internal object Read(Reader reader, Header header, AbstractCodec codec, Resolver resolver)
{
if (reader.IsReadToEnd())
{
return(string.Empty);
}
var result = new List <object>();
do
{
long itemsCount = reader.ReadLong();
var data = reader.ReadDataBlock(header.SyncData, codec);
reader = new Reader(new MemoryStream(data));
if (itemsCount > 1)
{
for (int i = 0; i < itemsCount; i++)
{
result.Add(resolver.Resolve(reader));
}
}
else
{
return(resolver.Resolve(reader));
}
} while (!reader.IsReadToEnd());
return(result);
}
internal object Read(IReader reader, Header header, AbstractCodec codec, Resolver resolver)
{
var remainingBlocks = reader.ReadLong();
var blockSize = reader.ReadLong();
var syncBuffer = new byte[DataFileConstants.SyncSize];
var dataBlock = new byte[blockSize];
reader.ReadFixed(dataBlock, 0, (int)blockSize);
reader.ReadFixed(syncBuffer);
if (!syncBuffer.SequenceEqual(header.SyncData))
{
throw new AvroRuntimeException("Invalid sync!");
}
dataBlock = codec.Decompress(dataBlock);
reader = new Reader(new MemoryStream(dataBlock));
if (remainingBlocks > 1)
{
var result = new List <object>();
for (int i = 0; i < remainingBlocks; i++)
{
result.Add(resolver.Resolve(reader));
}
return(result);
}
else
{
return(resolver.Resolve(reader));
}
}
internal void WriteHeader(string schema, CodecType codecType)
{
GenerateSyncData();
_header.AddMetadata(DataFileConstants.CodecMetadataKey, AbstractCodec.CreateCodec(codecType).Name);
_header.AddMetadata(DataFileConstants.SchemaMetadataKey, schema);
_writer.WriteHeader(_header);
}
internal byte[] ReadDataBlock(byte[] syncData, AbstractCodec codec)
{
var dataBlock = ReadRawBlock();
ReadAndValidateSync(syncData);
dataBlock = codec.Decompress(dataBlock);
return(dataBlock);
}
internal object Decode(Stream stream, Schema.Schema schema)
{
var reader = new Reader(stream);
var header = new Header();
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//does not contain header
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
if (schema == null)
{
throw new MissingSchemaException("Provide valid schema for the Avro data");
}
var resolver = new Resolver(schema);
stream.Seek(0, SeekOrigin.Begin);
return(resolver.Resolve(reader));
}
else
{
// read meta data
long len = reader.ReadMapStart();
if (len > 0)
{
do
{
for (long i = 0; i < len; i++)
{
string key = reader.ReadString();
byte[] val = reader.ReadBytes();
header.AddMetadata(key, val);
}
} while ((len = reader.ReadMapNext()) != 0);
}
schema = schema ?? Schema.Schema.Parse(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
var resolver = new Resolver(schema);
// read in sync data
reader.ReadFixed(header.SyncData);
var codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
return(Read(reader, header, codec, resolver));
}
}
internal AvroObjectContent ExtractAvroObjectContent(byte[] avroObject)
{
using (var stream = new MemoryStream(avroObject))
{
var reader = new Reader(stream);
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//does not contain header
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
throw new InvalidAvroObjectException("Object does not contain Avro Header");
}
else
{
AvroObjectContent result = new AvroObjectContent();
var header = reader.ReadHeader();
result.Codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
reader.ReadFixed(header.SyncData);
result.Header = header;
result.Header.Schema = Schema.Create(result.Header.GetMetadata(DataFileConstants.SchemaMetadataKey));
if (reader.IsReadToEnd())
{
return(result);
}
do
{
var blockContent = new DataBlock
{
ItemsCount = reader.ReadLong(),
Data = reader.ReadDataBlock(header.SyncData, result.Codec)
};
result.DataBlocks.Add(blockContent);
} while (!reader.IsReadToEnd());
return(result);
}
}
}
public void CreateCodecFromString_NonExistingString_DefaultCodecIsReturned()
{
//Arrange
//Act
var result = AbstractCodec.CreateCodecFromString("NonExistingCodec");
//Assert
Assert.IsType <NullCodec>(result);
}
internal T Decode <T>(Stream stream, Schema.Schema readSchema)
{
var reader = new Reader(stream);
var header = new Header();
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//does not contain header
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
throw new InvalidAvroObjectException("Object does not contain Avro Header");
}
else
{
// read meta data
long len = reader.ReadMapStart();
if (len > 0)
{
do
{
for (long i = 0; i < len; i++)
{
string key = reader.ReadString();
byte[] val = reader.ReadBytes();
header.AddMetadata(key, val);
}
} while ((len = reader.ReadMapNext()) != 0);
}
readSchema = readSchema ?? Schema.Schema.Parse(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
Schema.Schema writeSchema = Schema.Schema.Parse(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
var resolver = new Resolver(writeSchema, readSchema);
// read in sync data
reader.ReadFixed(header.SyncData);
var codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
return(Read <T>(reader, header, codec, resolver));
}
}
internal BaseLineReader(Reader reader, byte[] syncDate, AbstractCodec codec, TypeSchema writeSchema, TypeSchema readSchema)
{
_reader = reader;
_syncDate = syncDate;
_codec = codec;
_writeSchema = writeSchema;
_readSchema = readSchema;
if (_reader.IsReadToEnd())
{
return;
}
LoadNextDataBlock();
}
internal MergeEncoder(Stream outStream)
{
_codec = new NullCodec();
_stream = outStream;
_syncInterval = DataFileConstants.DefaultSyncInterval;
_blockCount = 0;
_writer = new Writer(_stream);
_tempBuffer = new MemoryStream();
_tempWriter = new Writer(_tempBuffer);
_isOpen = true;
_header = new Header();
}
private Decoder(Stream stream)
{
_stream = stream;
_header = new Header();
_reader = new Reader(stream);
_syncBuffer = new byte[DataFileConstants.SyncSize];
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
_reader.ReadFixed(firstBytes);
}
catch (Exception)
{
throw new InvalidAvroObjectException("Cannot read length of Avro Header");
}
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
throw new InvalidAvroObjectException("Cannot read Avro Header");
}
// read meta data
long len = _reader.ReadMapStart();
if (len > 0)
{
do
{
for (long i = 0; i < len; i++)
{
string key = _reader.ReadString();
byte[] val = _reader.ReadBytes();
_header.MetaData.Add(key, val);
}
} while ((len = _reader.ReadMapNext()) != 0);
}
// read in sync data
_reader.ReadFixed(_header.SyncData);
// parse schema and set codec
_header.Schema = Schema.Schema.Parse(GetMetaString(DataFileConstants.SchemaMetadataKey));
_resolver = new Resolver(_header.Schema, _readerSchema ?? _header.Schema);
_codec = AbstractCodec.CreateCodecFromString(GetMetaString(DataFileConstants.CodecMetadataKey));
}
internal Encoder(Schema.Schema schema, Stream outStream, CodecType codecType)
{
_codec = AbstractCodec.CreateCodec(codecType);
_stream = outStream;
_metadata = new Metadata();
_schema = schema;
_syncInterval = DataFileConstants.DefaultSyncInterval;
_blockCount = 0;
_encoder = new Writer(_stream);
_blockStream = new MemoryStream();
_blockEncoder = new Writer(_blockStream);
_writer = Resolver.ResolveWriter(schema);
_isOpen = true;
}
internal static ILineReader <T> OpenReader <T>(Stream stream, TypeSchema readSchema)
{
var reader = new Reader(stream);
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//headless
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
if (readSchema == null)
{
throw new MissingSchemaException("Provide valid schema for the Avro data");
}
var resolver = new Resolver(readSchema, readSchema);
stream.Seek(0, SeekOrigin.Begin);
return(new ListLineReader <T>(reader, resolver));
}
else
{
var header = reader.ReadHeader();
readSchema = readSchema ?? Schema.Create(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
TypeSchema writeSchema = Schema.Create(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
// read in sync data
reader.ReadFixed(header.SyncData);
var codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
return(new BaseLineReader <T>(reader, header.SyncData, codec, writeSchema, readSchema));
}
}
internal T Read <T>(IReader reader, Header header, AbstractCodec codec, Resolver resolver)
{
var remainingBlocks = reader.ReadLong();
var blockSize = reader.ReadLong();
var syncBuffer = new byte[DataFileConstants.SyncSize];
var dataBlock = new byte[blockSize];
reader.ReadFixed(dataBlock, 0, (int)blockSize);
reader.ReadFixed(syncBuffer);
if (!syncBuffer.SequenceEqual(header.SyncData))
{
throw new AvroRuntimeException("Invalid sync!");
}
dataBlock = codec.Decompress(dataBlock);
reader = new Reader(new MemoryStream(dataBlock));
return(resolver.Resolve <T>(reader, remainingBlocks));
}
internal Encoder(TypeSchema schema, Stream outStream, CodecType codecType)
{
_codec = AbstractCodec.CreateCodec(codecType);
_stream = outStream;
_header = new Header();
_schema = schema;
_syncInterval = DataFileConstants.DefaultSyncInterval;
_blockCount = 0;
_writer = new Writer(_stream);
_tempBuffer = new MemoryStream();
_tempWriter = new Writer(_tempBuffer);
GenerateSyncData();
_header.AddMetadata(DataFileConstants.CodecMetadataKey, _codec.Name);
_header.AddMetadata(DataFileConstants.SchemaMetadataKey, _schema.ToString());
_writeItem = Resolver.ResolveWriter(schema);
_isOpen = true;
}
internal object Decode(Stream stream, TypeSchema schema)
{
var reader = new Reader(stream);
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//does not contain header
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
if (schema == null)
{
throw new MissingSchemaException("Provide valid schema for the Avro data");
}
var resolver = new Resolver(schema);
stream.Seek(0, SeekOrigin.Begin);
return(resolver.Resolve(reader));
}
else
{
var header = reader.ReadHeader();
schema = schema ?? Schema.Create(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
var resolver = new Resolver(schema);
reader.ReadFixed(header.SyncData);
var codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
return(Read(reader, header, codec, resolver));
}
}
internal T Decode <T>(Stream stream, TypeSchema readSchema)
{
var reader = new Reader(stream);
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//does not contain header
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
throw new InvalidAvroObjectException("Object does not contain Avro Header");
}
else
{
var header = reader.ReadHeader();
TypeSchema writeSchema = Schema.Create(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
readSchema = readSchema ?? writeSchema;
var resolver = new Resolver(writeSchema, readSchema);
// read in sync data
reader.ReadFixed(header.SyncData);
var codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
return(Read <T>(reader, header, codec, resolver));
}
}
internal T Read <T>(Reader reader, Header header, AbstractCodec codec, Resolver resolver)
{
if (reader.IsReadToEnd())
{
return(default);
internal static ILineReader <T> OpenReader <T>(Stream stream, Schema.Schema readSchema)
{
var reader = new Reader(stream);
var header = new Header();
// validate header
byte[] firstBytes = new byte[DataFileConstants.AvroHeader.Length];
try
{
reader.ReadFixed(firstBytes);
}
catch (EndOfStreamException)
{
//stream shorter than AvroHeader
}
//does not contain header
if (!firstBytes.SequenceEqual(DataFileConstants.AvroHeader))
{
if (readSchema == null)
{
throw new MissingSchemaException("Provide valid schema for the Avro data");
}
var resolver = new Resolver(readSchema, readSchema);
stream.Seek(0, SeekOrigin.Begin);
return(new ListLineReader <T>(reader, resolver));
}
else
{
// read meta data
long len = reader.ReadMapStart();
if (len > 0)
{
do
{
for (long i = 0; i < len; i++)
{
string key = reader.ReadString();
byte[] val = reader.ReadBytes();
header.AddMetadata(key, val);
}
} while ((len = reader.ReadMapNext()) != 0);
}
readSchema = readSchema ?? Schema.Schema.Parse(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
Schema.Schema writeSchema = Schema.Schema.Parse(header.GetMetadata(DataFileConstants.SchemaMetadataKey));
var resolver = new Resolver(writeSchema, readSchema);
// read in sync data
reader.ReadFixed(header.SyncData);
var codec = AbstractCodec.CreateCodecFromString(header.GetMetadata(DataFileConstants.CodecMetadataKey));
var remainingBlocks = reader.ReadLong();
var blockSize = reader.ReadLong();
var syncBuffer = new byte[DataFileConstants.SyncSize];
var dataBlock = new byte[blockSize];
reader.ReadFixed(dataBlock, 0, (int)blockSize);
reader.ReadFixed(syncBuffer);
if (!syncBuffer.SequenceEqual(header.SyncData))
{
throw new AvroRuntimeException("Invalid sync!");
}
dataBlock = codec.Decompress(dataBlock);
reader = new Reader(new MemoryStream(dataBlock));
if (remainingBlocks > 1)
{
return(new BlockLineReader <T>(reader, resolver, remainingBlocks));
}
if (writeSchema.Tag == Schema.Schema.Type.Array)
{
return(new ListLineReader <T>(reader, new Resolver(((ArraySchema)writeSchema).ItemSchema, readSchema)));
}
return(new ListLineReader <T>(reader, new Resolver(writeSchema, readSchema)));
}
}