Newer
Older
}
}
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
/**
* Uploads specified data set for specified plate. The data set code will be returned.
* <p>
* Matlab example:
*
* <pre>
* % Upload data set /path/to/my-data-set with properties DESCRIPTION and NUMBER for
* % plate P005 in space SPACE
* % with data set 201007091122-928 as the parent
* properties = {'DESCRIPTION' 'hello example'; 'NUMBER' 3.14}
* parents = {'201007091122-928' }
* datasetcode = OpenBISScreeningML.uploadDataSetForPlateAndParents('/SPACE/P005', parents, '/path/to/my-data-set', 'HCS_IMAGE', properties)
* </pre>
*
* @param augmentedPlateCode The augmented plate code.
* @param parentDataSetCodeObjects The codes of the parents of this data set
* @param dataSetFilePath Path to the data set file/folder to be uploaded.
* @param dataSetType Data set type.
* @param dataSetProperties A two dimensional array where the first column contains the property
* codes and the second column the corresponding property values.
*/
public static Object uploadDataSetForPlateAndParents(String augmentedPlateCode,
Object[] parentDataSetCodeObjects, String dataSetFilePath, String dataSetType,
Object[][] dataSetProperties)
{
checkLoggedIn();
Plate plateIdentifier = getPlate(augmentedPlateCode);
List<String> dataSetCodes = createStringList(parentDataSetCodeObjects);
File dataSetFile = new File(dataSetFilePath);
if (dataSetFile.exists() == false)
{
throw new RuntimeException("Unknown data set file path '" + dataSetFilePath + "'.");
}
try
{
Map<String, String> map = createMap(dataSetProperties);
IDataSetDss dataSet =
openbis.putDataSet(plateIdentifier, dataSetFile, new NewDataSetMetadataDTO(
dataSetType, map, dataSetCodes));
return dataSet.getCode();
} catch (Exception ex)
{
throw new RuntimeException("Couldn't upload data set for plate '" + augmentedPlateCode
+ "'.", ex);
}
}
/**
* Uploads a data set to the specified experiment, setting the data set parents. The data set
* code will be returned.
* <p>
* Matlab example:
*
* <pre>
* % Upload data set /path/to/my-data-set with property DESCRIPTION and N
* % to experiment E103 in project PROJECT and space SPACE
* % with data set 201007091122-928 as the parent
* properties = {'DESCRIPTION' 'hello example' }
* parents = {'201007091122-928' }
* datasetcode = OpenBISScreeningML.uploadDataSetWithParents('/SPACE/PROJECT/E103', parents, '/path/to/my-data-set', 'HCS_IMAGE', properties)
* </pre>
*
* @param augmentedExperimentCode The augmented experiment code.
* @param parentDataSetCodeObjects The codes of the parents of this data set
* @param dataSetFilePath Path to the data set file/folder to be uploaded.
* @param dataSetType Data set type.
* @param dataSetProperties A two dimensional array where the first column contains the property
* codes and the second column the corresponding property values.
*/
public static Object uploadDataSetForExperimentAndParents(String augmentedExperimentCode,
Object[] parentDataSetCodeObjects, String dataSetFilePath, String dataSetType,
Object[][] dataSetProperties)
{
checkLoggedIn();
ExperimentIdentifier experimentIdentifier =
getExperimentIdentifierOrFail(augmentedExperimentCode);
List<String> dataSetCodes = createStringList(parentDataSetCodeObjects);
File dataSetFile = new File(dataSetFilePath);
if (dataSetFile.exists() == false)
{
throw new RuntimeException("Unknown data set file path '" + dataSetFilePath + "'.");
}
try
{
Map<String, String> map = createMap(dataSetProperties);
IDataSetDss dataSet =
openbis.putDataSet(experimentIdentifier, dataSetFile,
new NewDataSetMetadataDTO(dataSetType, map, dataSetCodes));
return dataSet.getCode();
} catch (Exception ex)
{
throw new RuntimeException("Couldn't upload data set for experiment '"
+ augmentedExperimentCode + "' and parents '"
+ Arrays.toString(parentDataSetCodeObjects) + "'.", ex);
}
}
private static Map<String, String> createMap(Object[][] properties)
{
Map<String, String> map = new HashMap<String, String>();
for (Object[] objects : properties)
{
if (objects.length == 2)
{
Object value = objects[1];
map.put(objects[0].toString(), value == null ? null : value.toString());
}
}
return map;
}
private static List<String> createStringList(Object[] identifiers)
{
List<String> list = new ArrayList<String>();
for (Object identifier : identifiers)
{
list.add(identifier.toString());
}
return list;
}
brinn
committed
//
// Images
//
/**
* Loads the TIFF images for the given well location, all tiles and all channels and stores them
* in temporary files. The temporary files will be removed automatically when the Java Virtual
brinn
committed
* Machine exits.
* <p>
* Matlab example:
*
* <pre>
* % Load the images for all channels of well B10 of plate P005 in space SPACE
* imginfo = OpenBISScreeningML.loadImages('/SPACE/P005', 2, 10)
brinn
committed
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
* % Get the plate-well descriptions of all locations
* imginfo(2,:,3)
* % Show the third image (assuming there are at least three images)
* imtool(imginfo(1,3))
* </pre>
*
* @param plate The augmented plate code
* @param row The row in the plate to get the images for
* @param col The column in the plate to get the images for
* @return <code>{ names of TIFF files, image annotation }</code>
* <p>
* Each of <code>names of TIFF files</code> and <code>image annotation</code> is a
* vector of length of the number of images.
* <p>
* <code>image annotation</code> contains
* <code>{ channel name, tile number, plate well description,
* plate augmented code, plate perm id, plate space code, plate code, row, column,
* experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code }</code>
*/
public static Object[][][] loadImages(String plate, int row, int col)
{
return loadImages(plate, row, col, (String[]) null);
}
/**
* Loads the TIFF images for the given well location, tile number, and all channels and stores
* them in temporary files. The temporary files will be removed automatically when the Java
* Virtual Machine exits.
* <p>
* Matlab example:
*
* <pre>
* % Load the images for all channels of well B10 and tile 3 of plate P005 in space SPACE
* imginfo = OpenBISScreeningML.loadImages('/SPACE/P005', 2, 10, 3)
* % Get the plate-well descriptions of all locations
* imginfo(2,:,3)
* % Show the third image (assuming there are at least three images)
* imtool(imginfo(1,3))
* </pre>
*
* @param plate The augmented plate code
* @param row The row in the plate to get the images for
* @param col The column in the plate to get the images for
* @param tile The tile number. Starts with 0.
* @return <code>{ names of TIFF files, image annotation }</code>
* <p>
* Each of <code>names of TIFF files</code> and <code>image annotation</code> is a
* vector of length of the number of images.
* <p>
* <code>image annotation</code> contains
* <code>{ channel name, tile number, plate well description,
* plate augmented code, plate perm id, plate space code, plate code, row, column,
* experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code }</code>
*/
public static Object[][][] loadImages(String plate, int row, int col, int tile)
{
return loadImages(plate, row, col, tile, (String[]) null);
}
/**
* Loads the TIFF images for the given well location, list of channels, and all tiles and stores
* them in temporary files. The temporary files will be removed automatically when the Java
* Virtual Machine exits.
brinn
committed
* <p>
* Matlab example:
*
* <pre>
* % Load the images for channel DAPI of well H10 of plate P005 in space SPACE
* imginfo=OpenBISScreeningML.loadImages('/SPACE/P005', 8, 10, 'DAPI')
brinn
committed
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
* % Get the channel names and tile numbers of all locations
* imginfo(2,:,1:2)
* % Show the second image (assuming there are at least two images)
* imtool(imginfo(1,2))
* </pre>
*
* @param plate The augmented plate code
* @param row The row in the plate to get the images for
* @param col The column in the plate to get the images for
* @param channels The names of the channels to get the images for
* @return <code>{ names of TIFF files, image annotation }</code>
* <p>
* Each of <code>names of TIFF files</code> and <code>image annotation</code> is a
* vector of length of the number of images.
* <p>
* <code>image annotation</code> contains
* <code>{ channel name, tile number, plate well description,
* plate augmented code, plate perm id, plate space code, plate code, row, column,
* experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code }</code>
*/
public static Object[][][] loadImages(String plate, int row, int col, String[] channels)
{
return loadRawImages(plate, row, col, channels, createAllTilesIterator());
}
/**
* Has the same effect as {@link #loadImages(String, int, int, String[])}, but instead of
* loading raw images loads their segmentation results if available.
* @param objectNamesOrNull The names of the segmentation objects to get the images for. If
* <code>null</code> (or <code>[]</code> in MatLab) no restriction applies.
* @param analysisProcedureOrNull The analysis procedure used to filter the result. That is, the
* result is restricted to feature vector data sets with a value of property
* <code>ANALYSIS_PROCEDURE</code> as specified. If <code>null</code> (or
* <code>[]</code> in MatLab) no restriction applies.
*/
public static Object[][][] loadSegmentationImages(String plate, int row, int col,
String[] objectNamesOrNull, String analysisProcedureOrNull)
return loadSegmentationImages(plate, row, col, objectNamesOrNull, createAllTilesIterator(),
analysisProcedureOrNull);
}
private static ITileNumberIterable createAllTilesIterator()
{
return new ITileNumberIterable()
{
private int maximumNumberOfTiles;
public void setMaximumNumberOfTiles(int numberOfTiles)
{
this.maximumNumberOfTiles = numberOfTiles;
}
public int getMaximumNumberOfTiles()
{
return maximumNumberOfTiles;
}
public Iterator<Integer> iterator()
{
return new Iterator<Integer>()
{
private int index;
public boolean hasNext()
{
return index < maximumNumberOfTiles;
}
public Integer next()
{
return index++;
}
public void remove()
{
throw new UnsupportedOperationException();
}
};
}
};
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
}
/**
* Loads the TIFF images for the given well location, tile number, and list of channels and
* stores them in temporary files. The temporary files will be removed automatically when the
* Java Virtual Machine exits.
* <p>
* Matlab example:
*
* <pre>
* % Load the images for channel DAPI of well H10 and tile 3 of plate P005 in space SPACE
* imginfo=OpenBISScreeningML.loadImages('/SPACE/P005', 8, 10, 3, 'DAPI')
* % Get the channel names of all locations
* imginfo(2,:,1)
* % Show the second image (assuming there are at least two images)
* imtool(imginfo(1,2))
* </pre>
*
* @param plate The augmented plate code
* @param row The row in the plate to get the images for
* @param col The column in the plate to get the images for
* @param tile The tile number. Starts with 0.
* @param channels The names of the channels to get the images for
* @return <code>{ names of TIFF files, image annotation }</code>
* <p>
* Each of <code>names of TIFF files</code> and <code>image annotation</code> is a
* vector of length of the number of images.
* <p>
* <code>image annotation</code> contains
* <code>{ channel name, tile number, plate well description,
* plate augmented code, plate perm id, plate space code, plate code, row, column,
* experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code }</code>
*/
public static Object[][][] loadImages(String plate, int row, int col, final int tile,
String[] channels)
{
return loadRawImages(plate, row, col, channels, createSingleTileIterator(tile));
}
/**
* Has the same effect as {@link #loadImages(String, int, int, int, String[])}, but instead of
* loading raw images loads their segmentation results if available.
* @param objectNamesOrNull The names of the segmentation objects to get the images for. If
* <code>null</code> (or <code>[]</code> in MatLab) no restriction applies.
* @param analysisProcedureOrNull The analysis procedure used to filter the result. That is, the
* result is restricted to feature vector data sets with a value of property
* <code>ANALYSIS_PROCEDURE</code> as specified. If <code>null</code> (or
* <code>[]</code> in MatLab) no restriction applies.
*/
public static Object[][][] loadSegmentationImages(String plate, int row, int col,
final int tile, String[] objectNamesOrNull, String analysisProcedureOrNull)
return loadSegmentationImages(plate, row, col, objectNamesOrNull,
createSingleTileIterator(tile), analysisProcedureOrNull);
}
private static ITileNumberIterable createSingleTileIterator(final int tile)
{
return new ITileNumberIterable()
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
{
public void setMaximumNumberOfTiles(int numberOfTiles)
{
if (tile >= numberOfTiles)
{
throw new IllegalArgumentException("Tile number " + tile
+ " is not less than number of tiles " + numberOfTiles + ".");
}
}
public int getMaximumNumberOfTiles()
{
return 1;
}
public Iterator<Integer> iterator()
{
return new Iterator<Integer>()
{
private boolean delivered;
public boolean hasNext()
{
return delivered == false;
}
public Integer next()
{
delivered = true;
return tile;
}
public void remove()
{
throw new UnsupportedOperationException();
}
};
}
};
}
/**
* Lists all segmentation objects for the <var>plate</var>.
* <p>
* Matlab example:
*
* <pre>
* % Get the segmentation objects of plate P005 in space SPACE.
* segmentationObjects = OpenBISScreeningML.listSegmentationObjects('/SPACE/P005', []);
* % Get the segmentation objects of plate P005 in space SPACE for data sets calculated
* % with analysis procedure AP-42.
* segmentationObjects = OpenBISScreeningML.listSegmentationObjects('/SPACE/P005', 'AP-42');
* % How many segmentation objects do we have?
* length(segmentationObjects)
* % What is the name of segmentation objects 1?
* segmentationObjects(1)
* </pre>
*
* @param plate augmented code of the plate
* @param analysisProcedureOrNull The analysis procedure used to filter the result. That is, the
* result is restricted to feature vector data sets with a value of property
* <code>ANALYSIS_PROCEDURE</code> as specified. If <code>null</code> (or
* <code>[]</code> in MatLab) no restriction applies.
* @return Each row contains information about one segmentation object. Currently the only
* information available is the segmentation object name.
*/
public static Object[][] listSegmentationObjects(String plate, String analysisProcedureOrNull)
{
checkLoggedIn();
Plate plateId = getPlate(plate);
final List<ImageDatasetReference> imageDatasets =
listSegmentationImageDatasets(plateId, analysisProcedureOrNull);
if (imageDatasets.isEmpty())
{
return new Object[0][];
}
return extractChannels(imageDatasets.get(0));
}
private static Object[][] extractChannels(ImageDatasetReference imageDatasetReference)
{
final List<ImageDatasetMetadata> meta =
openbis.listImageMetadata(Arrays.asList(imageDatasetReference));
if (meta.isEmpty())
{
return new Object[0][];
}
final List<String> channels = getChannelCodes(meta);
Object[][] result = new Object[channels.size()][1];
for (int i = 0; i < result.length; ++i)
{
result[i][0] = channels.get(i);
}
return result;
}
private static Object[][][] loadRawImages(String plate, int row, int col, String[] channels,
ITileNumberIterable tileNumberIterable)
brinn
committed
{
checkLoggedIn();
final Plate plateId = getPlate(plate);
final List<ImageDatasetReference> imageDatasets = listRawImageDatasets(plateId);
return loadImages(plateId, imageDatasets, row, col, channels, tileNumberIterable);
}
private static Object[][][] loadSegmentationImages(String plate, int row, int col,
String[] channelsOrNull, ITileNumberIterable tileNumberIterable,
String analysisProcedureOrNull)
{
checkLoggedIn();
final Plate plateId = getPlate(plate);
final List<ImageDatasetReference> imageDatasets =
listSegmentationImageDatasets(plateId, analysisProcedureOrNull);
return loadImages(plateId, imageDatasets, row, col, channelsOrNull, tileNumberIterable);
}
private static Object[][][] loadImages(Plate plate, List<ImageDatasetReference> imageDatasets,
int row, int col, String[] channelsOrNull, ITileNumberIterable tileNumberIterable)
final List<ImageDatasetMetadata> meta = openbis.listImageMetadata(imageDatasets);
if (meta.isEmpty())
{
return new Object[][][]
{ new Object[0][], new Object[0][] };
}
brinn
committed
final List<String> imageChannels;
if (channelsOrNull == null || channelsOrNull.length == 0)
brinn
committed
{
brinn
committed
} else
{
imageChannels = Arrays.asList(channelsOrNull);
brinn
committed
}
final List<ImageReferenceAndFile> imageReferencesAndFiles =
new ArrayList<ImageReferenceAndFile>(imageDatasets.size());
brinn
committed
final Object[][][] result = new Object[2][][];
tileNumberIterable.setMaximumNumberOfTiles(meta.get(0).getNumberOfTiles());
int numberOfTiles = tileNumberIterable.getMaximumNumberOfTiles();
result[0] = new Object[imageDatasets.size() * imageChannels.size() * numberOfTiles][1];
result[1] = new Object[imageDatasets.size() * imageChannels.size() * numberOfTiles][15];
brinn
committed
int dsIdx = 0;
int resultIdx = 0;
for (ImageDatasetReference ds : imageDatasets)
{
for (String channel : imageChannels)
{
for (Integer tile : tileNumberIterable)
brinn
committed
{
final PlateImageReference ref =
new PlateImageReference(row, col, tile, channel, ds);
final File imageFile = createImageFileName(plate, ref);
imageReferencesAndFiles.add(new ImageReferenceAndFile(ref, imageFile));
brinn
committed
result[0][resultIdx][0] = imageFile.getPath();
PlateIdentifier plateIdentifier = ds.getPlate();
ExperimentIdentifier expIdentifier = ds.getExperimentIdentifier();
brinn
committed
final Object[] annotations =
new Object[]
{ channel, tile,
createPlateWellDescription(plateIdentifier, row, col),
plateIdentifier.getAugmentedCode(),
plateIdentifier.getPermId(),
plateIdentifier.tryGetSpaceCode(),
plateIdentifier.getPlateCode(), row, col,
expIdentifier.getAugmentedCode(),
expIdentifier.getPermId(), expIdentifier.getSpaceCode(),
expIdentifier.getProjectCode(),
expIdentifier.getExperimentCode(), ds.getPermId(), };
brinn
committed
System.arraycopy(annotations, 0, result[1][resultIdx], 0, annotations.length);
resultIdx++;
}
}
dsIdx++;
}
try
{
loadImages(imageReferencesAndFiles);
brinn
committed
} catch (IOException ex)
{
throw new RuntimeException(ex);
}
return result;
}
private static List<ImageDatasetReference> listRawImageDatasets(final Plate plateId)
{
return openbis.listRawImageDatasets(Arrays.asList(plateId));
}
private static List<ImageDatasetReference> listSegmentationImageDatasets(final Plate plateId,
String analysisProcedureOrNull)
return openbis.listSegmentationImageDatasets(Arrays.asList(plateId),
analysisProcedureOrNull);
}
brinn
committed
/**
* Saves images for a given list of image references (given by data set code, well position,
* channel and tile) and files.
brinn
committed
*
* @throws IOException when reading images from the server or writing them to the files fails
*/
private static void loadImages(List<ImageReferenceAndFile> imageReferencesAndFiles)
throws IOException
brinn
committed
{
List<PlateImageReference> imageReferences = new ArrayList<PlateImageReference>();
final Map<PlateImageReference, ImageReferenceAndFile> imageRefToFileMap =
new HashMap<PlateImageReference, ImageReferenceAndFile>();
for (ImageReferenceAndFile imageReferenceAndFile : imageReferencesAndFiles)
{
PlateImageReference imageReference = imageReferenceAndFile.getImageReference();
File file = loadedImages.get(imageReference);
if (file == null)
{
imageReferences.add(imageReference);
imageRefToFileMap.put(imageReference, imageReferenceAndFile);
}
}
brinn
committed
try
{
openbis.loadImages(imageReferences, new IImageOutputStreamProvider()
{
public OutputStream getOutputStream(PlateImageReference imageReference)
throws IOException
{
return imageRefToFileMap.get(imageReference).open();
brinn
committed
}
}, false);
brinn
committed
} finally
{
Collection<ImageReferenceAndFile> values = imageRefToFileMap.values();
for (ImageReferenceAndFile imageReferenceAndFile : values)
{
imageReferenceAndFile.close();
PlateImageReference imageReference = imageReferenceAndFile.getImageReference();
loadedImages.put(imageReference, imageReferenceAndFile.getImageFile());
}
brinn
committed
}
}
brinn
committed
private static File createImageFileName(Plate plate, PlateImageReference image)
{
final WellPosition well = image.getWellPosition();
File imageDir = new File(temporarySessionDir, "images");
imageDir.mkdirs();
final File f =
new File(imageDir, "img_" + plate.getPlateCode() + "_" + image.getDatasetCode()
+ "_row" + well.getWellRow() + "_col" + well.getWellColumn() + "_"
+ image.getChannel() + "_tile" + image.getTile() + ".tiff");
f.deleteOnExit();
return f;
brinn
committed
}
brinn
committed
//
// Feature matrix
//
/**
* Returns the feature matrix of the specified features for all locations in
* <var>experiment</var> (a location is one well position in one feature vector data set) in
* <var>experiment</var> connected to <var>gene</var> in <code>[0]</code>, location annotations
* in <code>[1]</code> and feature annotation in <code>[2]</code>.
* <p>
* Matlab example:
*
* <pre>
* % Get feature matrix for experiment /SPACE/PROJ/MYEXP for locations connected to GENENAME
* fmatrix = OpenBISScreeningML.getFeatureMatrix('/SPACE/PROJ/MYEXP', 'GENENAME', [], []);
* % Get feature matrix for features F1, F2 and F3 for
brinn
committed
* % experiment /SPACE/PROJ/MYEXP for locations connected to GENENAME
* fmatrix = OpenBISScreeningML.getFeatureMatrix('/SPACE/PROJ/MYEXP', 'GENENAME', [], {'F1' 'F2' 'F3'));
* % Get feature matrix for features F1 and F2 for experiment /SPACE/PROJ/MYEXP for locations
* % connected to GENENAME calculated with analysis procedure AP-42.
* fmatrix = OpenBISScreeningML.getFeatureMatrix('/SPACE/PROJ/MYEXP', 'GENENAME', 'AP-42', {'F1' 'F2'));
* % Get the feature vector for the second location (assuming that there are at least two locations)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(1,:,2,3)
* % Get the values of the fourth feature for all locations (assuming that there are at least 4 features)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(1,4,:,3)
* % Get code of the fourth feature (assuming that there are at least 4 features)
* fmatrix(3,4)
* % Get the plate-well descriptions for the second location (assuming that there are at least two locations)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(2,2,3,:)
brinn
committed
* </pre>
*
* @param experiment The augmented experiment code
* @param gene The gene code (stored as material code in openBIS, usually it is gene id)
* @param analysisProcedureOrNull The code of the analysis procedure used to calculate requested
* features. That is, the result is restricted to feature vector data sets with a
* value of property <code>ANALYSIS_PROCEDURE</code> as specified. If
* <code>null</code> (or <code>[]</code> in MatLab) no restriction applies.
* @param featuresOrNull The codes of the features to contain the feature matrix. Unknown
* feature codes will be ignored. If <code>null</code> (or <code>[]</code> in MatLab)
* all features are returned.
* @return a four dimensional matrix. The first dimension denotes the type in the following
* order: <code>{feature matrix, annotations per location, feature codes}</code>. The
* other dimensions depend on the value of the first dimension:
* <ol>
* <li>feature matrix: 2. dimension is feature vector, 3. dimension is location number,
* 4. dimension is data set number. If for a particular location and a particular data
* set the corresponding feature value does not exists <code>NaN</code> will be
* returned. <li>annotations: 2. dimension is location number, 3. dimension is data set
* number, 4. dimension is location annotations in the following order: <code>{plate
* well description, plate augmented code, plate perm id, plate space code, plate code,
* row, column, experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code, data set type}</code> <li>
* feature codes: 2. dimension is feature codes in alphabetical order. 3. and 4.
* dimension are meaningless (i.e. they have length one)
* </ol>
brinn
committed
*/
public static Object[][][][] getFeatureMatrix(String experiment, String gene,
String analysisProcedureOrNull, String[] featuresOrNull)
brinn
committed
{
checkLoggedIn();
final ExperimentIdentifier experimentId = experimentCodeToExperimentMap.get(experiment);
if (experimentId == null)
{
throw new RuntimeException("No experiment with that code found.");
}
final List<Plate> experimentPlates = experimentToPlateMap.get(experiment);
if (experimentPlates == null || experimentPlates.isEmpty())
{
return new Object[][][][]
{ new Object[0][][], new Object[0][][], new Object[0][][] };
brinn
committed
}
final List<FeatureVectorWithDescription> featureVectors =
openbis.loadFeaturesForPlateWells(experimentId, new MaterialIdentifier(
MaterialTypeIdentifier.GENE, gene), analysisProcedureOrNull,
(featuresOrNull == null) ? null : Arrays.asList(featuresOrNull));
return getFeatureMatrix(featureVectors);
brinn
committed
}
/**
* Returns the feature matrix of the specified features for all locations (a location is one
* well position in one feature vector data set) in <var>experiment</var> connected to
* <var>gene</var> in <code>[0]</code>, location annotations in <code>[1]</code> and feature
* annotation in <code>[2]</code>.
* <p>
* Matlab example:
*
* <pre>
* % Get feature matrix for GENENAME
* fmatrix = OpenBISScreeningML.getFeatureMatrix('GENENAME', [], []);
* % Get feature matrix for features FEATURE1, FEATURE2 and FEATURE3 for GENENAME
* fmatrix = OpenBISScreeningML.getFeatureMatrix('GENENAME', [], {'FEATURE1' 'FEATURE2' 'FEATURE3'});
* % Get feature matrix for features FEATURE1 and FEATURE2 for GENENAME
* % computed with analysis procedure AP-42
* fmatrix = OpenBISScreeningML.getFeatureMatrix('GENENAME', 'AP-42', {'FEATURE1' 'FEATURE2'});
* % Get the feature vector for the second location (assuming that there are at least two locations)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(1,:,2,3)
* % Get the values of the fourth feature for all locations (assuming that there are at least 4 features)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(1,4,:,3)
* % Get code of the fourth feature (assuming that there are at least 4 features)
* fmatrix(3,4)
* % Get the plate-well descriptions for the second location (assuming that there are at least two locations)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(2,2,3,:)
brinn
committed
* </pre>
*
* @param gene The gene code (stored as material code in openBIS, usually it is gene id)
* @param analysisProcedureOrNull The code of the analysis procedure used to calculate requested
* features. That is, the result is restricted to feature vector data sets with a
* value of property <code>ANALYSIS_PROCEDURE</code> as specified. If
* <code>null</code> (or <code>[]</code> in MatLab) no restriction applies.
* @param featuresOrNull The codes of the features to contain the feature matrix. Unknown
* feature codes will be ignored. If <code>null</code> (or <code>[]</code> in MatLab)
* all features are returned.
* @return a four dimensional matrix. The first dimension denotes the type in the following
* order: <code>{feature matrix, annotations per location, feature codes}</code>. The
* other dimensions depend on the value of the first dimension:
* <ol>
* <li>feature matrix: 2. dimension is feature vector, 3. dimension is location number,
* 4. dimension is data set number. If for a particular location and a particular data
* set the corresponding feature value does not exists <code>NaN</code> will be
* returned. <li>annotations: 2. dimension is location number, 3. dimension is data set
* number, 4. dimension is location annotations in the following order: <code>{plate
* well description, plate augmented code, plate perm id, plate space code, plate code,
* row, column, experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code, data set type}</code> <li>
* feature codes: 2. dimension is feature codes in alphabetical order. 3. and 4.
* dimension are meaningless (i.e. they have length one)
* </ol>
brinn
committed
*/
public static Object[][][][] getFeatureMatrix(String gene, String analysisProcedureOrNull,
String[] featuresOrNull)
brinn
committed
{
checkLoggedIn();
final List<FeatureVectorWithDescription> featureVectors =
openbis.loadFeaturesForPlateWells(new MaterialIdentifier(
MaterialTypeIdentifier.GENE, gene), analysisProcedureOrNull,
(featuresOrNull == null) ? null : Arrays.asList(featuresOrNull));
return getFeatureMatrix(featureVectors);
}
private static Object[][][][] getFeatureMatrix(
final List<FeatureVectorWithDescription> featureVectors)
{
final Object[][][][] result = new Object[3][][][];
brinn
committed
if (featureVectors.isEmpty())
{
return result;
}
List<String> featureCodes = getFeatureCodes(featureVectors);
Map<String, Integer> featureCodeToIndexMap = new HashMap<String, Integer>();
result[2] = new Object[featureCodes.size()][1][1];
for (int i = 0; i < featureCodes.size(); ++i)
brinn
committed
{
String code = featureCodes.get(i);
result[2][i][0][0] = code;
featureCodeToIndexMap.put(code, i);
brinn
committed
}
IndexSet<String> dataSetCodes = new IndexSet<String>();
IndexSet<WellPosition> wellPositions = new IndexSet<WellPosition>();
for (FeatureVectorWithDescription featureVector : featureVectors)
brinn
committed
{
WellPosition wellPosition = featureVector.getWellPosition();
wellPositions.add(wellPosition);
dataSetCodes.add(featureVector.getDatasetWellReference().getDatasetCode());
brinn
committed
}
result[0] = new Object[featureCodes.size()][wellPositions.size()][dataSetCodes.size()];
result[1] = new Object[wellPositions.size()][dataSetCodes.size()][14];
for (FeatureVectorWithDescription vector : featureVectors)
{
WellPosition wellPosition = vector.getWellPosition();
int wellIndex = wellPositions.getIndex(wellPosition);
int[] featureIndexes = createFeatureIndexes(vector, featureCodeToIndexMap);
FeatureVectorDatasetWellReference dataSetRef = vector.getDatasetWellReference();
int dataSetIndex = dataSetCodes.getIndex(dataSetRef.getDatasetCode());
copyFeatureValuesInto(result, wellIndex, dataSetIndex, featureIndexes, vector);
copyAnnotationsInto(result, wellIndex, dataSetIndex, dataSetRef, vector);
}
return replaceFeatureNullValuesByNaN(result);
brinn
committed
}
/**
* Returns the feature matrix of the specified features for all locations (a location is one
* well position in one feature vector data set) of all feature vector data sets created by
* specified analysis procedure of the given <var>plate</var> in <code>[0]</code>, location
* annotations in <code>[1]</code> and feature annotation in <code>[2]</code>.
* % Get feature matrix for PLATECODE
* fmatrix = OpenBISScreeningML.getFeatureMatrixForPlate('PLATECODE', [], []);
* % Get feature matrix for features FEATURE1, FEATURE2 and FEATURE3 for PLATECODE.
* fmatrix = OpenBISScreeningML.getFeatureMatrixForPlate('PLATECODE', [], {'FEATURE1' 'FEATURE2' 'FEATURE3'});
* % Get feature matrix for features FEATURE1 and FEATURE2 for PLATECODE calculated by analysis procedure AP-42.
* fmatrix = OpenBISScreeningML.getFeatureMatrixForPlate('PLATECODE', 'AP-42', {'FEATURE1' 'FEATURE2'});
* % Get the feature vector for the second location (assuming that there are at least two locations)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(1,:,2,3)
* % Get the values of the fourth feature for all locations (assuming that there are at least 4 features)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(1,4,:,3)
* % Get code of the fourth feature (assuming that there are at least 4 features)
* fmatrix(3,4)
* % Get the plate-well descriptions for the second location (assuming that there are at least two locations)
* % of third data set (assuming that there are at least three data sets)
* fmatrix(2,2,3,:)
* @param plate augmented code of the plate for which features should be loaded
* @param analysisProcedureOrNull The code of the analysis procedure used to calculate requested
* features. That is, the result is restricted to feature vector data sets with a
* value of property <code>ANALYSIS_PROCEDURE</code> as specified. If
* <code>null</code> (or <code>[]</code> in MatLab) no restriction applies.
* @param featuresOrNull The codes of the features to contain the feature matrix. Unknown
* feature codes will be ignored. If <code>null</code> (or <code>[]</code> in MatLab)
* all features are returned.
* @return a four dimensional matrix. The first dimension denotes the type in the following
* order: <code>{feature matrix, annotations per location, feature codes}</code>. The
* other dimensions depend on the value of the first dimension:
* <ol>
* <li>feature matrix: 2. dimension is feature vector, 3. dimension is location number,
* 4. dimension is data set number. If for a particular location and a particular data
* set the corresponding feature value does not exists <code>NaN</code> will be
* returned. <li>annotations: 2. dimension is location number, 3. dimension is data set
* number, 4. dimension is location annotations in the following order: <code>{plate
* well description, plate augmented code, plate perm id, plate space code, plate code,
* row, column, experiment augmented code, experiment perm id, experiment space code,
* experiment project code, experiment code, data set code, data set type}</code> <li>
* feature codes: 2. dimension is feature codes in alphabetical order. 3. and 4.
* dimension are meaningless (i.e. they have length one)
* </ol>
public static Object[][][][] getFeatureMatrixForPlate(String plate,
String analysisProcedureOrNull, String[] featuresOrNull)
final List<FeatureVectorDataset> dataSets =
openbis.loadFeaturesForPlates(
Arrays.asList(PlateIdentifier.createFromAugmentedCode(plate)),
(featuresOrNull == null) ? null : Arrays.asList(featuresOrNull),
analysisProcedureOrNull);
final Object[][][][] result = new Object[3][][][];
if (dataSets.isEmpty())
List<String> featureCodes = getFeatureCodes(dataSets);
Map<String, Integer> featureCodeToIndexMap = new HashMap<String, Integer>();
result[2] = new Object[featureCodes.size()][1][1];
for (int i = 0; i < featureCodes.size(); ++i)
String code = featureCodes.get(i);
result[2][i][0][0] = code;
featureCodeToIndexMap.put(code, i);
IndexSet<WellPosition> wellPositions = new IndexSet<WellPosition>();
for (FeatureVectorDataset dataSet : dataSets)
List<FeatureVector> featureVectors = dataSet.getFeatureVectors();
for (FeatureVector featureVector : featureVectors)
WellPosition wellPosition = featureVector.getWellPosition();
wellPositions.add(wellPosition);
int numberOfDataSets = dataSets.size();
result[0] = new Object[featureCodes.size()][wellPositions.size()][numberOfDataSets];
result[1] = new Object[wellPositions.size()][numberOfDataSets][14];
for (int dataSetIndex = 0; dataSetIndex < numberOfDataSets; dataSetIndex++)
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
FeatureVectorDataset dataSet = dataSets.get(dataSetIndex);
FeatureVectorDatasetReference dataSetRef = dataSet.getDataset();
int[] featureIndexes = createFeatureIndexes(dataSet, featureCodeToIndexMap);
for (FeatureVector vector : dataSet.getFeatureVectors())
{
WellPosition wellPosition = vector.getWellPosition();
int wellIndex = wellPositions.getIndex(wellPosition);
copyFeatureValuesInto(result, wellIndex, dataSetIndex, featureIndexes, vector);
copyAnnotationsInto(result, wellIndex, dataSetIndex, dataSetRef, vector);
}
}
return replaceFeatureNullValuesByNaN(result);
}
private static Object[][][][] replaceFeatureNullValuesByNaN(final Object[][][][] result)
{
Double nan = Double.NaN;
for (int i = 0; i < result[0].length; i++)
{
Object[][] r0i = result[0][i];
for (int j = 0; j < r0i.length; j++)
{
Object[] r0ij = r0i[j];
for (int k = 0; k < r0ij.length; k++)
{
if (r0ij[k] == null)
{
r0ij[k] = nan;
}
}
}
private static int[] createFeatureIndexes(IFeatureCodesProvider dataSet,
Map<String, Integer> featureCodeToIndexMap)
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
List<String> codes = dataSet.getFeatureCodes();
int[] featureIndexes = new int[codes.size()];
for (int i = 0; i < featureIndexes.length; i++)
{
featureIndexes[i] = featureCodeToIndexMap.get(codes.get(i));
}
return featureIndexes;
}
private static void copyFeatureValuesInto(final Object[][][][] result, int wellIndex,
int dataSetIndex, int[] featureIndexes, FeatureVector vector)
{
List<Object> valueObjects = vector.getValueObjects();
for (int i = 0, n = valueObjects.size(); i < n; i++)
{
int featureIndex = featureIndexes[i];
try
{
Object[][] f = result[0][featureIndex];
Object[] d = f[wellIndex];
d[dataSetIndex] = valueObjects.get(i);
} catch (Exception ex)
{
throw new RuntimeException(featureIndex + "." + wellIndex + "." + dataSetIndex, ex);
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
}
}
}
private static void copyAnnotationsInto(final Object[][][][] result, int wellIndex,
int dataSetIndex, FeatureVectorDatasetReference dataSetRef, FeatureVector vector)
{
Object[] annotations =
new Object[]
{ createPlateWellDescription(dataSetRef.getPlate(), vector),
dataSetRef.getPlate().getAugmentedCode(),
dataSetRef.getPlate().getPermId(),
dataSetRef.getPlate().tryGetSpaceCode(),
dataSetRef.getPlate().getPlateCode(),
vector.getWellPosition().getWellRow(),
vector.getWellPosition().getWellColumn(),
dataSetRef.getExperimentIdentifier().getAugmentedCode(),
dataSetRef.getExperimentIdentifier().getPermId(),
dataSetRef.getExperimentIdentifier().getSpaceCode(),
dataSetRef.getExperimentIdentifier().getProjectCode(),
dataSetRef.getExperimentIdentifier().getExperimentCode(),
dataSetRef.getDatasetCode(), dataSetRef.getDataSetType() };
System.arraycopy(annotations, 0, result[1][wellIndex][dataSetIndex], 0, annotations.length);
}
private static List<String> getFeatureCodes(List<? extends IFeatureCodesProvider> dataSets)
{
Set<String> codes = new HashSet<String>();
for (IFeatureCodesProvider featureVectorDataset : dataSets)
{
codes.addAll(featureVectorDataset.getFeatureCodes());
}
List<String> result = new ArrayList<String>(codes);
Collections.sort(result);
return result;
/**
* Returns the gene mapping for the given <var>plateCodes</var> in <code>[0]</code> and location
* annotations in <code>[1]</code>.
* <p>
* One row in the matrix corresponds to one well.
* <p>
* Matlab example:
*
* <pre>
* % Get feature matrix for features FEATURE1, FEATURE2 and FEATURE for PLATECODE
* genes = getGeneMappingForPlate('PLATECODE');
* % Get the plate well location description of the 10th wells
* loc2 = genes(2,10,1)
* % Get the gene ids that are in the 10th well
* geneIds = genes(1,10,:)
* </pre>
*
* @param platesCodes The augmented codes of the plates to get the mapping for
* @return <code>{ gene ids, annotations per well }</code> where <code>gene ids</code> can be 0,
* 1 or more gene ids. <code>annotations per location</code> contain: