY="$1"
+            [ ! -r "$ONTOLOGY" ] && \
+                msg_error "Unable to read '$ONTOLOGY'" && \
+                print_help && \
+                exit 2
+            ;;
+        -d|--data)
+            shift
+            DATA="$1"
+            [ ! -d "$DATA" ] && \
+                msg_error "'$DATA' is not a directory" && \
+                print_help && \
+                exit 2
+            ;;
+        -q|--queries)
+            shift
+            QUERIES="$1"
+            [ ! -d "$QUERIES" ] && \
+                msg_error "'$QUERIES' is not a directory" && \
+                print_help && \
+                exit 2
+            ;;
+        -p|--prefix)
+            shift
+            PREFIX="$1"
+            ;;
+        -h|-?|--help)
+            print_help
+            exit 0
+            ;;
+        *)
+            msg_error "$OPTION: invalid option"
+            print_help
+            exit 1
+            ;;
+    esac
+    shift
+done
+
+[ -z "$ONTOLOGY" ] && \
+    msg_error "Use -o | --ontology to provide an ontology file" && \
+    print_help && \
+    exit 3
+
+[ -z "$DATA" ] && \
+    msg_error "Use -d | --data to provide a data folder" && \
+    print_help && \
+    exit 3
+
+[ -z "$QUERIES" ] && \
+    msg_error "Use -q | --queries to provide a query folder" && \
+    print_help && \
+    exit 3
+
+
+DATAS=`\ls $DATA/*`
+mkdir -p "$PREFIX"
+for QUERY in "$QUERIES"/*.sparql
+do
+    sbt "run $QUERY $ONTOLOGY $DATAS" 2>&1 | tee "$PREFIX/answers_$(basename $QUERY .sparql).txt"
+done
+
+OUTPUT="$PREFIX/results.csv"
+echo "NAME, TBOX, RBOX, ABOX, \
+    CANONICAL MODEL GENERATION, \
+    CANONICAL MODEL RULES, CANONICAL MODEL RULES LOADING, \
+    CANONICAL MODEL FACTS, CANONICAL MODEL FACTS LOADING, \
+    CANONICAL MODEL IDB, CANONICAL MODEL EDB, \
+    FILTERING PROGRAM GENERATION, \
+    FILTERING PROGRAM RULES, FILTERING PROGRAM RULES LOADING, \
+    FILTERING PROGRAM FACTS, FILTERING PROGRAM FACTS LOADING, \
+    FILTERING PROGRAM IDB, FILTERING PROGRAM EDB, \
+    ANSWERING TIME, #ANSWERS, #UNFILTERED, #SPURIOUS, %SPURIOUS" > "$OUTPUT"
+
+for RESULT in "$PREFIX"/*.txt
+do
+    awk -v filename="$RESULT" '
+        BEGIN { 
+            OFS = ", "
+            name = filename
+            sub("^.*answers_", "", name)
+            sub(".txt$", "", name)
+        }
+        /Original TBox/ { tbox_size = $NF }
+        /Original RBox/ { rbox_size = $NF }
+        /Original ABox/ { abox_size = $NF }
+        /Generating canonical model program \(END\)/ { canon_gen_time = $NF }
+        /Generating filtering program \(END\)/ { filter_gen_time = $NF }
+        /Canonical model rules/ {
+            canon_rules = $NF
+            canon = 1
+        }
+        /Canonical model facts/ {
+            canon_facts = $NF
+            canon = 1
+        }
+        /Filtering program rules/ {
+            filter_rules = $NF
+        }
+        /Filtering program facts/ {
+            filter_facts = $NF
+        }
+        /Loading rules \(END\)/ {
+            if (canon) {
+                canon_rules_load = $NF
+            } else {
+                filter_rules_load = $NF
+            }
+        }
+        /Loading facts/ {
+            if (canon) {
+                canon_facts_load = $NF
+            } else {
+                filter_facts_load = $NF
+            }
+        }
+        /Aggregate number of IDB facts/ {
+            sub("^.*=", "")
+            sub(",$", "")
+            if (canon) {
+                canon_idb = $0
+            } else {
+                filter_idb = $0
+            }
+        }
+        /Aggregate number of EDB facts/ {
+            sub("^.*=", "")
+            sub(",$", "")
+            if (canon) {
+                canon_edb = $0
+                canon = 0
+            } else {
+                filter_edb = $0
+            }
+        }
+        /Answers computation \(END\)/ { answers_time = $NF }
+        /Number of answers/ { answers = $(NF-1) }
+        /Number of unfiltered answers/ { unfiltered = $(NF-1) }
+        /Number of spurious answers/ { spurious = $(NF-1) }
+        /Percentage of spurious answers/ { spurious_perc = $NF }
+        END { print name, tbox_size, rbox_size, abox_size, \
+                    canon_gen_time, canon_rules, canon_rules_load, canon_facts, canon_facts_load, \
+                    canon_idb, canon_edb, \
+                    filter_gen_time, filter_rules, filter_rules_load, filter_facts, filter_facts_load, \
+                    filter_idb, filter_edb, \
+                    answers_time, answers, unfiltered, spurious, spurious_perc
+        }
+    ' "$RESULT" >> "$OUTPUT"
+done
+
+exit 0
diff --git a/src/main/scala/uk/ac/ox/cs/rsacomb/FilteringProgram.scala b/src/main/scala/uk/ac/ox/cs/rsacomb/FilteringProgram.scala
index 4e533c6..9427735 100644
--- a/src/main/scala/uk/ac/ox/cs/rsacomb/FilteringProgram.scala
+++ b/src/main/scala/uk/ac/ox/cs/rsacomb/FilteringProgram.scala
@@ -126,9 +126,9 @@ class FilteringProgram(query: ConjunctiveQuery, constants: List[Term]) {
         if index2 >= 0
       } yield Rule.create(
         RSA.FK,
+        RSA.ID(RSA(index1), RSA(index2)),
         role1 << Forward,
         role2 << Forward,
-        RSA.ID(RSA(index1), RSA(index2)),
         not(RSA.Congruent(role1.getArguments get 0, role2.getArguments get 0))
       )
       val r4b = for {
@@ -140,9 +140,9 @@ class FilteringProgram(query: ConjunctiveQuery, constants: List[Term]) {
         if index2 >= 0
       } yield Rule.create(
         RSA.FK,
+        RSA.ID(RSA(index1), RSA(index2)),
         role1 << Forward,
         role2 << Backward,
-        RSA.ID(RSA(index1), RSA(index2)),
         not(RSA.Congruent(role1.getArguments get 0, role2.getArguments get 2))
       )
       val r4c = for {
@@ -154,9 +154,9 @@ class FilteringProgram(query: ConjunctiveQuery, constants: List[Term]) {
         if index2 >= 0
       } yield Rule.create(
         RSA.FK,
+        RSA.ID(RSA(index1), RSA(index2)),
         role1 << Backward,
         role2 << Backward,
-        RSA.ID(RSA(index1), RSA(index2)),
         not(RSA.Congruent(role1.getArguments get 2, role2.getArguments get 2))
       )
 
@@ -180,13 +180,13 @@ class FilteringProgram(query: ConjunctiveQuery, constants: List[Term]) {
           RSA(query.bounded indexOf r1arg0),
           RSA(query.bounded indexOf r2arg0)
         ),
-        role1 << Forward,
-        role2 << Forward,
         RSA.ID(
           RSA(query.bounded indexOf r1arg2),
           RSA(query.bounded indexOf r2arg2)
         ),
         RSA.Congruent(r1arg0, r2arg0),
+        role1 << Forward,
+        role2 << Forward,
         not(RSA.NI(r1arg0))
       )
       val r5b = for {
@@ -205,13 +205,13 @@ class FilteringProgram(query: ConjunctiveQuery, constants: List[Term]) {
           RSA(query.bounded indexOf r1arg0),
           RSA(query.bounded indexOf r2arg2)
         ),
-        role1 << Forward,
-        role2 << Backward,
         RSA.ID(
           RSA(query.bounded indexOf r1arg2),
           RSA(query.bounded indexOf r2arg0)
         ),
         RSA.Congruent(r1arg0, r2arg2),
+        role1 << Forward,
+        role2 << Backward,
         not(RSA.NI(r1arg0))
       )
       val r5c = for {
@@ -230,13 +230,13 @@ class FilteringProgram(query: ConjunctiveQuery, constants: List[Term]) {
           RSA(query.bounded indexOf r1arg2),
           RSA(query.bounded indexOf r2arg2)
         ),
-        role1 << Backward,
-        role2 << Backward,
         RSA.ID(
           RSA(query.bounded indexOf r1arg0),
           RSA(query.bounded indexOf r2arg0)
         ),
         RSA.Congruent(r1arg2, r2arg2),
+        role1 << Backward,
+        role2 << Backward,
         not(RSA.NI(r1arg2))
       )
 
diff --git a/src/main/scala/uk/ac/ox/cs/rsacomb/Main.scala b/src/main/scala/uk/ac/ox/cs/rsacomb/Main.scala
index f54884f..bf96a31 100644
--- a/src/main/scala/uk/ac/ox/cs/rsacomb/Main.scala
+++ b/src/main/scala/uk/ac/ox/cs/rsacomb/Main.scala
@@ -1,37 +1,41 @@
 package uk.ac.ox.cs.rsacomb
 
-/* Java imports */
 import java.io.File
 import java.util.HashMap
 import scala.collection.JavaConverters._
-
 import tech.oxfordsemantic.jrdfox.client.UpdateType
-import tech.oxfordsemantic.jrdfox.logic.sparql.statement.SelectQuery
 import tech.oxfordsemantic.jrdfox.logic.expression.{IRI, Term}
+import tech.oxfordsemantic.jrdfox.logic.sparql.statement.SelectQuery
 
-/* Local imports */
 import util.{Logger, RDFoxUtil, RSA}
 import sparql.ConjunctiveQuery
 
+/** Entry point of the program.
+  *
+  * The executable expects a SPARQL query and a non-empty sequence of
+  * ontology files as arguments. The query file is expected to contain
+  * exactly one query, while the ontology files will be programmatically
+  * merged in a single ontology.
+  *
+  * @todo better arguments handling is needed. Look into some library
+  * for this.
+  * @todo at the moment the input ontology is assumed to be Horn-ALCHOIQ.
+  * This might not be the case.
+  */
 object RSAComb extends App {
 
   val help: String = """
   rsacomb - combined approach for CQ answering for RSA ontologies.
 
   USAGE
-    rsacomb <query> <ontology> ...
+    rsacomb <query> <ontology> [...]
 
   where
-    - query: a (single) SPARQL query file.
-    - ontology: one or more ontologies.
+    - query: path to a file containing a single SPARQL query
+    - ontology: one or more ontology files
 
   """
 
-  /* Simple arguments handling
-   *
-   * TODO: use something better later on
-   */
-
   if (args.length < 2) {
     println(help)
     sys.exit;
@@ -46,52 +50,41 @@ object RSAComb extends App {
     sys.exit;
   }
 
-  /* TODO: It might be required to check if the ontology in input is
-   * Horn-ALCHOIQ. At the moment we are assuming this is always the
-   * case.
-   */
-
   val ontology = RSAOntology(ontoPaths: _*)
   if (ontology.isRSA) {
 
     Logger print "Ontology is RSA!"
 
-    /** Read SPARQL query from file */
-    val strQuery = RDFoxUtil.loadQueryFromFile(queryPath.getAbsoluteFile)
-    val query = ConjunctiveQuery parse strQuery
+    val query = RDFoxUtil.loadQueryFromFile(queryPath.getAbsoluteFile)
 
-    query match {
+    ConjunctiveQuery.parse(query) match {
       case Some(query) => {
         val answers = ontology ask query
         Logger.print(s"$answers", Logger.QUIET)
-        Logger print s"Number of answer: ${answers.length} (${answers.lengthWithMultiplicity})"
-
-        val unfiltered = ontology askUnfiltered query
-        unfiltered map { u =>
-          Logger.print(
-            s"Number of unfiltered answers: ${u.length} (${u.map(_._1).sum}).",
-            Logger.DEBUG
-          )
-          //u foreach println
-          val spurious = {
-            val sp =
-              RDFoxUtil.buildDescriptionQuery("SP", query.variables.length)
-            ontology.queryDataStore(query, sp, RSA.Prefixes)
-          }
-          spurious map { s =>
-            Logger.print(
-              s"Number of spurious answers: ${s.length} (${s.map(_._1).sum})",
-              Logger.DEBUG
-            )
-            //s foreach println
-            val perc =
-              if (u.length > 0) (s.length / u.length.toFloat) * 100 else 0
-            Logger.print(
-              s"Percentage of spurious answers: $perc%",
-              Logger.DEBUG
-            )
+        Logger print s"Number of answers: ${answers.length} (${answers.lengthWithMultiplicity})"
+
+        /* Additional DEBUG information */
+        if (Logger.level >= Logger.DEBUG) {
+          /* Unfiltered rules */
+          val unfiltered = ontology askUnfiltered query
+          unfiltered map { u =>
+            Logger print s"Number of unfiltered answers: ${u.length} (${u.map(_._1).sum})."
+
+            /* Spurious answers */
+            val spurious = {
+              val variables = query.variables.length
+              val sp = RDFoxUtil.buildDescriptionQuery("SP", variables)
+              ontology.queryDataStore(query, sp, RSA.Prefixes)
+            }
+            spurious map { s =>
+              Logger print s"Number of spurious answers: ${s.length} (${s.map(_._1).sum})"
+
+              /* Spurious/unfiltered percentage */
+              val perc =
+                if (u.length > 0) (s.length / u.length.toFloat) * 100 else 0
+              Logger print s"Percentage of spurious answers: $perc%"
+            }
           }
-
         }
       }
       case None =>
diff --git a/src/main/scala/uk/ac/ox/cs/rsacomb/RSAOntology.scala b/src/main/scala/uk/ac/ox/cs/rsacomb/RSAOntology.scala
index 4ac5a77..8d5bf4c 100644
--- a/src/main/scala/uk/ac/ox/cs/rsacomb/RSAOntology.scala
+++ b/src/main/scala/uk/ac/ox/cs/rsacomb/RSAOntology.scala
@@ -114,9 +114,9 @@ class RSAOntology(val ontology: OWLOntology) {
 
   val axioms: List[OWLLogicalAxiom] = abox ::: tbox ::: rbox
 
-  Logger.print(s"Original TBox: ${tbox.length} axioms", Logger.DEBUG)
-  Logger.print(s"Original RBox: ${rbox.length} axioms", Logger.DEBUG)
-  Logger.print(s"Original ABox: ${abox.length} axioms", Logger.DEBUG)
+  Logger.print(s"Original TBox: ${tbox.length}", Logger.DEBUG)
+  Logger.print(s"Original RBox: ${rbox.length}", Logger.DEBUG)
+  Logger.print(s"Original ABox: ${abox.length}", Logger.DEBUG)
 
   /* Retrieve individuals in the original ontology
    */
@@ -358,16 +358,18 @@ class RSAOntology(val ontology: OWLOntology) {
 
       //data.beginTransaction(TransactionType.READ_WRITE)
 
-      Logger print s"Canonical model: ${canon.rules.length} rules"
+      Logger print s"Canonical model rules: ${canon.rules.length}"
       RDFoxUtil.addRules(data, this.canonicalModel.rules)
 
-      Logger print s"Canonical model: ${canon.facts.length} facts"
+      Logger print s"Canonical model facts: ${canon.facts.length}"
       RDFoxUtil.addFacts(data, this.canonicalModel.facts)
 
-      Logger print s"Filtering program: ${filter.facts.length} facts"
+      RDFoxUtil printStatisticsFor data
+
+      Logger print s"Filtering program facts: ${filter.facts.length}"
       RDFoxUtil.addFacts(data, filter.facts)
 
-      Logger print s"Filtering program: ${filter.rules.length} rules"
+      Logger print s"Filtering program rules: ${filter.rules.length}"
       RDFoxUtil.addRules(data, filter.rules)
 
       //data.commitTransaction()
diff --git a/src/main/scala/uk/ac/ox/cs/rsacomb/util/Logger.scala b/src/main/scala/uk/ac/ox/cs/rsacomb/util/Logger.scala
index 74797a2..56e9de0 100644
--- a/src/main/scala/uk/ac/ox/cs/rsacomb/util/Logger.scala
+++ b/src/main/scala/uk/ac/ox/cs/rsacomb/util/Logger.scala
@@ -38,7 +38,7 @@ object Logger {
     print(s"$desc (START)", lvl)
     val result = expr
     val t1 = System.currentTimeMillis()
-    print(s"$desc (END): ${(t1 - t0).toFloat / 1000}s", lvl)
+    print(s"$desc (END): ${(t1 - t0).toFloat / 1000}", lvl)
     result
   }