Merge pull request #4 in PGX/pgx-samples from feature/paper-classific…

…ation to master

* commit 'ec86daa8e47382432535e1b444948d6a0876d33f':
  Add Cora example

README.md

@@ -12,6 +12,7 @@
 6. [Article Ranking](#article-ranking)
 7. [Movie Recommendation](#movie-recommender)
 8. [Entity Linking](#entity-linking)
+9. [Research Paper Classification](#paper-classification)
 
 ****
 
@@ -59,3 +60,7 @@ More details regarding this application are available [here](movie-recommendatio
 ## Entity Linking <a name="entity-linking"></a>
 Entity Linking allows to connect Named Entities (for example, names of famous people) to their Wikipedia/DBpedia page.
  This application leverages vertex embeddings to provide high-quality results. More details available [here](entity-linking/README.md) and in our [paper](https://dl.acm.org/citation.cfm?doid=3327964.3328499).
+
+## Research Paper Classification <a name="paper-classification"></a>
+This application demonstrates how graph data can be used to enhance classification performance of a research paper classifier.
+More details regarding this application are available [here](paper-classification/README.md).

paper-classification/README.md

@@ -0,0 +1,40 @@
+Note: this example makes use of the new `SupervisedGraphWise` classifer, which is available in PGX 19.4.0.
+# Research Paper Classification - Cora Dataset
+Graph analytics often allows you to leverage relational information that other, classical models simply have no
+effective way of capturing. One example of such data is the citation network of research papers.
+
+## The Cora dataset
+The Cora dataset consists of 2708 machine-learning research papers which have been labeled with one of seven topics.
+Each paper has a 1433 dimensional binary feature vector, where 0/1 indicates the absence/presence of selected words from
+the dictionary. Additionally, the dataset provides the citation graph, where each edge represents a citation between two
+research papers.
+
+To run this example, download [the dataset](https://linqs-data.soe.ucsc.edu/public/lbc/cora.tgz) and extract the
+files `cora.cites` and `cora.content` into the `data/graph/cora` directory. 
+`cora.content` contains the vertex (research paper) feature vectors, and `cora.cites` contains the edges.
+
+## Running the example
+The example can be run using the command `./gradlew run`. This will perform the following steps:
+1. Load the Cora graph into the PGX server
+2. Create a train-test split
+3. Train a `SupervisedGraphWise` model on the training graph
+4. Evaluate the trained model on the unseen test nodes
+5. Infer and save embeddings for all vertices into `embeddings.csv`
+
+## Evaluating the results
+We can see that the model performs quite well on the test set: the following is example evaluation output:
+
+|--------------------|--------------------|--------------------|--------------------|
+| Accuracy           | Precision          | Recall             | F1-Score           |
+|--------------------|--------------------|--------------------|--------------------|
+| 0.869              | 0.867              | 0.846              | 0.854              |
+|--------------------|--------------------|--------------------|--------------------|
+
+The hyper parameters of the model have been left to the default settings, but can of course be tuned to optimize the
+performance.
+
+## Visualizing the Embeddings
+The embeddings themselves can be used for other downstream tasks, such as cluster or link prediction. To visualize them
+in an understandable manner, we can generate and plot t-SNE vectors for the vertices. The following figure is the result.
+
+![Cora Embeddings t-SNE plot](cora_tsne.png "Cora Embeddings TNSE plot")

paper-classification/build.gradle

@@ -0,0 +1,34 @@
+apply plugin: 'java'
+jar.enabled = false
+
+ext {
+  mainClassName = 'Apps.CoraApp'
+}
+
+def pgxDir = "$projectDir/../libs/pgx-19.4.0-SNAPSHOT"
+def commonDeps = "$pgxDir/shared-lib/common"
+def embeddedDeps = "$pgxDir/shared-lib/embedded"
+def serverDeps = "$pgxDir/shared-lib/server"
+def thirdPartyDeps = "$pgxDir/shared-lib/third-party"
+def smCommonDeps = "$pgxDir/shared-memory/common"
+def smClientDeps = "$pgxDir/shared-memory/client"
+def smEmbeddedDeps = "$pgxDir/shared-memory/embedded"
+def smThirdPartyDeps = "$pgxDir/shared-memory/third-party"
+
+dependencies {
+  compile fileTree(dir: commonDeps, includes: ['*.jar'])
+  compile fileTree(dir: embeddedDeps, includes: ['*.jar'])
+  compile fileTree(dir: serverDeps, includes: ['*.jar'])
+  compile fileTree(dir: thirdPartyDeps, includes: ['*.jar'])
+  compile fileTree(dir: smCommonDeps, includes: ['*.jar'])
+  compile fileTree(dir: smClientDeps, includes: ['*.jar'])
+  compile fileTree(dir: smEmbeddedDeps, includes: ['*.jar'])
+  compile fileTree(dir: smThirdPartyDeps, includes: ['*.jar'])
+}
+
+task run(type: JavaExec, dependsOn: jar) {
+  jvmArgs = ['-Dpgx_conf=conf/pgx.conf']
+  classpath = sourceSets.main.runtimeClasspath
+  main = mainClassName
+  args = [project.projectDir]
+}

paper-classification/conf/pgx.conf

@@ -0,0 +1,5 @@
+{
+  "datasource_dir_whitelist" : ["PUT_THE_ABSOLUTE_PATH_TO_DATA_DIRECTORY_HERE"],
+  "allow_local_filesystem": true,
+  "tmp_dir": "tmp_data"
+}

paper-classification/data/graph/cora/DATA.md

@@ -0,0 +1 @@
+Place the files `cora.content` and `cora.cites` here. See `README.md` for more information.

paper-classification/gradle/wrapper/gradle-wrapper.properties

@@ -0,0 +1,5 @@
+distributionBase=GRADLE_USER_HOME
+distributionPath=wrapper/dists
+zipStoreBase=GRADLE_USER_HOME
+zipStorePath=wrapper/dists
+distributionUrl=https\://services.gradle.org/distributions/gradle-5.2.1-bin.zip

paper-classification/gradlew

@@ -0,0 +1,172 @@
+#!/usr/bin/env sh
+
+##############################################################################
+##
+##  Gradle start up script for UN*X
+##
+##############################################################################
+
+# Attempt to set APP_HOME
+# Resolve links: $0 may be a link
+PRG="$0"
+# Need this for relative symlinks.
+while [ -h "$PRG" ] ; do
+    ls=`ls -ld "$PRG"`
+    link=`expr "$ls" : '.*-> \(.*\)$'`
+    if expr "$link" : '/.*' > /dev/null; then
+        PRG="$link"
+    else
+        PRG=`dirname "$PRG"`"/$link"
+    fi
+done
+SAVED="`pwd`"
+cd "`dirname \"$PRG\"`/" >/dev/null
+APP_HOME="`pwd -P`"
+cd "$SAVED" >/dev/null
+
+APP_NAME="Gradle"
+APP_BASE_NAME=`basename "$0"`
+
+# Add default JVM options here. You can also use JAVA_OPTS and GRADLE_OPTS to pass JVM options to this script.
+DEFAULT_JVM_OPTS=""
+
+# Use the maximum available, or set MAX_FD != -1 to use that value.
+MAX_FD="maximum"
+
+warn () {
+    echo "$*"
+}
+
+die () {
+    echo
+    echo "$*"
+    echo
+    exit 1
+}
+
+# OS specific support (must be 'true' or 'false').
+cygwin=false
+msys=false
+darwin=false
+nonstop=false
+case "`uname`" in
+  CYGWIN* )
+    cygwin=true
+    ;;
+  Darwin* )
+    darwin=true
+    ;;
+  MINGW* )
+    msys=true
+    ;;
+  NONSTOP* )
+    nonstop=true
+    ;;
+esac
+
+CLASSPATH=$APP_HOME/gradle/wrapper/gradle-wrapper.jar
+
+# Determine the Java command to use to start the JVM.
+if [ -n "$JAVA_HOME" ] ; then
+    if [ -x "$JAVA_HOME/jre/sh/java" ] ; then
+        # IBM's JDK on AIX uses strange locations for the executables
+        JAVACMD="$JAVA_HOME/jre/sh/java"
+    else
+        JAVACMD="$JAVA_HOME/bin/java"
+    fi
+    if [ ! -x "$JAVACMD" ] ; then
+        die "ERROR: JAVA_HOME is set to an invalid directory: $JAVA_HOME
+
+Please set the JAVA_HOME variable in your environment to match the
+location of your Java installation."
+    fi
+else
+    JAVACMD="java"
+    which java >/dev/null 2>&1 || die "ERROR: JAVA_HOME is not set and no 'java' command could be found in your PATH.
+
+Please set the JAVA_HOME variable in your environment to match the
+location of your Java installation."
+fi
+
+# Increase the maximum file descriptors if we can.
+if [ "$cygwin" = "false" -a "$darwin" = "false" -a "$nonstop" = "false" ] ; then
+    MAX_FD_LIMIT=`ulimit -H -n`
+    if [ $? -eq 0 ] ; then
+        if [ "$MAX_FD" = "maximum" -o "$MAX_FD" = "max" ] ; then
+            MAX_FD="$MAX_FD_LIMIT"
+        fi
+        ulimit -n $MAX_FD
+        if [ $? -ne 0 ] ; then
+            warn "Could not set maximum file descriptor limit: $MAX_FD"
+        fi
+    else
+        warn "Could not query maximum file descriptor limit: $MAX_FD_LIMIT"
+    fi
+fi
+
+# For Darwin, add options to specify how the application appears in the dock
+if $darwin; then
+    GRADLE_OPTS="$GRADLE_OPTS \"-Xdock:name=$APP_NAME\" \"-Xdock:icon=$APP_HOME/media/gradle.icns\""
+fi
+
+# For Cygwin, switch paths to Windows format before running java
+if $cygwin ; then
+    APP_HOME=`cygpath --path --mixed "$APP_HOME"`
+    CLASSPATH=`cygpath --path --mixed "$CLASSPATH"`
+    JAVACMD=`cygpath --unix "$JAVACMD"`
+
+    # We build the pattern for arguments to be converted via cygpath
+    ROOTDIRSRAW=`find -L / -maxdepth 1 -mindepth 1 -type d 2>/dev/null`
+    SEP=""
+    for dir in $ROOTDIRSRAW ; do
+        ROOTDIRS="$ROOTDIRS$SEP$dir"
+        SEP="|"
+    done
+    OURCYGPATTERN="(^($ROOTDIRS))"
+    # Add a user-defined pattern to the cygpath arguments
+    if [ "$GRADLE_CYGPATTERN" != "" ] ; then
+        OURCYGPATTERN="$OURCYGPATTERN|($GRADLE_CYGPATTERN)"
+    fi
+    # Now convert the arguments - kludge to limit ourselves to /bin/sh
+    i=0
+    for arg in "$@" ; do
+        CHECK=`echo "$arg"|egrep -c "$OURCYGPATTERN" -`
+        CHECK2=`echo "$arg"|egrep -c "^-"`                                 ### Determine if an option
+
+        if [ $CHECK -ne 0 ] && [ $CHECK2 -eq 0 ] ; then                    ### Added a condition
+            eval `echo args$i`=`cygpath --path --ignore --mixed "$arg"`
+        else
+            eval `echo args$i`="\"$arg\""
+        fi
+        i=$((i+1))
+    done
+    case $i in
+        (0) set -- ;;
+        (1) set -- "$args0" ;;
+        (2) set -- "$args0" "$args1" ;;
+        (3) set -- "$args0" "$args1" "$args2" ;;
+        (4) set -- "$args0" "$args1" "$args2" "$args3" ;;
+        (5) set -- "$args0" "$args1" "$args2" "$args3" "$args4" ;;
+        (6) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" ;;
+        (7) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" "$args6" ;;
+        (8) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" "$args6" "$args7" ;;
+        (9) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" "$args6" "$args7" "$args8" ;;
+    esac
+fi
+
+# Escape application args
+save () {
+    for i do printf %s\\n "$i" | sed "s/'/'\\\\''/g;1s/^/'/;\$s/\$/' \\\\/" ; done
+    echo " "
+}
+APP_ARGS=$(save "$@")
+
+# Collect all arguments for the java command, following the shell quoting and substitution rules
+eval set -- $DEFAULT_JVM_OPTS $JAVA_OPTS $GRADLE_OPTS "\"-Dorg.gradle.appname=$APP_BASE_NAME\"" -classpath "\"$CLASSPATH\"" org.gradle.wrapper.GradleWrapperMain "$APP_ARGS"
+
+# by default we should be in the correct project dir, but when run from Finder on Mac, the cwd is wrong
+if [ "$(uname)" = "Darwin" ] && [ "$HOME" = "$PWD" ]; then
+  cd "$(dirname "$0")"
+fi
+
+exec "$JAVACMD" "$@"

paper-classification/gradlew.bat

@@ -0,0 +1,84 @@
+@if "%DEBUG%" == "" @echo off
+@rem ##########################################################################
+@rem
+@rem  Gradle startup script for Windows
+@rem
+@rem ##########################################################################
+
+@rem Set local scope for the variables with windows NT shell
+if "%OS%"=="Windows_NT" setlocal
+
+set DIRNAME=%~dp0
+if "%DIRNAME%" == "" set DIRNAME=.
+set APP_BASE_NAME=%~n0
+set APP_HOME=%DIRNAME%
+
+@rem Add default JVM options here. You can also use JAVA_OPTS and GRADLE_OPTS to pass JVM options to this script.
+set DEFAULT_JVM_OPTS=
+
+@rem Find java.exe
+if defined JAVA_HOME goto findJavaFromJavaHome
+
+set JAVA_EXE=java.exe
+%JAVA_EXE% -version >NUL 2>&1
+if "%ERRORLEVEL%" == "0" goto init
+
+echo.
+echo ERROR: JAVA_HOME is not set and no 'java' command could be found in your PATH.
+echo.
+echo Please set the JAVA_HOME variable in your environment to match the
+echo location of your Java installation.
+
+goto fail
+
+:findJavaFromJavaHome
+set JAVA_HOME=%JAVA_HOME:"=%
+set JAVA_EXE=%JAVA_HOME%/bin/java.exe
+
+if exist "%JAVA_EXE%" goto init
+
+echo.
+echo ERROR: JAVA_HOME is set to an invalid directory: %JAVA_HOME%
+echo.
+echo Please set the JAVA_HOME variable in your environment to match the
+echo location of your Java installation.
+
+goto fail
+
+:init
+@rem Get command-line arguments, handling Windows variants
+
+if not "%OS%" == "Windows_NT" goto win9xME_args
+
+:win9xME_args
+@rem Slurp the command line arguments.
+set CMD_LINE_ARGS=
+set _SKIP=2
+
+:win9xME_args_slurp
+if "x%~1" == "x" goto execute
+
+set CMD_LINE_ARGS=%*
+
+:execute
+@rem Setup the command line
+
+set CLASSPATH=%APP_HOME%\gradle\wrapper\gradle-wrapper.jar
+
+@rem Execute Gradle
+"%JAVA_EXE%" %DEFAULT_JVM_OPTS% %JAVA_OPTS% %GRADLE_OPTS% "-Dorg.gradle.appname=%APP_BASE_NAME%" -classpath "%CLASSPATH%" org.gradle.wrapper.GradleWrapperMain %CMD_LINE_ARGS%
+
+:end
+@rem End local scope for the variables with windows NT shell
+if "%ERRORLEVEL%"=="0" goto mainEnd
+
+:fail
+rem Set variable GRADLE_EXIT_CONSOLE if you need the _script_ return code instead of
+rem the _cmd.exe /c_ return code!
+if  not "" == "%GRADLE_EXIT_CONSOLE%" exit 1
+exit /b 1
+
+:mainEnd
+if "%OS%"=="Windows_NT" endlocal
+
+:omega