org.molgenis.data.csv.CsvRepository Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of molgenis-data-csv Show documentation
Comma-separated data reading/writing.
There is a newer version: 8.4.5
package org.molgenis.data.csv;

import static java.util.Objects.requireNonNull;
import static org.molgenis.data.meta.AttributeType.STRING;

import com.google.common.collect.Iterables;
import java.io.File;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;
import java.util.Set;
import javax.annotation.CheckForNull;
import javax.annotation.Nullable;
import javax.validation.constraints.NotNull;
import org.molgenis.data.Entity;
import org.molgenis.data.RepositoryCapability;
import org.molgenis.data.file.processor.CellProcessor;
import org.molgenis.data.meta.model.Attribute;
import org.molgenis.data.meta.model.AttributeFactory;
import org.molgenis.data.meta.model.EntityType;
import org.molgenis.data.meta.model.EntityTypeFactory;
import org.molgenis.data.support.AbstractRepository;
import org.springframework.util.StringUtils;

/**
 * Repository implementation for csv files.
 *
 * The filename without the extension is considered to be the entityname
 */
public class CsvRepository extends AbstractRepository {
  private final File file;
  private final EntityTypeFactory entityTypeFactory;
  private final AttributeFactory attrMetaFactory;
  private final String sheetName;
  private List cellProcessors;
  private EntityType entityType;
  private Character separator = null;

  public CsvRepository(
      File file,
      EntityTypeFactory entityTypeFactory,
      AttributeFactory attrMetaFactory,
      @Nullable @CheckForNull List cellProcessors,
      Character separator) {
    this(
        file,
        entityTypeFactory,
        attrMetaFactory,
        StringUtils.stripFilenameExtension(file.getName()),
        cellProcessors);
    this.separator = separator;
  }

  public CsvRepository(
      File file,
      EntityTypeFactory entityTypeFactory,
      AttributeFactory attrMetaFactory,
      @Nullable @CheckForNull List cellProcessors) {
    this(
        file,
        entityTypeFactory,
        attrMetaFactory,
        StringUtils.stripFilenameExtension(file.getName()),
        cellProcessors);
  }

  public CsvRepository(
      File file,
      EntityTypeFactory entityTypeFactory,
      AttributeFactory attrMetaFactory,
      String sheetName,
      @Nullable @CheckForNull List cellProcessors) {
    this.file = file;
    this.entityTypeFactory = requireNonNull(entityTypeFactory);
    this.attrMetaFactory = requireNonNull(attrMetaFactory);
    this.sheetName = sheetName;
    this.cellProcessors = cellProcessors;
  }

  @NotNull
  @Override
  public Iterator iterator() {
    return new CsvIterator(file, sheetName, cellProcessors, separator, getEntityType());
  }

  public EntityType getEntityType() {
    if (entityType == null) {
      entityType = entityTypeFactory.create(sheetName);

      for (String attrName :
          new CsvIterator(file, sheetName, null, separator).getColNamesMap().keySet()) {
        Attribute attr = attrMetaFactory.create().setName(attrName).setDataType(STRING);
        entityType.addAttribute(attr);
      }
    }

    return entityType;
  }

  public void addCellProcessor(CellProcessor cellProcessor) {
    if (cellProcessors == null) cellProcessors = new ArrayList<>();
    cellProcessors.add(cellProcessor);
  }

  @Override
  public Set getCapabilities() {
    return Collections.emptySet();
  }

  @Override
  public long count() {
    return Iterables.size(this);
  }
}