def test_readAndWriteTextualFileHeader(self): """ Reading and writing should not change the textual file header. """ for file, attribs in self.files.items(): endian = attribs['endian'] header_enc = attribs['textual_header_enc'] file = os.path.join(self.path, file) # Read the file. with open(file, 'rb') as f: org_header = f.read(3200) f.seek(0, 0) # Initialize an empty SEGY object and set certain attributes. segy = SEGYFile() segy.endian = endian segy.file = f segy.textual_header_encoding = None # Read the textual header. segy._readTextualHeader() # Assert the encoding and compare with known values. self.assertEqual(segy.textual_header_encoding, header_enc) # The header writes to a file like object. new_header = io.BytesIO() segy._writeTextualHeader(new_header) new_header.seek(0, 0) new_header = new_header.read() # Assert the correct length. self.assertEqual(len(new_header), 3200) # Assert the actual header. self.assertEqual(org_header, new_header)
def writeSEGY(stream, filename, data_encoding=None, byteorder=None, textual_header_encoding=None, **kwargs): # @UnusedVariable """ Writes a SEG Y file from given ObsPy Stream object. .. warning:: This function should NOT be called directly, it registers via the the :meth:`~obspy.core.stream.Stream.write` method of an ObsPy :class:`~obspy.core.stream.Stream` object, call this instead. :type stream: :class:`~obspy.core.stream.Stream` :param stream: The ObsPy Stream object to write. :type filename: str :param filename: Name of file to write. :type data_encoding: int :param data_encoding: The data encoding is an integer with the following currently supported meaning: ``1`` 4 byte IBM floating points (float32) ``2`` 4 byte Integers (int32) ``3`` 2 byte Integer (int16) ``5`` 4 byte IEEE floating points (float32) The value in the brackets is the necessary dtype of the data. ObsPy will now automatically convert the data because data might change/loose precision during the conversion so the user has to take care of the correct dtype. If it is ``None``, the value of the first Trace will be used for all consecutive Traces. If it is None for the first Trace, 1 (IBM floating point numbers) will be used. Different data encodings for different traces are currently not supported because these will most likely not be readable by other software. :type byteorder: str or ``None`` :param byteorder: Determines the endianness of the file. Either ``'>'`` for big endian or ``'<'`` for little endian. If is ``None``, it will either be the endianness of the first Trace or if that is also not set, it will be big endian. A mix between little and big endian for the headers and traces is currently not supported. :type textual_header_encoding: str or ``None`` :param textual_header_encoding: The encoding of the textual header. Can be ``'EBCDIC'``, ``'ASCII'`` or ``None``. If it is ``None``, the textual_file_header_encoding attribute in the stats.segy dictionary of the first Trace is used and if that is not set, ASCII will be used. This function will automatically set the data encoding field of the binary file header so the user does not need to worry about it. The starttime of every trace is not a required field in the SEG Y specification. If the starttime of a trace is UTCDateTime(0) it will be interpreted as a not-set starttime and no time is written to the trace header. Every other time will be written. SEG Y supports a sample interval from 1 to 65535 microseconds in steps of 1 microsecond. Larger intervals cannot be supported due to the definition of the SEG Y format. Therefore the smallest possible sampling rate is ~ 15.26 Hz. Please keep that in mind. """ # Some sanity checks to catch invalid arguments/keyword arguments. if data_encoding is not None and data_encoding not in VALID_FORMATS: msg = "Invalid data encoding." raise SEGYCoreWritingError(msg) # Figure out the data encoding if it is not set. if data_encoding is None: if hasattr(stream, 'stats') and hasattr(stream.stats, 'data_encoding'): data_encoding = stream.stats.data_encoding if hasattr(stream, 'stats') and hasattr(stream.stats, 'binary_file_header'): data_encoding = \ stream.stats.binary_file_header.data_sample_format_code # Set it to float if it in not given. else: data_encoding = 1 # Create empty file wide headers if they do not exist. if not hasattr(stream, 'stats'): stream.stats = AttribDict() if not hasattr(stream.stats, 'textual_file_header'): stream.stats.textual_file_header = b"" if not hasattr(stream.stats, 'binary_file_header'): stream.stats.binary_file_header = SEGYBinaryFileHeader() # Valid dtype for the data encoding. valid_dtype = DATA_SAMPLE_FORMAT_CODE_DTYPE[data_encoding] # Makes sure that the dtype is for every Trace is correct. for trace in stream: # Check the dtype. if trace.data.dtype != valid_dtype: msg = """ The dtype of the data and the chosen data_encoding do not match. You need to manually convert the dtype if you want to use that data_encoding. Please refer to the obspy.segy manual for more details. """.strip() raise SEGYCoreWritingError(msg) # Check the sample interval. if trace.stats.delta > MAX_INTERVAL_IN_SECONDS: msg = """ SEG Y supports a maximum interval of %s seconds in between two samples (trace.stats.delta value). """.strip() msg = msg % MAX_INTERVAL_IN_SECONDS raise SEGYSampleIntervalError(msg) # Figure out endianness and the encoding of the textual file header. if byteorder is None: if hasattr(stream, 'stats') and hasattr(stream.stats, 'endian'): byteorder = stream.stats.endian else: byteorder = '>' # Map the byteorder. byteorder = ENDIAN[byteorder] if textual_header_encoding is None: if hasattr(stream, 'stats') and hasattr( stream.stats, 'textual_file_header_encoding'): textual_header_encoding = \ stream.stats.textual_file_header_encoding else: textual_header_encoding = 'ASCII' # Loop over all Traces and create a SEGY File object. segy_file = SEGYFile() # Set the file wide headers. segy_file.textual_file_header = stream.stats.textual_file_header segy_file.textual_header_encoding = \ textual_header_encoding binary_header = SEGYBinaryFileHeader() this_binary_header = stream.stats.binary_file_header # Loop over all items and if they exists set them. Ignore all other # attributes. for _, item, _ in BINARY_FILE_HEADER_FORMAT: if hasattr(this_binary_header, item): setattr(binary_header, item, getattr(this_binary_header, item)) # Set the data encoding. binary_header.data_sample_format_code = data_encoding segy_file.binary_file_header = binary_header # Add all traces. for trace in stream: new_trace = SEGYTrace() new_trace.data = trace.data # Create empty trace header if none is there. if not hasattr(trace.stats, 'segy'): warnings.warn("CREATING TRACE HEADER") trace.stats.segy = {} trace.stats.segy.trace_header = SEGYTraceHeader(endian=byteorder) elif not hasattr(trace.stats.segy, 'trace_header'): warnings.warn("CREATING TRACE HEADER") trace.stats.segy.trace_header = SEGYTraceHeader() this_trace_header = trace.stats.segy.trace_header new_trace_header = new_trace.header # Again loop over all field of the trace header and if they exists, set # them. Ignore all additional attributes. for _, item, _, _ in TRACE_HEADER_FORMAT: if hasattr(this_trace_header, item): setattr(new_trace_header, item, getattr(this_trace_header, item)) starttime = trace.stats.starttime # Set the date of the Trace if it is not UTCDateTime(0). if starttime == UTCDateTime(0): new_trace.header.year_data_recorded = 0 new_trace.header.day_of_year = 0 new_trace.header.hour_of_day = 0 new_trace.header.minute_of_hour = 0 new_trace.header.second_of_minute = 0 else: new_trace.header.year_data_recorded = starttime.year new_trace.header.day_of_year = starttime.julday new_trace.header.hour_of_day = starttime.hour new_trace.header.minute_of_hour = starttime.minute new_trace.header.second_of_minute = starttime.second # Set the sampling rate. new_trace.header.sample_interval_in_ms_for_this_trace = \ int(trace.stats.delta * 1E6) # Set the data encoding and the endianness. new_trace.data_encoding = data_encoding new_trace.endian = byteorder # Add the trace to the SEGYFile object. segy_file.traces.append(new_trace) # Write the file segy_file.write(filename, data_encoding=data_encoding, endian=byteorder)
def writeSEGY(stream, filename, data_encoding=None, byteorder=None, textual_header_encoding=None, **kwargs): # @UnusedVariable """ Writes a SEG Y file from given ObsPy Stream object. .. warning:: This function should NOT be called directly, it registers via the the :meth:`~obspy.core.stream.Stream.write` method of an ObsPy :class:`~obspy.core.stream.Stream` object, call this instead. :type stream: :class:`~obspy.core.stream.Stream` :param stream: The ObsPy Stream object to write. :type filename: str :param filename: Name of file to write. :type data_encoding: int :param data_encoding: The data encoding is an integer with the following currently supported meaning: ``1`` 4 byte IBM floating points (float32) ``2`` 4 byte Integers (int32) ``3`` 2 byte Integer (int16) ``5`` 4 byte IEEE floating points (float32) The value in the brackets is the necessary dtype of the data. ObsPy will now automatically convert the data because data might change/loose precision during the conversion so the user has to take care of the correct dtype. If it is ``None``, the value of the first Trace will be used for all consecutive Traces. If it is None for the first Trace, 1 (IBM floating point numbers) will be used. Different data encodings for different traces are currently not supported because these will most likely not be readable by other software. :type byteorder: ``'<'``, ``'>'``, or ``None`` :param byteorder: Determines the endianness of the file. Either ``'>'`` for big endian or ``'<'`` for little endian. If is ``None``, it will either be the endianness of the first Trace or if that is also not set, it will be big endian. A mix between little and big endian for the headers and traces is currently not supported. :type textual_header_encoding: ``'EBCDIC'``, ``'ASCII'`` or ``None`` :param textual_header_encoding: The encoding of the textual header. If it is ``None``, the textual_file_header_encoding attribute in the stats.segy dictionary of the first Trace is used and if that is not set, ASCII will be used. This function will automatically set the data encoding field of the binary file header so the user does not need to worry about it. The starttime of every trace is not a required field in the SEG Y specification. If the starttime of a trace is UTCDateTime(0) it will be interpreted as a not-set starttime and no time is written to the trace header. Every other time will be written. SEG Y supports a sample interval from 1 to 65535 microseconds in steps of 1 microsecond. Larger intervals cannot be supported due to the definition of the SEG Y format. Therefore the smallest possible sampling rate is ~ 15.26 Hz. Please keep that in mind. """ # Some sanity checks to catch invalid arguments/keyword arguments. if data_encoding is not None and data_encoding not in VALID_FORMATS: msg = "Invalid data encoding." raise SEGYCoreWritingError(msg) # Figure out the data encoding if it is not set. if data_encoding is None: if hasattr(stream, 'stats') and hasattr(stream.stats, 'data_encoding'): data_encoding = stream.stats.data_encoding if hasattr(stream, 'stats') and hasattr(stream.stats, 'binary_file_header'): data_encoding = \ stream.stats.binary_file_header.data_sample_format_code # Set it to float if it in not given. else: data_encoding = 1 # Create empty file wide headers if they do not exist. if not hasattr(stream, 'stats'): stream.stats = AttribDict() if not hasattr(stream.stats, 'textual_file_header'): stream.stats.textual_file_header = "" if not hasattr(stream.stats, 'binary_file_header'): stream.stats.binary_file_header = SEGYBinaryFileHeader() # Valid dtype for the data encoding. valid_dtype = DATA_SAMPLE_FORMAT_CODE_DTYPE[data_encoding] # Makes sure that the dtype is for every Trace is correct. for trace in stream: # Check the dtype. if trace.data.dtype != valid_dtype: msg = """ The dtype of the data and the chosen data_encoding do not match. You need to manually convert the dtype if you want to use that data_encoding. Please refer to the obspy.segy manual for more details. """.strip() raise SEGYCoreWritingError(msg) # Check the sample interval. if trace.stats.delta > MAX_INTERVAL_IN_SECONDS: msg = """ SEG Y supports a maximum interval of %s seconds in between two samples (trace.stats.delta value). """.strip() msg = msg % MAX_INTERVAL_IN_SECONDS raise SEGYSampleIntervalError(msg) # Figure out endianness and the encoding of the textual file header. if byteorder is None: if hasattr(stream, 'stats') and hasattr(stream.stats, 'endian'): byteorder = stream.stats.endian else: byteorder = '>' # Map the byteorder. byteorder = ENDIAN[byteorder] if textual_header_encoding is None: if hasattr(stream, 'stats') and hasattr( stream.stats, 'textual_file_header_encoding'): textual_header_encoding = \ stream.stats.textual_file_header_encoding else: textual_header_encoding = 'ASCII' # Loop over all Traces and create a SEGY File object. segy_file = SEGYFile() # Set the file wide headers. segy_file.textual_file_header = stream.stats.textual_file_header segy_file.textual_header_encoding = \ textual_header_encoding binary_header = SEGYBinaryFileHeader() this_binary_header = stream.stats.binary_file_header # Loop over all items and if they exists set them. Ignore all other # attributes. for _, item, _ in BINARY_FILE_HEADER_FORMAT: if hasattr(this_binary_header, item): setattr(binary_header, item, getattr(this_binary_header, item)) # Set the data encoding. binary_header.data_sample_format_code = data_encoding segy_file.binary_file_header = binary_header # Add all traces. for trace in stream: new_trace = SEGYTrace() new_trace.data = trace.data # Create empty trace header if none is there. if not hasattr(trace.stats, 'segy'): print "CREATING TRACE HEADER" trace.stats.segy = {} trace.stats.segy.trace_header = SEGYTraceHeader(endian=byteorder) elif not hasattr(trace.stats.segy, 'trace_header'): print "CREATING TRACE HEADER" trace.stats.segy.trace_header = SEGYTraceHeader() this_trace_header = trace.stats.segy.trace_header new_trace_header = new_trace.header # Again loop over all field of the trace header and if they exists, set # them. Ignore all additional attributes. for _, item, _, _ in TRACE_HEADER_FORMAT: if hasattr(this_trace_header, item): setattr(new_trace_header, item, getattr(this_trace_header, item)) starttime = trace.stats.starttime # Set the date of the Trace if it is not UTCDateTime(0). if starttime == UTCDateTime(0): new_trace.header.year_data_recorded = 0 new_trace.header.day_of_year = 0 new_trace.header.hour_of_day = 0 new_trace.header.minute_of_hour = 0 new_trace.header.second_of_minute = 0 else: new_trace.header.year_data_recorded = starttime.year new_trace.header.day_of_year = starttime.julday new_trace.header.hour_of_day = starttime.hour new_trace.header.minute_of_hour = starttime.minute new_trace.header.second_of_minute = starttime.second # Set the sampling rate. new_trace.header.sample_interval_in_ms_for_this_trace = \ int(trace.stats.delta * 1E6) # Set the data encoding and the endianness. new_trace.data_encoding = data_encoding new_trace.endian = byteorder # Add the trace to the SEGYFile object. segy_file.traces.append(new_trace) # Write the file segy_file.write(filename, data_encoding=data_encoding, endian=byteorder)