java.lang.Object
- com.tibco.patterns.learn.api.project.ProjectFiles

```
public final class ProjectFiles
extends java.lang.Object
```
Manages all files for one Patterns Learn project. This class: 1. Knows and maintains a specific folder and file naming structure that the Learn application uses to store projects. 2. Contains factory methods that construct ProjectSettings and ModelSettings objects while creating the necessary files and folders. 3. Provides methods that modify both project and model XML files. 4. Ensures consistency between the name of project/model stored in XML file and the project/model directory name.

Method Summary

All Methods Static Methods Instance Methods Concrete Methods
Modifier and Type	Method	Description
`DataPartition`	`addRecPair(RecPair recPair)`	Adds a new record pair randomly to the training or validation dataset of the active model.
`void`	`addRecPair(RecPair recPair, DataPartition partition)`	Adds a new record pair to the specified dataset of the active model.
`void`	`addRecPairsFromCsv(java.io.File csvFile)`	Adds all record pairs from the specified CSV file to the active model settings.
`void`	`addThesFile(java.lang.String filePath)`	Adds the thesaurus file to the "thesauri subdirectory of the active model's directory, overwriting any previous file with same name.
`AddPairsResult`	`addUsefulRecPairs(java.util.List<RecPair> recPairs, int maxTrainSize, int maxReserveSize)`	Adds each record pair if it is likely to improve model training while limiting the grwth of datasets.
`static ProjectFiles`	`create(java.lang.String projectName, java.lang.String projectPath)`	Creates a new project.
`LowConfPairFinder`	`createLowConfPairFinder()`	Creates and stores an object that finds low confidence pairs in the data table.
`ModelTrainConfig`	`createModelTrainConfig(ModelConfig mConfig, java.lang.Boolean randomOrder, java.util.Map<Partition,ConvergenceObserver> observers)`	Creates and remembers a new ModelTrainConfig, replacing any previous one.
`RecPairScorer`	`createScorer(com.netrics.likeit.NetricsServerInterface server)`	Creates a RecPairScorer that can be used only for scoring any record pairs with known field values.
`ModelTrainer`	`createTrainer(com.netrics.likeit.NetricsServerInterface server)`	Creates a ModelTrainer for training the active model.
`void`	`deleteModel(int index)`	Deletes the non-active model specified by index.
`void`	`deleteTrainedModel()`	Deletes binary file of the trained model.
`void`	`exportModelAndQuery(int modelIndex, java.nio.file.Path exportedModelFile, java.lang.String modelVersion, java.nio.file.Path queryOutputDir, java.lang.String javaClassName, java.lang.String methodName)`	Exports all files needed to use the trained model on the Patterns server: the trained model binary file, Java source file that builds the Learn query, and any thesaurus files used by the query.
`void`	`exportThesFiles(java.nio.file.Path outputDir)`	Exports all thesaurus files used by the feature query of the active model.
`void`	`exportTrainedModel(java.nio.file.Path exportedFile)`	Exports the trained model binary file with model score threshold.
`void`	`exportTrainedModel(java.nio.file.Path exportedFile, java.lang.String version)`	Exports the trained model binary file from the active ModelSettings to the given location.
`ModelSettings`	`getActiveModelSettings()`
`java.nio.file.Path`	`getDataFile(int fileIndex)`
`ModelTrainConfig`	`getModelTrainConfig()`	Get model training configuration.
`int`	`getNModels()`
`ProjectSettings`	`getProjectSettings()`
`boolean`	`hasLowConfPairFinder()`
`boolean`	`hasModelTrainConfig()`	Checks whether a custom model training configuration is used to train the model.
`static ProjectFiles`	`load(java.lang.String projectName, java.lang.String projectPath)`	Loads an existing project from the specified directory.
`ModelSettings`	`loadModel(int index)`	Loads ModelSettings from model.xml file specified by the index.
`void`	`loadTrainedModel()`	Loads the trained model from the model.bin file to ModelTrainer of the active model.
`boolean`	`localDataFileExists(java.lang.String filePath)`
`AddPairsResult`	`moveUsefulRecPairs(DataPartition sourcePartition, int maxTrainSize)`	Moves the record pairs that are likely to improve model training from the given dataset (Reserve) to the training or validation dataset of the active model.
`void`	`replaceDataFile(int fileIndex, java.lang.String filePath, java.util.List<java.lang.String> fieldNames, boolean copyFile, boolean overwriteFile)`	Replaces the currently assigned data file with a compatible data file.
`void`	`save()`	Saves project.xml and active model.xml files to the existing project location
`void`	`saveAs(java.lang.String projectName, java.lang.String projectPath)`	Saves project files to the new project location and/or project name.
`void`	`saveLowConfPairFinder(int maxNPairs)`	Saves the state of the low confidence pair finder that was created before and used to find pairs.
`void`	`saveModelCopy(java.lang.String copiedModelName)`	Saves the active model, then saves its copy (creates model dir, model.xml and model.bin).
`void`	`saveTrainedModel()`	Saves trained model binary file to the active model directory.
`int`	`setDataFile(java.lang.String filePath, java.util.List<java.lang.String> fieldNames, boolean copyFile, boolean overwriteFile)`	Sets the data file info.
`void`	`setDebugOutput(boolean value)`	Sets the debug output.
`void`	`setKeyField(int fieldIndex)`	Sets the index of the key field in the data file.
`ModelSettings`	`switchActiveModel(int index)`	Saves the currently active model, then loads another model from file, makes it active, saves the new active model index.
`boolean`	`thesFileExists(java.lang.String filePath)`
`java.lang.String`	`toString()`
`void`	`updateDataFileLastModified(int fileIndex)`	Updates the stored last modified date of the data file.
`void`	`verifyCanSetDataFile()`	Verifies that a data file can be assigned or replaced.
`static void`	`verifyDir(java.lang.String projectName, java.lang.String projectPath)`	Verifies that the structure of files and directories of the given project is correct without loading the entire project.

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

- Method Detail
  - create
```
public static ProjectFiles create(java.lang.String projectName,
                                  java.lang.String projectPath)
                           throws java.io.IOException,
                                  org.xml.sax.SAXException,
                                  javax.xml.bind.JAXBException
```
    Creates a new project. Creates project directory structure, saves the project settings and current model settings. Uses random partitioning for pairs.
    
    Parameters:
    
    projectName - - the name of the new project directory. Must not exist. Not null.
    
    projectPath - - path where the project directory will be created. The path must exist. Not null.
    
    Returns:
    
    newly constructed ProjectFiles object.
    
    Throws:
    
    org.xml.sax.SAXException - if the schema file cannot be parsed.
    
    javax.xml.bind.JAXBException - when XML validation fails.
    
    java.nio.file.FileAlreadyExistsException - if project directory exists (optional).
    
    java.io.IOException - if project directory cannot be created.
  - load
```
public static ProjectFiles load(java.lang.String projectName,
                                java.lang.String projectPath)
                         throws org.xml.sax.SAXException,
                                javax.xml.bind.JAXBException,
                                java.io.FileNotFoundException
```
    Loads an existing project from the specified directory. Uses random partitioning for pairs.
    
    Parameters:
    
    projectName - - the name of the project directory (must exist). Not null.
    
    projectPath - - path to the project directory. Not null.
    
    Returns:
    
    newly constructed ProjectFiles object with data loaded from files.
    
    Throws:
    
    javax.xml.bind.JAXBException - if an XML file validation fails, or if an XML file was created with newer Learn API.
    
    org.xml.sax.SAXException - if the schema file cannot be parsed.
    
    java.io.FileNotFoundException - if a required project file or directory is not found.
    
    java.lang.IndexOutOfBoundsException - if active model index is invalid.
  - verifyDir
```
public static void verifyDir(java.lang.String projectName,
                             java.lang.String projectPath)
                      throws java.io.FileNotFoundException
```
    Verifies that the structure of files and directories of the given project is correct without loading the entire project. This only verifies the presence of files and directories that are required by any Learn project. It does not verify the content of each project file (loading the project is required to do that).
    
    Parameters:
    
    projectName - - the name of the project directory. Not null.
    
    projectPath - - path to the project directory. Not null.
    
    Throws:
    
    java.io.FileNotFoundException - if a required project directory or file does not exist.
  - toString
```
public java.lang.String toString()
```
    Overrides:
    
    toString in class java.lang.Object
  - getProjectSettings
```
public ProjectSettings getProjectSettings()
```
    Returns:
    
    internal reference to the ProjectSettings object.
  - setDebugOutput
```
public void setDebugOutput(boolean value)
```
    Sets the debug output. Used for LowConfPairFinder.
    
    Parameters:
    
    value - - if true, debug information is outputted to console.
  - save
```
public void save()
          throws javax.xml.bind.JAXBException
```
    Saves project.xml and active model.xml files to the existing project location
    
    Throws:
    
    javax.xml.bind.JAXBException - when XML validation fails.
  - saveAs
```
public void saveAs(java.lang.String projectName,
                   java.lang.String projectPath)
            throws javax.xml.bind.JAXBException,
                   java.io.IOException
```
    Saves project files to the new project location and/or project name. New project directory will be created (must not exist). This ProjectFiles object refers to the new location after the call. The files in the old project location are not modified. Note: any data file is NOT copied (currently client code manages data file location).
    
    Parameters:
    
    projectName - - name of the project directory. Not null.
    
    projectPath - - path to the project directory. Not null.
    
    Throws:
    
    javax.xml.bind.JAXBException - when XML validation fails.
    
    java.nio.file.FileAlreadyExistsException - if project directory exists (optional).
    
    java.io.IOException - if directory cannot be created or other I/O error occurs.
  - getActiveModelSettings
```
public ModelSettings getActiveModelSettings()
```
    Returns:
    
    internal reference to the ModelSettings object for the active model that was selected by the user. By default the current (most recent) model is active.
  - getNModels
```
public int getNModels()
```
    Returns:
    
    the number of saved models (model.xml files, ModelSettings objects).
  - loadModel
```
public ModelSettings loadModel(int index)
                        throws javax.xml.bind.JAXBException,
                               java.io.FileNotFoundException
```
    Loads ModelSettings from model.xml file specified by the index. Can be used to read info for any model. Any changes to the returned object cannot be saved (the returned object is not retained by ProjectFiles, and ModelSettings does not have any public save methods. Use getActiveModelSettings() to make and save changes).
    
    Parameters:
    
    index - - index of the model in the list of models (in ProjectSettings).
    
    Returns:
    
    newly created ModelSettings object.
    
    Throws:
    
    javax.xml.bind.JAXBException - if model XML file validation fails or if the XML file was created with newer Learn API.
    
    java.io.FileNotFoundException - if the model XML file is not found.
    
    java.lang.IndexOutOfBoundsException - if model index is invalid.
  - switchActiveModel
```
public ModelSettings switchActiveModel(int index)
                                throws javax.xml.bind.JAXBException,
                                       java.io.FileNotFoundException
```
    Saves the currently active model, then loads another model from file, makes it active, saves the new active model index. Any previously obtained references to the active ModelSettings object and its derived objects (FeatureQuery, ModelTrainer) become obsolete and should no longer be used.
    
    Parameters:
    
    index - - index of the model in the list of models (in ProjectSettings).
    
    Returns:
    
    ModelSettings object for the new active model.
    
    Throws:
    
    javax.xml.bind.JAXBException - if an XML file validation fails, or if an XML file was created with newer Learn API.
    
    java.io.FileNotFoundException - if the model XML file is not found.
  - saveModelCopy
```
public void saveModelCopy(java.lang.String copiedModelName)
                   throws javax.xml.bind.JAXBException,
                          java.io.IOException
```
    Saves the active model, then saves its copy (creates model dir, model.xml and model.bin). Adds the copied model to the list of models in ProjectSettings. After the call, the active model is still the same model.xml, not the new copy.
    
    Parameters:
    
    copiedModelName - - the name of the new copy of the active model. Must be different from the name of any other saved model.
    
    Throws:
    
    javax.xml.bind.JAXBException - when XML validation fails.
    
    java.nio.file.FileAlreadyExistsException - if the new model directory exists (optional).
    
    java.io.IOException - if the new model directory cannot be created.
    
    java.nio.file.InvalidPathException - if copiedModelName is not a valid directory name (optional).
    
    java.lang.IllegalArgumentException - if copiedModelName is null, empty, or 'current'
  - deleteModel
```
public void deleteModel(int index)
                 throws java.io.IOException,
                        javax.xml.bind.JAXBException
```
    Deletes the non-active model specified by index. Deletes model files and directory, saves the updated project. The index of the active model may change. Note that deleting a model may break the continuity of the tree structure of saved models.
    
    Parameters:
    
    index - - index of the model in the list of models (in ProjectSettings).
    
    Throws:
    
    java.lang.IllegalArgumentException - if the model to be deleted is currently active.
    
    javax.xml.bind.JAXBException - when XML validation fails when saving project. Model files were not deleted.
    
    java.io.IOException - if one of the model files or directory could not be deleted. In this case the model list no longer contains the deleted model, but model files may still exist.
  - localDataFileExists
```
public boolean localDataFileExists(java.lang.String filePath)
```
    Returns:
    
    true if the file with the specified name exists in project's "tables" directory.
  - setDataFile
```
public int setDataFile(java.lang.String filePath,
                       java.util.List<java.lang.String> fieldNames,
                       boolean copyFile,
                       boolean overwriteFile)
                throws java.io.IOException
```
    Sets the data file info. Assumes only 1 data file can be used. Fills in the fieldDefs in both project and active model XML files based on given fieldNames taken from the data file. Assigns field type Searchable Text to all fields. Removes key field index, feature query, pair finder settings. Must be followed by setKeyField(int) that creates feature query.
    
    Parameters:
    
    filePath - - the data file including the absolute path.
    
    fieldNames - defines all field names and field order. Indexes of these field names are later used to identify a specific field.
    
    copyFile - - if true, copies the data file to the project directory.
    
    overwriteFile - - overwrite any existing local data file while copying.
    
    Returns:
    
    index of the added file (currently always 0).
    
    Throws:
    
    java.io.FileNotFoundException - if the specified data file is not found (optional).
    
    java.nio.file.FileAlreadyExistsException - if local file exists during copying and overwriteFile==false (optional).
    
    java.io.IOException - if file I/O error occurs during copying or other operation.
    
    java.lang.IllegalArgumentException - - if fieldNames is not given or has empty or duplicate names
    
    java.lang.IllegalStateException - - if existing data file cannot be replaced: features, record pairs or trained model already assigned, or several models have been saved for the same project.
  - replaceDataFile
```
public void replaceDataFile(int fileIndex,
                            java.lang.String filePath,
                            java.util.List<java.lang.String> fieldNames,
                            boolean copyFile,
                            boolean overwriteFile)
                     throws java.io.IOException
```
    Replaces the currently assigned data file with a compatible data file. The new data file must have the same field names, types, and same key field. Replacement can be performed even when features, pairs or trained models exist. Does not change already assigned fields (types, statistics, key field, ignored fields). Removes pair finder settings. Field statistics should be reassigned after replacing the data file.
    
    Parameters:
    
    fileIndex - - index of the data file: from 0 to ProjectSettings.getNDataFiles()-1
    
    filePath - - the data file including the absolute path.
    
    fieldNames - - field names read from the new data file. Must match old field names.
    
    copyFile - - if true, copies the data file to the project directory.
    
    overwriteFile - - overwrite any existing local data file while copying.
    
    Throws:
    
    java.io.FileNotFoundException - if the specified data file is not found (optional).
    
    java.nio.file.FileAlreadyExistsException - if local file exists during copying and overwriteFile==false (optional).
    
    java.io.IOException - if file I/O error occurs during copying or other operation.
    
    java.lang.IndexOutOfBoundsException - if there is no data file with given index.
    
    java.lang.IllegalArgumentException - if fieldNames do not match already assigned field names.
  - updateDataFileLastModified
```
public void updateDataFileLastModified(int fileIndex)
                                throws java.io.IOException
```
    Updates the stored last modified date of the data file. Determines the current date of the existing data file and writes it to XML. Removes pair finder settings. Use with caution: existing pairs must match records in the data file, otherwise operations that use the server table may fail.
    
    Parameters:
    
    fileIndex - - index of the data file: from 0 to ProjectSettings.getNDataFiles()-1
    
    Throws:
    
    java.io.IOException - if I/O error occurs while obtaining file date
    
    java.lang.IndexOutOfBoundsException - if there is no data file with given index.
  - verifyCanSetDataFile
```
public void verifyCanSetDataFile()
```
    Verifies that a data file can be assigned or replaced.
    
    Throws:
    
    java.lang.IllegalStateException - - if existing data file cannot be replaced (features, record pairs or trained model already assigned). Also if several models have been saved for the same project.
  - getDataFile
```
public java.nio.file.Path getDataFile(int fileIndex)
```
    Parameters:
    
    fileIndex - - index of the data file: from 0 to ProjectSettings.getNDataFiles()-1
    
    Returns:
    
    object containing the absolute path to the data file.
    
    Throws:
    
    java.lang.IndexOutOfBoundsException - if there is no data file with given index.
  - setKeyField
```
public void setKeyField(int fieldIndex)
```
    Sets the index of the key field in the data file. Creates empty feature query.
    
    Parameters:
    
    fieldIndex - - the index of key field in the list of field names of the data file.
    
    Throws:
    
    java.lang.IllegalStateException - if data file is not assigned; or several models have been saved; or active model has features/pairs assigned.
    
    java.lang.IllegalArgumentException - if the given field is ignored.
  - thesFileExists
```
public boolean thesFileExists(java.lang.String filePath)
```
    Returns:
    
    true if the file with the specified name exists in active model's "thesauri" directory.
  - addThesFile
```
public void addThesFile(java.lang.String filePath)
                 throws java.io.IOException
```
    Adds the thesaurus file to the "thesauri subdirectory of the active model's directory, overwriting any previous file with same name. Overwriting will cause all features that use old thesaurus file to use the new file. Check it first using thesFileExists(String).
    
    Parameters:
    
    filePath - - the thesaurus file including the absolute path.
    
    Throws:
    
    java.io.IOException - if file copying fails.
  - exportThesFiles
```
public void exportThesFiles(java.nio.file.Path outputDir)
                     throws java.io.IOException
```
    Exports all thesaurus files used by the feature query of the active model. If a file with the same name exists in output directory, this method overwrites it.
    
    Parameters:
    
    outputDir - - the directory where thesaurus files are copied to.
    
    Throws:
    
    java.io.IOException - if I/O error occurs
  - addRecPair
```
public DataPartition addRecPair(RecPair recPair)
```
    Adds a new record pair randomly to the training or validation dataset of the active model. Saves info about the currently assigned data file as the source of the pair. Data file must be assigned and contain the given records.
    
    Parameters:
    
    recPair - - information on the two records. Not null.
    
    Returns:
    
    the partition to which the record pair was added (training or validation)
    
    Throws:
    
    java.lang.IllegalStateException - if data file or key field is not assigned.
    
    java.lang.IllegalArgumentException - if rec pair with same keys has already been added to one of the datasets, or if the record pair contains incorrect number of fields or feature values.
  - addRecPair
```
public void addRecPair(RecPair recPair,
                       DataPartition partition)
```
    Adds a new record pair to the specified dataset of the active model. Saves info about the currently assigned data file as the source of the pair. Data file must be assigned and contain the given records. To add a pair to the training or validation dataset addRecPair(RecPair) must be used.
    
    Parameters:
    
    recPair - - information on the two records. Not null.
    
    partition - - the dataset to add the pair to. It must not be used in training.
    
    Throws:
    
    java.lang.IllegalStateException - if data file or key field is not assigned.
    
    java.lang.IllegalArgumentException - if rec pair with same keys has already been added to one of the datasets, or if the record pair contains incorrect number of fields or feature values, or if partition is used in training.
  - addUsefulRecPairs
```
public AddPairsResult addUsefulRecPairs(java.util.List<RecPair> recPairs,
                                        int maxTrainSize,
                                        int maxReserveSize)
                                 throws com.netrics.likeit.NetricsException,
                                        java.io.IOException
```
    Adds each record pair if it is likely to improve model training while limiting the grwth of datasets. Useful pairs are added randomly to the training or validation dataset of the active model. Before calling this method, create model trainer with table definition and load trained model if it exists. If trained model does not exist, the growth of datasets is potentially unlimited.
    
    Parameters:
    
    recPairs - - all record pairs to be added.
    
    maxTrainSize - - no more pairs will be added to Train/Vld datasets if the Training dataset reaches this maximum size. Should be as high as possible (3000 and more) as long as the model training time is acceptable.
    
    maxReserveSize - - the maximum number of pairs that can be stored to the Reserve dataset. Pairs that are deemed not useful are added to the Reserve dataset up to this limit. If 0, no pairs are added. If -1, an unlimited number of pairs are added.
    
    Returns:
    
    an object containing groups of pairs added and not added to project.
    
    Throws:
    
    java.lang.IllegalStateException - if data file or key field is not assigned, or if model trainer has not been created, or if neither the server table nor table definition is assigned, or if the server table is assigned but not loaded, or if trained model exists but was not loaded to model trainer, or if the feature query is empty.
    
    java.lang.IllegalArgumentException - if any record pair contains incorrect number of fields or feature values.
    
    com.netrics.likeit.NetricsException - if the server indicates that an error has occured
    
    java.io.IOException - if an I/O error occurs while communicating with server
  - moveUsefulRecPairs
```
public AddPairsResult moveUsefulRecPairs(DataPartition sourcePartition,
                                         int maxTrainSize)
                                  throws com.netrics.likeit.NetricsException,
                                         java.io.IOException
```
    Moves the record pairs that are likely to improve model training from the given dataset (Reserve) to the training or validation dataset of the active model. Before calling this, create model trainer with table definition and load trained model if it exists.
    
    Parameters:
    
    sourcePartition - - the dataset with the pairs to be considered. Must be a dataset not used for training (currently only the RESERVE dataset).
    
    maxTrainSize - - no more pairs will be added to Train/Vld datasets if the Training dataset reaches this maximum size. Should be as high as possible (3000 and more) as long as the model training time is acceptable.
    
    Returns:
    
    an object containing groups of pairs added to Training and Validation datasets (remaining lists of pairs are not used).
    
    Throws:
    
    java.lang.IllegalStateException - if data file or key field is not assigned, or if model trainer has not been created, or if neither the server table nor table definition is assigned, or if the server table is assigned but not loaded, or if trained model exists but was not loaded to model trainer, or if the feature query is empty.
    
    java.lang.IllegalArgumentException - if sourcePartition is not RESERVE.
    
    com.netrics.likeit.NetricsException - if the server indicates that an error has occured
    
    java.io.IOException - if an I/O error occurs while communicating with server
  - addRecPairsFromCsv
```
public void addRecPairsFromCsv(java.io.File csvFile)
                        throws com.netrics.likeit.NetricsFileFormatException,
                               java.io.IOException,
                               com.netrics.likeit.NetricsException
```
    Adds all record pairs from the specified CSV file to the active model settings. Format of each line in CSV file:"key1,key2,label", cannot contain spaces before/after comma. The keys must exist in the assigned data file. Field values are read from the server table. Label is "0" (False), "1" (True), "-" (null), "?" or any other value (Unsure).
    
    Parameters:
    
    csvFile - - the CSV file with record pairs
    
    Throws:
    
    java.lang.IllegalStateException - if data file or key field is not assigned; or the ModelTrainer is not created or does not have a server table assigned.
    
    java.io.FileNotFoundException - if the CSV file is not found (optional).
    
    com.netrics.likeit.NetricsFileFormatException - if CSV file format is incorrect
    
    com.netrics.likeit.NetricsException - if the server indicates that an error has occured
    
    java.io.IOException - if pairs CSV file has wrong number of entries per line; or I/O error occurs while communicating with server
  - createTrainer
```
public ModelTrainer createTrainer(com.netrics.likeit.NetricsServerInterface server)
                           throws com.netrics.likeit.NetricsException,
                                  java.io.IOException
```
    Creates a ModelTrainer for training the active model. Closes any existing trainer. Note: client must still call ModelTrainer.close(boolean) after scorer or RLink model has been created and the ModelTrainer object will not be replaced by calling this method.
    
    Parameters:
    
    server - - Patterns server must be running and accessible.
    
    Throws:
    
    com.netrics.likeit.NetricsException - if an error occured on server.
    
    java.io.IOException - if the server does not exist, is not accessible or if an I/O error occurs while communicating with server.
  - createScorer
```
public RecPairScorer createScorer(com.netrics.likeit.NetricsServerInterface server)
                           throws com.netrics.likeit.NetricsException,
                                  java.io.IOException
```
    Creates a RecPairScorer that can be used only for scoring any record pairs with known field values. Client must close the returned scorer object when it is no longer used. Scorer will use the field names, types and feature query defined for the active model. This scorer object does not load or require a table on server. Use ModelTrainer to score pairs in the project, train models, etc. See createTrainer(NetricsServerInterface).
    
    Parameters:
    
    server - - Patterns server must be running and accessible.
    
    Returns:
    
    the newly created RecPairScorer.
    
    Throws:
    
    java.lang.IllegalStateException - if data file or key field is not assigned, or if no features are defined.
    
    com.netrics.likeit.NetricsException - if an error occured on server.
    
    java.io.IOException - if the server does not exist, is not accessible or if an I/O error occurs while communicating with server.
  - saveTrainedModel
```
public void saveTrainedModel()
                      throws javax.xml.bind.JAXBException,
                             com.netrics.likeit.NetricsException,
                             java.io.IOException
```
    Saves trained model binary file to the active model directory. Stores in active model settings: binary file name, training date (current timestamp), training iteration, model score for each pair. Saves project/model XML files to disk (to be in sync with the saved binary file).
    
    Throws:
    
    java.lang.IllegalStateException - if ModelTrainer does not exist, or if no training iterations were performed.
    
    javax.xml.bind.JAXBException - - if XML validation fails or other JAXB error occurs.
    
    com.netrics.likeit.NetricsException - if the server error occured when deleting old loaded model.
    
    java.io.IOException - if the model file cannot be created.
  - hasModelTrainConfig
```
public boolean hasModelTrainConfig()
```
    Checks whether a custom model training configuration is used to train the model.
    
    Returns:
    
    true if the modelTrainConfig.xml file exists in the project.
  - getModelTrainConfig
```
public ModelTrainConfig getModelTrainConfig()
```
    Get model training configuration. Changes to the returned object will be visible in ProjectFiles and ModelTrainer (if it exists).
    
    Returns:
    
    the custom model training configuration (internal reference), or null if the project does not contain the modelTrainConfig.xml file.
  - createModelTrainConfig
```
public ModelTrainConfig createModelTrainConfig(ModelConfig mConfig,
                                               java.lang.Boolean randomOrder,
                                               java.util.Map<Partition,ConvergenceObserver> observers)
                                        throws org.xml.sax.SAXException
```
    Creates and remembers a new ModelTrainConfig, replacing any previous one. After this call, modelTrainConfig.xml will be saved to project directory when the project is saved. Updates ModelTrainer to use the new ModelTrainConfig.
    
    Parameters:
    
    mConfig - - values to initialize the model configuration in ModelTrainConfig. If null, modelConfig information will be blank (default).
    
    randomOrder - - value for random order of training examples. Default if null.
    
    observers - - convergence observers to initialize training configuration in ModelTrainConfig. Must contain COErrorRateMin for Validation dataset and COErrorRate for Training dataset.
    
    Returns:
    
    the new ModelTrainConfig object with default settings.
    
    Throws:
    
    org.xml.sax.SAXException - if the schema file cannot be parsed.
    
    java.lang.IllegalArgumentException - if observers contain objects other than COErrorRate or COErrorRateMin.
  - deleteTrainedModel
```
public void deleteTrainedModel()
                        throws javax.xml.bind.JAXBException,
                               java.io.IOException
```
    Deletes binary file of the trained model. Deletes training statistics, predictions for all pairs, pair finder state from active ModelSettings. Saves project.
    
    Throws:
    
    java.lang.IllegalStateException - if trained model has not been saved.
    
    javax.xml.bind.JAXBException - - if XML validation fails.
    
    java.io.IOException - - if deleting the binary file fails.
  - loadTrainedModel
```
public void loadTrainedModel()
                      throws java.io.FileNotFoundException
```
    Loads the trained model from the model.bin file to ModelTrainer of the active model.
    
    Throws:
    
    java.io.FileNotFoundException - if the model file is missing.
    
    java.lang.IllegalStateException - if ModelTrainer has not been created; or trained model was not saved in ModelSettings; or if a model is already created or loaded to ModelTrainer.
  - createLowConfPairFinder
```
public LowConfPairFinder createLowConfPairFinder()
                                          throws com.netrics.likeit.NetricsException,
                                                 java.io.IOException
```
    Creates and stores an object that finds low confidence pairs in the data table. If needed, creates scorer, loads the trained model to server and to the ModelTrainer. Any pairs saved from the last LowConfPairFinder are removed from the active model, and those that still have low confidence are stored in the returned object.
    
    Returns:
    
    the new pair finder object.
    
    Throws:
    
    java.lang.IllegalStateException - if trained model has not been saved, or if ModelTrainer has not been created, or if server table has not been assigned/loaded or if the table is empty, or if feature confidence measure is not used.
    
    java.io.FileNotFoundException - if the model file does not exist.
    
    com.netrics.likeit.NetricsException - if the server indicates that an error has occured.
    
    java.io.IOException - if an I/O error occurs while communicating with server.
  - hasLowConfPairFinder
```
public boolean hasLowConfPairFinder()
```
    Returns:
    
    true of the pair finder has been created.
  - saveLowConfPairFinder
```
public void saveLowConfPairFinder(int maxNPairs)
```
    Saves the state of the low confidence pair finder that was created before and used to find pairs. Saves the table position and confidence threshold in the active model. The found unprocessed pairs are saved in the LOW_CONF dataset. Any previous pairs in that dataset are deleted. Unprocessed pairs are not saved if they already exist in the project or are very similar to the last returned pair. Thus the LOW_CONF dataset will empty if all the unprocessed pairs were skipped.
    
    Parameters:
    
    maxNPairs - - maximum number of found pairs to save.
    
    Throws:
    
    java.lang.IllegalStateException - if a pair finder has not been created.
  - exportTrainedModel
```
public void exportTrainedModel(java.nio.file.Path exportedFile,
                               java.lang.String version)
                        throws java.io.IOException
```
    Exports the trained model binary file from the active ModelSettings to the given location. Saves model score threshold and the given version in the exported file.
    
    Parameters:
    
    exportedFile - - destination binary model file.
    
    version - - new version value for model file. If null, version is not assigned.
    
    Throws:
    
    java.nio.file.FileAlreadyExistsException - if the given file already exists (optional).
    
    java.io.IOException - for various file copy errors
    
    java.lang.IllegalStateException - if trained model has not been saved for the active model.
  - exportTrainedModel
```
public void exportTrainedModel(java.nio.file.Path exportedFile)
                        throws java.io.IOException
```
    Exports the trained model binary file with model score threshold.
    
    Parameters:
    
    exportedFile - - destination binary model file.
    
    Throws:
    
    java.nio.file.FileAlreadyExistsException - if the given file already exists (optional).
    
    java.io.IOException - for various file copy errors
    
    java.lang.IllegalStateException - if trained model has not been saved for the active model.
    
    See Also:
    
    exportTrainedModel(Path, String)
  - exportModelAndQuery
```
public void exportModelAndQuery(int modelIndex,
                                java.nio.file.Path exportedModelFile,
                                java.lang.String modelVersion,
                                java.nio.file.Path queryOutputDir,
                                java.lang.String javaClassName,
                                java.lang.String methodName)
                         throws javax.xml.bind.JAXBException,
                                java.io.IOException
```
    Exports all files needed to use the trained model on the Patterns server: the trained model binary file, Java source file that builds the Learn query, and any thesaurus files used by the query. Exports the files for any given saved model. Any previously obtained references to the active ModelSettings and related objects remain valid.
    
    Parameters:
    
    modelIndex - - the index of the model in the list of models (in ProjectSettings).
    
    exportedModelFile - - destination binary model file.
    
    modelVersion - - new version value for model file. If null, version is not assigned.
    
    queryOutputDir - - the directory where Java source file and thesaurus files are saved. Not null.
    
    javaClassName - - name of the Java class in the exported source. May include package name. If package is not included, the default package is used. Not null.
    
    methodName - - the name of the method that returns NetricsQuery in the exported Java source. Not null.
    
    Throws:
    
    javax.xml.bind.JAXBException - when XML validation fails, or if an XML file was created with newer Learn API. Previous references to the active ModelSettings are not guaranteed to be valid in this case.
    
    java.nio.file.FileAlreadyExistsException - if the given file already exists (optional).
    
    java.io.FileNotFoundException - if the model XML file for modelIndex is not found (optional).
    
    java.io.IOException - if I/O error occurs.
    
    java.lang.IllegalStateException - if trained model has not been saved for the active model.
    
    java.lang.IllegalArgumentException - if javaClassName is invalid.

Class ProjectFiles

Method Summary

Methods inherited from class java.lang.Object

Method Detail

create

load

verifyDir

toString

getProjectSettings

setDebugOutput

save

saveAs

getActiveModelSettings

getNModels

loadModel

switchActiveModel

saveModelCopy

deleteModel

localDataFileExists

setDataFile

replaceDataFile

updateDataFileLastModified

verifyCanSetDataFile

getDataFile

setKeyField

thesFileExists

addThesFile

exportThesFiles

addRecPair

addRecPair

addUsefulRecPairs

moveUsefulRecPairs

addRecPairsFromCsv

createTrainer

createScorer

saveTrainedModel

hasModelTrainConfig

getModelTrainConfig

createModelTrainConfig

deleteTrainedModel

loadTrainedModel

createLowConfPairFinder

hasLowConfPairFinder

saveLowConfPairFinder

exportTrainedModel

exportTrainedModel

exportModelAndQuery