Formatted I/O

In addition to passing raw strings, the yggdrasil framework also has support for formatting/processing messages from/to language native objects.

Scalars

The format of messages containing scalar variables (strings, integers, and floating point numbers) can be specified using a C-style format string (See C-Style Format Strings for details). In the example below, the format string "%6s\t%d\t%f\n" indicates that each message will contain a 6 character string, an integer, and a float. In addition to providing a format string, the C API requires the use of different functions for initializing channels and sending/receiving to/from them.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputA')
 6out_channel = YggOutput('outputA')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, msg = in_channel.recv()
14    if not flag:
15        print("Model A: No more input.")
16        break
17
18    # Print received message
19    print('Model A: %s' % msg)
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(msg)
24    if not flag:
25        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputB')
 6out_channel = YggOutput('outputB')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, msg = in_channel.recv()
14    if not flag:
15        print("Model B: No more input.")
16        break
17
18    # Print received message
19    print('Model B: %s' % msg)
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(msg)
24    if not flag:
25        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The same YAML can be used as was used for the example from Getting started with the modification that the files are now read/written line-by-line (filetype: ascii) rather than all at once:

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io1_modelA.py
 5    inputs: inputA
 6    outputs: outputA
 7
 8  - name: python_modelB
 9    language: python
10    args: ./src/formatted_io1_modelB.py
11    inputs: inputB
12    outputs: outputB
13
14connections:
15  - input: outputA  # Connection between model A output & model B input
16    output: inputB
17  - input: ./Input/input.txt  # Connection between file and model A input
18    output: inputA
19    filetype: ascii
20  - input: outputB  # Connection between model B output and file
21    output: ./output.txt
22    filetype: ascii

(Example in other languages)

Tables by Row

Tables can also be passed in a similar manner. For input from a table, the format string does not need to be provided and will be determined by the source of the table. There are different API classes/functions for I/O from/to table channels versus standard channels in each language. (e.g. YggInput vs. YggAsciiTableInput in Python)

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggAsciiTableInput, YggAsciiTableOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggAsciiTableInput('inputA')
 7out_channel = YggAsciiTableOutput('outputA', '%6s\t%d\t%f\n')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, msg = in_channel.recv()
15    if not flag:
16        print("Model A: No more input.")
17        break
18    name, count, size = msg
19
20    # Print received message
21    print('Model A: %s, %d, %f' % (name, count, size))
22
23    # Send output to output channel
24    # If there is an error, the flag will be False
25    flag = out_channel.send(name, count, size)
26    if not flag:
27        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggAsciiTableInput, YggAsciiTableOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggAsciiTableInput('inputB')
 7out_channel = YggAsciiTableOutput('outputB', '%6s\t%d\t%f\n')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, msg = in_channel.recv()
15    if not flag:
16        print("Model B: No more input.")
17        break
18    name, count, size = msg
19
20    # Print received message
21    print('Model B: %s, %d, %f' % (name, count, size))
22
23    # Send output to output channel
24    # If there is an error, the flag will be False
25    flag = out_channel.send(name, count, size)
26    if not flag:
27        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The filetype: table options in the YAML tell the yggdrasil framework that the file should be read/written as a table row-by-row including verification that each row conforms with the table.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io2_modelA.py
 5    inputs: inputA
 6    outputs: outputA
 7
 8  - name: python_modelB
 9    language: python
10    args: ./src/formatted_io2_modelB.py
11    inputs: inputB
12    outputs: outputB
13
14connections:
15  - input: outputA  # Connection between model A output & model B input
16    output: inputB
17  - input: ./Input/input.txt  # Connection between file and model A input
18    output: inputA
19    filetype: table
20  - input: outputB  # Connection between model B output and file
21    output: ./output.txt
22    filetype: table

(Example in other languages)

Tables as Arrays

Tables can also be passed as arrays. In Python and Matlab, this is done using separate classes/functions. In C and C++, this is done by passing 1 to the as_array arguments of the table API classes/functions.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggAsciiArrayInput, YggAsciiArrayOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggAsciiArrayInput('inputA')
 7out_channel = YggAsciiArrayOutput('outputA', '%6s\t%d\t%f\n')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, arr = in_channel.recv_array()
15    if not flag:
16        print("Model A: No more input.")
17        break
18
19    # Print received message
20    print('Model A: (%d rows)' % len(arr))
21    for i in range(len(arr)):
22        print('   %s, %d, %f' % tuple(arr[i]))
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send_array(arr)
27    if not flag:
28        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggAsciiArrayInput, YggAsciiArrayOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggAsciiArrayInput('inputB')
 7out_channel = YggAsciiArrayOutput('outputB', '%6s\t%d\t%f\n')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, arr = in_channel.recv_array()
15    if not flag:
16        print("Model B: No more input.")
17        break
18
19    # Print received message
20    print('Model B: (%d rows)' % len(arr))
21    for i in range(len(arr)):
22        print('   %s, %d, %f' % tuple(arr[i]))
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send_array(arr)
27    if not flag:
28        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML only differs in that as_array: True for the connections to the files to indicate that the files should be interpreted as tables and that the tables should be read/written in their entirety as arrays.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io3_modelA.py
 5    inputs: inputA
 6    outputs:
 7      name: outputA
 8      field_names: name,count,size
 9
10  - name: python_modelB
11    language: python
12    args: ./src/formatted_io3_modelB.py
13    inputs: inputB
14    outputs:
15      name: outputB
16      field_names: name,count,size
17
18connections:
19  - input: outputA  # Connection between model A output & model B input
20    output: inputB
21  - input: ./Input/input.txt  # Connection between file and model A input
22    output: inputA
23    as_array: True
24    filetype: table
25  - input: outputB  # Connection between model B output and file
26    output: ./output.txt
27    as_array: True
28    filetype: table

(Example in other languages)

Tables as Pandas Data Frames

In Python, tables can also be passed as Pandas data frames.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggPandasInput, YggPandasOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggPandasInput('inputA')
 7out_channel = YggPandasOutput('outputA')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, frame = in_channel.recv()
15    if not flag:
16        print("Model A: No more input.")
17        break
18
19    # Print received message
20    nrows = len(frame.index)
21    print('Model A: (%d rows)' % len(frame.index))
22    print(frame)
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send(frame)
27    if not flag:
28        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggPandasInput, YggPandasOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggPandasInput('inputB')
 7out_channel = YggPandasOutput('outputB')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, frame = in_channel.recv()
15    if not flag:
16        print("Model B: No more input.")
17        break
18
19    # Print received message
20    nrows = len(frame.index)
21    print('Model B: (%d rows)' % nrows)
22    print(frame)
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send(frame)
27    if not flag:
28        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML specifies filetype: pandas for the connections to files to indicate that the files should be interpreted as CSV tables using Pandas.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io4_modelA.py
 5    inputs: inputA
 6    outputs:
 7      name: outputA
 8      field_names: name,count,size
 9
10  - name: python_modelB
11    language: python
12    args: ./src/formatted_io4_modelB.py
13    inputs: inputB
14    outputs:
15      name: outputB
16      field_names: name,count,size
17
18connections:
19  - input: outputA  # Connection between model A output & model B input
20    output: inputB
21  - input: ./Input/input.txt  # Connection between file and model A input
22    output: inputA
23    filetype: pandas
24  - input: outputB  # Connection between model B output and file
25    output: ./output.txt
26    filetype: pandas

(Example in other languages)

As Pandas data frames are a Python specific construction, they cannot be used within models written in other languages. However, the files can still be read using Pandas. The data format returned to models on the receiving end of sent Pandas data frames will receive arrays in the proper native data type. In addition, a model written in Python can receive any array sent by another model (whether or not it is Python) as a Pandas data frame.

3D Structures as Ply/Obj

3D structures can be passed around in Ply or Obj formats.

Model Code:

 1import pprint
 2# Import classes for input/output channels
 3from yggdrasil.interface.YggInterface import (
 4    YggPlyInput, YggPlyOutput)
 5
 6# Initialize input/output channels
 7in_channel = YggPlyInput('inputA')
 8out_channel = YggPlyOutput('outputA')
 9
10# Loop until there is no longer input or the queues are closed
11while True:
12
13    # Receive input from input channel
14    # If there is an error, the flag will be False
15    flag, ply = in_channel.recv()
16    if not flag:
17        print("Model A: No more input.")
18        break
19
20    # Print received message
21    print('Model A: (%d verts, %d faces)' % (ply.nvert, ply.nface))
22    pprint.pprint(ply)
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send(ply)
27    if not flag:
28        raise RuntimeError("Model A: Error sending output.")

 1import pprint
 2# Import classes for input/output channels
 3from yggdrasil.interface.YggInterface import (
 4    YggPlyInput, YggPlyOutput)
 5
 6# Initialize input/output channels
 7in_channel = YggPlyInput('inputB')
 8out_channel = YggPlyOutput('outputB')
 9
10# Loop until there is no longer input or the queues are closed
11while True:
12
13    # Receive input from input channel
14    # If there is an error, the flag will be False
15    flag, ply = in_channel.recv()
16    if not flag:
17        print("Model B: No more input.")
18        break
19
20    # Print received message
21    print('Model B: (%d verts, %d faces)' % (ply.nvert, ply.nface))
22    pprint.pprint(ply)
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send(ply)
27    if not flag:
28        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML specifies filetype: ply (filetype: obj for Obj) for the connections to files to indicate that the files should be interpreted as Ply/Obj file formats.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io5_modelA.py
 5    inputs: inputA
 6    outputs: outputA
 7
 8  - name: python_modelB
 9    language: python
10    args: ./src/formatted_io5_modelB.py
11    inputs: inputB
12    outputs: outputB
13
14connections:
15  - input: outputA  # Connection between model A output & model B input
16    output: inputB
17  - input: ./Input/input.ply  # Connection between file and model A input
18    output: inputA
19    filetype: ply
20  - input: outputB  # Connection between model B output and file
21    output: ./output.ply
22    filetype: ply

(Example in other languages)

In Python the data is returned as a dictionary subclass (yggdrasil.serialize.PlySerialize.PlyDict or yggdrasil.serialize.ObjSerialize.ObjDict) while in C/C++ it is returned as a structure (ply_t or obj_t).

Tables as Pandas Data Frames

In Python, tables can also be passed as Pandas data frames.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggPandasInput, YggPandasOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggPandasInput('inputA')
 7out_channel = YggPandasOutput('outputA')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, frame = in_channel.recv()
15    if not flag:
16        print("Model A: No more input.")
17        break
18
19    # Print received message
20    nrows = len(frame.index)
21    print('Model A: (%d rows)' % len(frame.index))
22    print(frame)
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send(frame)
27    if not flag:
28        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import (
 3    YggPandasInput, YggPandasOutput)
 4
 5# Initialize input/output channels
 6in_channel = YggPandasInput('inputB')
 7out_channel = YggPandasOutput('outputB')
 8
 9# Loop until there is no longer input or the queues are closed
10while True:
11
12    # Receive input from input channel
13    # If there is an error, the flag will be False
14    flag, frame = in_channel.recv()
15    if not flag:
16        print("Model B: No more input.")
17        break
18
19    # Print received message
20    nrows = len(frame.index)
21    print('Model B: (%d rows)' % nrows)
22    print(frame)
23
24    # Send output to output channel
25    # If there is an error, the flag will be False
26    flag = out_channel.send(frame)
27    if not flag:
28        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML specifies filetype: pandas for the connections to files to indicate that the files should be interpreted as CSV tables using Pandas.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io4_modelA.py
 5    inputs: inputA
 6    outputs:
 7      name: outputA
 8      field_names: name,count,size
 9
10  - name: python_modelB
11    language: python
12    args: ./src/formatted_io4_modelB.py
13    inputs: inputB
14    outputs:
15      name: outputB
16      field_names: name,count,size
17
18connections:
19  - input: outputA  # Connection between model A output & model B input
20    output: inputB
21  - input: ./Input/input.txt  # Connection between file and model A input
22    output: inputA
23    filetype: pandas
24  - input: outputB  # Connection between model B output and file
25    output: ./output.txt
26    filetype: pandas

(Example in other languages)

As Pandas data frames are a Python specific construction, they cannot be used within models written in other languages. However, the files can still be read using Pandas. The data format returned to models on the receiving end of sent Pandas data frames will receive arrays in the proper native data type. In addition, a model written in Python can receive any array sent by another model (whether or not it is Python) as a Pandas data frame.

Config style Mappings

Simple mapping (key/value) data can be passed around in a colon/tab delimited format.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputA')
 6out_channel = YggOutput('outputA')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, obj = in_channel.recv()
14    if not flag:
15        print("Model A: No more input.")
16        break
17
18    # Print received message
19    print('Model A: %s' % str(obj))
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(obj)
24    if not flag:
25        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputB')
 6out_channel = YggOutput('outputB')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, obj = in_channel.recv()
14    if not flag:
15        print("Model B: No more input.")
16        break
17
18    # Print received message
19    print('Model B: %s' % str(obj))
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(obj)
24    if not flag:
25        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML specifies filetype: map for the connections to files to indicate that the files should be interpreted as config file formats with single, unique keys on each line followed by a delimiter and then a value.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io7_modelA.py
 5    inputs:
 6      - name: inputA
 7        type: object
 8    outputs:
 9      - name: outputA
10        type: object
11
12  - name: python_modelB
13    language: python
14    args: ./src/formatted_io7_modelB.py
15    inputs:
16      - name: inputB
17        type: object
18    outputs:
19      - name: outputB
20        type: object
21
22connections:
23  - input: outputA  # Connection between model A output & model B input
24    output: inputB
25  - input: ./Input/input.txt  # Connection between file and model A input
26    output: inputA
27    filetype: map
28  - input: outputB  # Connection between model B output and file
29    output: ./output.txt
30    filetype: map

(Example in other languages)

In Python the data is returned as a dictionary with one key/value for each line in the file.

JSON Files

More complex nested structures can be passed around using JSON serialization.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputA')
 6out_channel = YggOutput('outputA')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, obj = in_channel.recv()
14    if not flag:
15        print("Model A: No more input.")
16        break
17
18    # Print received message
19    print('Model A: %s' % str(obj))
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(obj)
24    if not flag:
25        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputB')
 6out_channel = YggOutput('outputB')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, obj = in_channel.recv()
14    if not flag:
15        print("Model B: No more input.")
16        break
17
18    # Print received message
19    print('Model B: %s' % str(obj))
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(obj)
24    if not flag:
25        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML specifies filetype: json for the connections to files to indicate that the files should be interpreted as JSON files.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io8_modelA.py
 5    inputs:
 6      - name: inputA
 7        type: object
 8    outputs:
 9      - name: outputA
10        type: object
11
12  - name: python_modelB
13    language: python
14    args: ./src/formatted_io8_modelB.py
15    inputs:
16      - name: inputB
17        type: object
18    outputs:
19      - name: outputB
20        type: object
21
22connections:
23  - input: outputA  # Connection between model A output & model B input
24    output: inputB
25  - input: ./Input/input.txt  # Connection between file and model A input
26    output: inputA
27    filetype: json
28  - input: outputB  # Connection between model B output and file
29    output: ./output.txt
30    filetype: json

(Example in other languages)

In Python, the data is returned in the type determined by the json package.

YAML Files

More complex nested structures can also be represented using the YAML syntax.

Model Code:

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputA')
 6out_channel = YggOutput('outputA')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, obj = in_channel.recv()
14    if not flag:
15        print("Model A: No more input.")
16        break
17
18    # Print received message
19    print('Model A: %s' % str(obj))
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(obj)
24    if not flag:
25        raise RuntimeError("Model A: Error sending output.")

 1# Import classes for input/output channels
 2from yggdrasil.interface.YggInterface import YggInput, YggOutput
 3
 4# Initialize input/output channels
 5in_channel = YggInput('inputB')
 6out_channel = YggOutput('outputB')
 7
 8# Loop until there is no longer input or the queues are closed
 9while True:
10
11    # Receive input from input channel
12    # If there is an error, the flag will be False
13    flag, obj = in_channel.recv()
14    if not flag:
15        print("Model B: No more input.")
16        break
17
18    # Print received message
19    print('Model B: %s' % str(obj))
20
21    # Send output to output channel
22    # If there is an error, the flag will be False
23    flag = out_channel.send(obj)
24    if not flag:
25        raise RuntimeError("Model B: Error sending output.")

(Example in other languages)

The YAML input file options specifies filetype: yaml for the connections to files to indicate that the files should be interpreted as YAML files.

Model YAML:

 1models:
 2  - name: python_modelA
 3    language: python
 4    args: ./src/formatted_io9_modelA.py
 5    inputs:
 6      - name: inputA
 7        type: any
 8    outputs:
 9      - name: outputA
10        type: any
11
12  - name: python_modelB
13    language: python
14    args: ./src/formatted_io9_modelB.py
15    inputs:
16      - name: inputB
17        type: any
18    outputs:
19      - name: outputB
20        type: any
21
22connections:
23  - input: outputA  # Connection between model A output & model B input
24    output: inputB
25  - input: ./Input/input.txt  # Connection between file and model A input
26    output: inputA
27    filetype: yaml
28  - input: outputB  # Connection between model B output and file
29    output: ./output.txt
30    filetype: yaml

(Example in other languages)

In Python, the data is returned in the type determined by the PyYAML package.