Examples with SparkFileSchemaParser com.thinkbiganalytics.discovery.parser.SparkFileSchemaParser used on opensource projects

Example 1 with SparkFileSchemaParser

use of com.thinkbiganalytics.discovery.parser.SparkFileSchemaParser in project kylo by Teradata.

the class FileMetadataResultModifier method modifySuccessfulResults.

@Override
public void modifySuccessfulResults(ModifiedTransformResponse<FileMetadataResponse> modifiedTransformResponse) {
    FileMetadataResponse fileMetadataResponse = new FileMetadataResponse();
    Map<String, List<FileMetadataResponse.ParsedFileMetadata>> groupedMetadata = response.getResults().getRows().stream().map(objects -> {
        FileMetadata fileMetadata = new FileMetadata();
        String mimeType = (String) getRowValue("mimeType", objects);
        String delimiter = (String) getRowValue("delimiter", objects);
        Integer headerCount = getRowValue("headerCount", objects, Integer.class, 0);
        String filePath = (String) getRowValue("resource", objects);
        String encoding = (String) getRowValue("encoding", objects);
        FileMetadataResponse.ParsedFileMetadata metadata = new FileMetadataResponse.ParsedFileMetadata();
        metadata.setMimeType(mimeType);
        metadata.setDelimiter(delimiter);
        metadata.setEncoding(encoding);
        metadata.setFilePath(filePath);
        return metadata;
    }).collect(Collectors.groupingBy(FileMetadataResponse.ParsedFileMetadata::getKey));
    // create user friendly names from the groupings
    // the key here is going to be
    Map<String, FileMetadataResponse.FileDataSet> fileDatasets = groupedMetadata.keySet().stream().collect(Collectors.toMap(key -> groupedMetadata.get(key).get(0).getFilePath(), key -> {
        FileMetadataResponse.FileDataSet dataSet = new FileMetadataResponse.FileDataSet();
        List<FileMetadataResponse.ParsedFileMetadata> files = groupedMetadata.get(key);
        dataSet.setFiles(files);
        FileMetadataResponse.ParsedFileMetadata firstFile = files.get(0);
        dataSet.setMimeType(firstFile.getMimeType());
        findSchemaParserForMimeType(firstFile.getMimeType()).ifPresent(schemaParserDescriptor -> {
            dataSet.setSchemaParser(schemaParserDescriptor);
            // if the parser has a delimiter property set it
            schemaParserDescriptor.getProperties().stream().filter(property -> property.getObjectProperty().equals("separatorChar")).findFirst().ifPresent(property -> {
                property.setValue(firstFile.getDelimiter());
            });
            Optional<FileSchemaParser> fileSchemaParser = fileSchemaParser(schemaParserDescriptor);
            if (fileSchemaParser.isPresent() && fileSchemaParser.get() instanceof SparkFileSchemaParser) {
                List<String> paths = files.stream().map(parsedFileMetadata -> parsedFileMetadata.getFilePath()).collect(Collectors.toList());
                SampleFileSparkScript sparkScript = ((SparkFileSchemaParser) fileSchemaParser.get()).getSparkScript(paths);
                dataSet.setSparkScript(sparkScript);
            }
        });
        return dataSet;
    }));
    fileMetadataResponse.setDatasets(fileDatasets);
    modifiedTransformResponse.setResults(fileMetadataResponse);
}

Example 2 with SparkFileSchemaParser

use of com.thinkbiganalytics.discovery.parser.SparkFileSchemaParser in project kylo by Teradata.

the class SchemaDiscoveryRestController method uploadFileSpark.

/**
 * Generate the spark script that can parse the passed in file using the passed in "parserDescriptor"
 *
 * @param parserDescriptor  metadata about how the file should be parsed
 * @param dataFrameVariable the name of the dataframe variable in the generate spark code
 * @param limit             a number indicating how many rows the script should limit the output
 * @param fileInputStream   the file
 * @param fileMetaData      metadata about the file
 * @return an object including the name of the file on disk and the generated spark script
 */
@POST
@Path("/spark/sample-file")
@Consumes(MediaType.MULTIPART_FORM_DATA)
@Produces(MediaType.APPLICATION_JSON)
@ApiOperation("Determines the schema of the provided file.")
@ApiResponses({ @ApiResponse(code = 200, message = "Returns the spark script that parses the sample file.", response = Schema.class), @ApiResponse(code = 500, message = "The schema could not be determined.", response = RestResponseStatus.class) })
public Response uploadFileSpark(@FormDataParam("parser") String parserDescriptor, @FormDataParam("dataFrameVariable") @DefaultValue("df") String dataFrameVariable, @FormDataParam("limit") @DefaultValue("-1") Integer limit, @FormDataParam("file") InputStream fileInputStream, @FormDataParam("file") FormDataContentDisposition fileMetaData) throws Exception {
    SampleFileSparkScript sampleFileSparkScript = null;
    SchemaParserAnnotationTransformer transformer = new SchemaParserAnnotationTransformer();
    try {
        SchemaParserDescriptor descriptor = ObjectMapperSerializer.deserialize(parserDescriptor, SchemaParserDescriptor.class);
        FileSchemaParser p = transformer.fromUiModel(descriptor);
        SparkFileSchemaParser sparkFileSchemaParser = (SparkFileSchemaParser) p;
        sparkFileSchemaParser.setDataFrameVariable(dataFrameVariable);
        sparkFileSchemaParser.setLimit(limit);
        sampleFileSparkScript = sparkFileSchemaParser.getSparkScript(fileInputStream);
    } catch (IOException e) {
        throw new WebApplicationException(e.getMessage());
    } catch (PolicyTransformException e) {
        log.warn("Failed to convert parser", e);
        throw new InternalServerErrorException(STRINGS.getString("discovery.transformError"), e);
    }
    if (sampleFileSparkScript == null) {
        log.warn("Failed to convert parser");
        throw new InternalServerErrorException(STRINGS.getString("discovery.transformError"));
    }
    return Response.ok(sampleFileSparkScript).build();
}

Example 3 with SparkFileSchemaParser

use of com.thinkbiganalytics.discovery.parser.SparkFileSchemaParser in project kylo by Teradata.

the class SchemaDiscoveryRestController method sparkScriptForFilesList.

/**
 * Generate the spark script that can parse the passed in file using the passed in "parserDescriptor"
 *
 * @return an object including the name of the file on disk and the generated spark script
 */
@POST
@Path("/spark/files-list")
@Consumes(MediaType.APPLICATION_JSON)
@Produces(MediaType.APPLICATION_JSON)
@ApiOperation("Determines the schema of the provided file.")
@ApiResponses({ @ApiResponse(code = 200, message = "Returns the spark script that parses the sample file.", response = Schema.class), @ApiResponse(code = 500, message = "The schema could not be determined.", response = RestResponseStatus.class) })
public Response sparkScriptForFilesList(SparkFilesScript sparkFilesScript) throws Exception {
    SampleFileSparkScript sampleFileSparkScript = null;
    SchemaParserAnnotationTransformer transformer = new SchemaParserAnnotationTransformer();
    try {
        SchemaParserDescriptor descriptor = ObjectMapperSerializer.deserialize(sparkFilesScript.getParserDescriptor(), SchemaParserDescriptor.class);
        FileSchemaParser p = transformer.fromUiModel(descriptor);
        SparkFileSchemaParser sparkFileSchemaParser = (SparkFileSchemaParser) p;
        sparkFileSchemaParser.setDataFrameVariable(sparkFilesScript.getDataFrameVariable());
        sparkFileSchemaParser.setLimit(-1);
        sampleFileSparkScript = sparkFileSchemaParser.getSparkScript(sparkFilesScript.getFiles());
    } catch (PolicyTransformException e) {
        log.warn("Failed to convert parser", e);
        throw new InternalServerErrorException(STRINGS.getString("discovery.transformError"), e);
    }
    if (sampleFileSparkScript == null) {
        log.warn("Failed to convert parser");
        throw new InternalServerErrorException(STRINGS.getString("discovery.transformError"));
    }
    return Response.ok(sampleFileSparkScript).build();
}

Example 4 with SparkFileSchemaParser

use of com.thinkbiganalytics.discovery.parser.SparkFileSchemaParser in project kylo by Teradata.

the class FileMetadataResultModifier method fileSchemaParser.

private Optional<FileSchemaParser> fileSchemaParser(SchemaParserDescriptor descriptor) {
    try {
        SchemaParserAnnotationTransformer transformer = new SchemaParserAnnotationTransformer();
        FileSchemaParser p = transformer.fromUiModel(descriptor);
        if (p instanceof SparkFileSchemaParser) {
            SparkFileSchemaParser sparkFileSchemaParser = (SparkFileSchemaParser) p;
            sparkFileSchemaParser.setDataFrameVariable("df");
            sparkFileSchemaParser.setLimit(-1);
        }
        return Optional.of(p);
    } catch (Exception e) {
        return Optional.empty();
    }
}

Example 5 with SparkFileSchemaParser

use of com.thinkbiganalytics.discovery.parser.SparkFileSchemaParser in project kylo by Teradata.

the class FileMetadataTransformResponseModifier method fileSchemaParser.

private Optional<FileSchemaParser> fileSchemaParser(SchemaParserDescriptor descriptor) {
    try {
        SchemaParserAnnotationTransformer transformer = new SchemaParserAnnotationTransformer();
        FileSchemaParser p = transformer.fromUiModel(descriptor);
        if (p instanceof SparkFileSchemaParser) {
            SparkFileSchemaParser sparkFileSchemaParser = (SparkFileSchemaParser) p;
            sparkFileSchemaParser.setDataFrameVariable("df");
            sparkFileSchemaParser.setLimit(-1);
        }
        return Optional.of(p);
    } catch (Exception e) {
        return Optional.empty();
    }
}