rampage

draeger-lab · Aug 14, 2024 · 8829f61 · 8829f61
1 parent f294dbf
commit 8829f61
Show file tree

Hide file tree

Showing 129 changed files with 2,164 additions and 2,570 deletions.
diff --git a/build.gradle b/build.gradle
@@ -79,17 +79,18 @@ repositories {
     // snapshot versions of artifacts (i.e. dependency JARs)
     // Sonatype is the company behind the Maven Central repository
     maven { url "https://oss.sonatype.org/content/repositories/snapshots" }
-    // resolve dependencies from a local directory
-    flatDir {
-	dirs "lib/de/zbit/SysBio/1390"
+    maven {
+        name "biodata-reposilite"
+        url "https://biodata.informatik.uni-halle.de/maven/releases"
     }
+
 }
 
 /** ==== Dependencies ==== **/
 // variables used in dependency definitions
 def jacksonVersion = "2.17.1"
 def jupiterVersion = "5.10.2"
-def testContainersVersion = "1.19.8"
+def testContainersVersion = "1.20.1"
 
 /*
  - `implementation`: Dependencies required for the main source set.
@@ -99,28 +100,35 @@ def testContainersVersion = "1.19.8"
  - `testRuntimeOnly`: Dependencies required only at test runtime, not at compile time.
  */
 dependencies {
+    // systems biology tools
+    implementation "de.zbit.SysBio:SysBio:1390"
     // reading, writing, and manipulating SBML files
-    implementation "org.sbml.jsbml:jsbml:1.6.1"
-    // local dependency for systems biology tools
-    implementation "de.zbit:SysBio:1390"
-    // official JDBC driver for PostgreSQL
-    implementation "org.postgresql:postgresql:42.7.3"
+    implementation("org.sbml.jsbml:jsbml:1.6.1") {
+        exclude group: "org.apache.logging.log4j", module: "log4j-core"
+        exclude group: "org.apache.logging.log4j", module: "log4j-slf4j-impl"
+    }
     // ontology support within BioJava
     implementation "org.biojava:biojava-ontology:7.1.1"
+    // handle Combine Archives (packages of biological models)
+    implementation "de.uni-rostock.sbi:CombineArchive:1.4.1"
     // interacting with MATLAB files
     implementation 'us.hebi.matlab.mat:mfl-core:0.5.15'
+
+    // JDBC connection pool
+    implementation "com.zaxxer:HikariCP:5.1.0"
+    // https://mvnrepository.com/artifact/commons-io/commons-io
+    implementation group: 'commons-io', name: 'commons-io', version: '2.16.1'
+    // https://mvnrepository.com/artifact/org.apache.commons/commons-lang3
+    implementation group: 'org.apache.commons', name: 'commons-lang3', version: '3.15.0'
     // JSON processor
     implementation "com.fasterxml.jackson.core:jackson-core:${jacksonVersion}"
-    // data binding for Jackson
+// data binding for Jackson
     implementation "com.fasterxml.jackson.core:jackson-databind:${jacksonVersion}"
     // Java port of HTML Tidy, for cleaning up malformed HTML
     implementation "net.sf.jtidy:jtidy:r938"
-    // handle Combine Archives (packages of biological models)
-    implementation "de.uni-rostock.sbi:CombineArchive:1.4.1"
-    // JDBC connection pool
-    implementation "com.zaxxer:HikariCP:5.1.0"
-
-    implementation 'com.opentable.components:otj-pg-embedded:1.0.1'
+    // official JDBC driver for PostgreSQL
+    implementation "org.postgresql:postgresql:42.7.3"
+
     // Test implementation dependencies
     // API for writing tests with JUnit 5
     testImplementation "org.junit.jupiter:junit-jupiter-api:${jupiterVersion}"
@@ -133,6 +141,7 @@ dependencies {
     // runtime engine for executing tests with JUnit 5
     testRuntimeOnly 'org.junit.platform:junit-platform-launcher:1.10.2'
     testRuntimeOnly "org.junit.jupiter:junit-jupiter-engine:${jupiterVersion}"
+
 }
 
 /** === Configuration for pre-existing Tasks === **/
@@ -162,6 +171,7 @@ tasks.withType(JavaCompile).configureEach {
 
 // config for all test tasks
 tasks.withType(Test).configureEach {
+    environment "TESTCONTAINERS_RYUK_DISABLED",     "true"
     useJUnitPlatform()
     testLogging {
         showStandardStreams = true

diff --git a/...ain/java/edu/ucsd/sbrg/eco/ECOparser.java → ...a/edu/ucsd/sbrg/ECOParserCLILauncher.java b/...ain/java/edu/ucsd/sbrg/eco/ECOparser.java → ...a/edu/ucsd/sbrg/ECOParserCLILauncher.java
@@ -1,4 +1,4 @@
-package edu.ucsd.sbrg.eco;
+package edu.ucsd.sbrg;
 
 import java.io.BufferedReader;
 import java.io.IOException;
@@ -7,14 +7,19 @@
 import java.text.ParseException;
 import java.util.Set;
 
+import edu.ucsd.sbrg.eco.DAG;
+import edu.ucsd.sbrg.eco.Node;
 import org.biojava.nbio.ontology.Ontology;
 import org.biojava.nbio.ontology.Term;
 import org.biojava.nbio.ontology.Triple;
 import org.biojava.nbio.ontology.io.OboParser;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
-public class ECOparser {
+public class ECOParserCLILauncher {
 
   private static Ontology ontology;
+  private static final Logger logger = LoggerFactory.getLogger(ECOParserCLILauncher.class);
 
   /**
    * @see <a href="https://github.com/draeger-lab/ModelPolisher/issues/5">Issue 5</>
@@ -57,8 +62,8 @@ public static String getECOTermFromScore(int confidenceScore) {
   }
 
 
-  public static void main(String[] args) {
-    ECOparser.parseECO();
+  public static void main(String[] args) throws IOException, ParseException {
+    ECOParserCLILauncher.parseECO();
   }
 
 
@@ -82,18 +87,16 @@ public static int getConfidenceScoreForTerm(String query) {
   }
 
 
-  public static void parseECO() {
+  public static void parseECO() throws IOException, ParseException {
     String ecoName = "Evidence and Conclusion Ontology";
     String ecoDesc = "TODO";
     OboParser parser = new OboParser();
-    try (InputStream inputStream = ECOparser.class.getResourceAsStream("eco.obo");
-        BufferedReader oboReader = new BufferedReader(new InputStreamReader(inputStream))) {
+    try (InputStream inputStream = ECOParserCLILauncher.class.getResourceAsStream("eco.obo");
+         BufferedReader oboReader = new BufferedReader(new InputStreamReader(inputStream))) {
       ontology = parser.parseOBO(oboReader, ecoName, ecoDesc);
       Term rootTerm = ontology.getTerm("ECO:0000000");
       DAG ontologyDAG = new DAG(rootTerm);
       traverse(ontologyDAG.getRoot());
-    } catch (IOException | ParseException e) {
-      e.printStackTrace();
     }
   }
 

diff --git a/src/main/java/edu/ucsd/sbrg/ModelPolisherCLILauncher.java b/src/main/java/edu/ucsd/sbrg/ModelPolisherCLILauncher.java
@@ -9,26 +9,29 @@
 import java.net.URISyntaxException;
 import java.net.URL;
 import java.sql.SQLException;
-import java.util.Calendar;
-import java.util.LinkedList;
+import java.util.*;
 import java.util.List;
-import java.util.ResourceBundle;
 import java.util.concurrent.TimeUnit;
-import java.util.logging.Logger;
 
+import de.zbit.util.prefs.SBProperties;
 import edu.ucsd.sbrg.annotation.AnnotationException;
 import edu.ucsd.sbrg.annotation.adb.ADBSBMLAnnotator;
 import edu.ucsd.sbrg.annotation.bigg.BiGGSBMLAnnotator;
 import edu.ucsd.sbrg.io.*;
 import edu.ucsd.sbrg.parameters.CommandLineParameters;
 import edu.ucsd.sbrg.db.adb.AnnotateDB;
 import edu.ucsd.sbrg.db.bigg.BiGGDB;
+import edu.ucsd.sbrg.parameters.ModelPolisherOptions;
 import edu.ucsd.sbrg.polishing.SBMLPolisher;
 import edu.ucsd.sbrg.reporting.PolisherProgressBar;
 import edu.ucsd.sbrg.reporting.ProgressInitialization;
 import edu.ucsd.sbrg.reporting.ProgressObserver;
 import edu.ucsd.sbrg.resolver.Registry;
 import edu.ucsd.sbrg.resolver.identifiersorg.IdentifiersOrg;
+import edu.ucsd.sbrg.validation.ModelValidator;
+import edu.ucsd.sbrg.validation.ModelValidatorException;
+import org.apache.commons.io.FileUtils;
+import org.apache.commons.lang3.tuple.Pair;
 import org.sbml.jsbml.Model;
 import org.sbml.jsbml.SBMLDocument;
 
@@ -37,11 +40,12 @@
 import de.zbit.util.ResourceManager;
 import de.zbit.util.logging.LogOptions;
 import de.zbit.util.prefs.KeyProvider;
-import de.zbit.util.prefs.SBProperties;
 import edu.ucsd.sbrg.db.adb.AnnotateDBOptions;
 import edu.ucsd.sbrg.db.bigg.BiGGDBOptions;
 import org.sbml.jsbml.ext.fbc.FBCConstants;
 import org.sbml.jsbml.ext.fbc.FBCModelPlugin;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
 /**
  * The ModelPolisher class is the entry point of this application.
@@ -68,7 +72,7 @@ public class ModelPolisherCLILauncher extends Launcher {
 
   private static final ResourceBundle baseBundle = ResourceManager.getBundle("edu.ucsd.sbrg.Messages");
   private static final ResourceBundle MESSAGES = ResourceManager.getBundle("edu.ucsd.sbrg.polisher.Messages");
-  private static final Logger logger = Logger.getLogger(ModelPolisherCLILauncher.class.getName());
+  private static final Logger logger = LoggerFactory.getLogger(ModelPolisherCLILauncher.class);
 
   private CommandLineParameters parameters;
   private Registry registry;
@@ -117,37 +121,39 @@ public void commandLineMode(AppConf appConf) {
     parameters = new CommandLineParameters(args);
     registry = new IdentifiersOrg();
 
-    if (parameters.annotationParameters().biggAnnotationParameters().annotateWithBiGG()) {
-      this.bigg = new BiGGDB(parameters.annotationParameters().biggAnnotationParameters().dbParameters());
-    }
-
-    if (parameters.annotationParameters().adbAnnotationParameters().addADBAnnotations()) {
-      this.adb = new AnnotateDB(parameters.annotationParameters().adbAnnotationParameters().dbParameters());
-    }
-
     try {
-      var input = parameters.input();
-      var output = parameters.output();
+      validateIOParameters();
 
-      // Check if the input exists, throw an exception if it does not
-      if (!input.exists()) {
-        throw new IOException(format(MESSAGES.getString("READ_FILE_ERROR"),
-                input.toString()));
+      if (parameters.annotationParameters().biggAnnotationParameters().annotateWithBiGG()) {
+        this.bigg = new BiGGDB(parameters.annotationParameters().biggAnnotationParameters().dbParameters());
       }
 
-      // If the output is not a directory but the input is, log an error and return
-      if (!output.isDirectory() && input.isDirectory()) {
-        throw new IOException(format(MESSAGES.getString("WRITE_DIR_TO_FILE_ERROR"),
-                input.getAbsolutePath(),
-                output.getAbsolutePath()));
+      if (parameters.annotationParameters().adbAnnotationParameters().addADBAnnotations()) {
+        this.adb = new AnnotateDB(parameters.annotationParameters().adbAnnotationParameters().dbParameters());
       }
 
-      // Ensure the output directory or file's parent directory exists
-      SBMLFileUtils.checkCreateOutDir(output);
-
-      batchProcess(input, parameters.output());
+      var inputFiles = FileUtils.listFiles(parameters.input(), new String[]{"xml", "sbml", "json", "mat"}, true);
+      List<Pair<File, File>> inputOutputPairs = new ArrayList<>();
+
+      if (parameters.input().isDirectory()) {
+        for (var input: inputFiles) {
+          inputOutputPairs.add(Pair.of(input, SBMLFileUtils.getOutputFileName(input, parameters.output())));
+        }
+
+        // TODO: this is a placeholder for a parallel implementation
+        for (var pair: inputOutputPairs) {
+            try {
+                processFile(pair.getLeft(), pair.getRight());
+            } catch (ModelReaderException e) {
+                logger.debug(format("Skipping unreadable file \"{0}\".", pair.getLeft()));
+            }
+        }
+      }
+      else {
+        processFile(parameters.input(), parameters.output());
+      }
 
-    } catch (ModelReaderException | ModelValidatorException | ModelWriterException | IOException |
+    } catch (ModelValidatorException | ModelWriterException | IOException |
              AnnotationException e) {
       // TODO: produce some user-friendly output and log to a file that can be provided for trouble-shooting
       throw new RuntimeException(e);
@@ -157,52 +163,48 @@ public void commandLineMode(AppConf appConf) {
     } catch (SQLException e) {
       // TODO: produce some user-friendly output and log to a file that can be provided for trouble-shooting
       throw new RuntimeException(e);
+    } catch (ModelReaderException e) {
+        throw new RuntimeException(e);
     }
   }
 
+  private void validateIOParameters() throws IOException {
+    var input = parameters.input();
+    var output = parameters.output();
 
-  /**
-   * Processes the specified input and output paths. If the input is a directory, it recursively processes each file within.
-   *
-   * @param input  Path to the input file or directory to be processed. This should correspond to {@link CommandLineParameters#input()}.
-   * @param output Path to the output file or directory where processed files should be saved. This should correspond to {@link CommandLineParameters#output()}.
-   */
-  private void batchProcess(File input, File output) throws ModelReaderException, ModelWriterException, ModelValidatorException, AnnotationException, SQLException {
-    // If the input is a directory, process each file within it
-    if (input.isDirectory()) {
-      File[] files = input.listFiles();
-
-      if (files == null || files.length < 1) {
-        logger.info(MESSAGES.getString("NO_FILES_ERROR"));
-        return;
-      }
-      // Recursively process each file in the directory
-      for (File file : files) {
-        File target = SBMLFileUtils.getOutputFileName(file, output);
-        batchProcess(file, target);
-      }
-    } else {
-      // NOTE: input is a single file, but output can be a file or a directory
-      // Adjust output file name if the output is a directory
-      var newOutput = output.isDirectory() ? SBMLFileUtils.getOutputFileName(input, output) : output;
+    // Check if the input exists, throw an exception if it does not
+    if (!input.exists()) {
+      throw new IOException(format(MESSAGES.getString("READ_FILE_ERROR"),
+              input.toString()));
+    }
 
-      processFile(input, newOutput);
+    // If the input is a directory but the output is not, exit with error
+    if (input.isDirectory() && !output.isDirectory()) {
+      throw new IOException(format(MESSAGES.getString("WRITE_DIR_TO_FILE_ERROR"),
+              input.getAbsolutePath(),
+              output.getAbsolutePath()));
     }
+
+    // If the output is a directory but the input is not, exit with error
+    if (output.isDirectory() && !input.isDirectory()) {
+      throw new IOException(format("Output \"{0}\" is a directory, but Input \"{1}\" is not",
+              output.getAbsolutePath(),
+              input.getAbsolutePath()));
+    }
+
+    // Ensure the output directory or file's parent directory exists
+    SBMLFileUtils.checkCreateOutDir(output);
   }
 
 
   private void processFile(File input, File output) throws ModelReaderException, ModelWriterException, ModelValidatorException, AnnotationException, SQLException {
     long startTime = System.currentTimeMillis();
 
     SBMLDocument doc = new ModelReader(parameters.sboParameters(), registry).read(input);
-    // TODO: we should be doing better sanity checking here; e.g.: just validate the SBML
-    if (doc == null) return;
+
+    // TODO: hier wäre es jetzt angebracht das Ding zu validieren, geht aber nicht, weil es keinen Validator in JSBML gibt
+
     Model model = doc.getModel();
-    if (model == null) {
-      logger.severe(MESSAGES.getString("MODEL_MISSING"));
-      // TODO: handle this better
-      throw new RuntimeException(MESSAGES.getString("MODEL_MISSING"));
-    }
 
     List<ProgressObserver> polishingObservers = List.of(new PolisherProgressBar());
     int count = getPolishingTaskCount(model);
@@ -240,6 +242,7 @@ private void processFile(File input, File output) throws ModelReaderException, M
 
     output = new ModelWriter(parameters.outputType()).write(doc, output);
 
+    // TODO: das ist keine anständige Validierung!
     if (parameters.SBMLValidation()) {
       var mv = new ModelValidator();
       // use offline validation