The Python support

The Python support#

MTL leverage SWIG, found at swig/swig, to transform C APIs into a binding layer that Python can utilize.

Before using the Python binding, please ensure that MTL is built and the NIC is set up correctly.

1. Installation of SWIG Dependency#

It is recommended to install the SWIG dependency from your operating system’s software repository.

For Ubuntu:

sudo apt-get install swig

For CentOS:

sudo yum install swig

If you encounter issues installing SWIG through your operating system’s package manager, follow the build and installation guide from the SWIG GitHub repository at swig/swig.

Below are the example steps to build the v4.1.1 release. Replace the tag with a newer one if there is a more recent release of SWIG available:

git clone https://github.com/swig/swig.git
cd swig/
git checkout v4.1.1
./autogen.sh
./configure
make
sudo make install

2. Build and install MTL python binding layer#

2.1. Create MTL binding layer code based on swig#

cd $mtl_source_code/python/swig/
swig -python -I/usr/local/include -o pymtl_wrap.c pymtl.i

If you encounter the error pymtl.i:15: Error: Unable to find 'mtl/mtl_api.h', this is typically due to an incorrect include path. Use the following command to locate the correct path for mtl_api.h:

find /usr/ -name mtl_api.h

Once you have obtained the correct path, you can update your SWIG interface file or your build configuration to reference the correct location of mtl_api.h.

2.2. Build#

python3 setup.py build_ext --inplace

2.3. Install#

sudo python3 setup.py install

Checking the log to see the path installed.

creating /usr/local/lib/python3.10/dist-packages/pymtl-0.1-py3.10-linux-x86_64.egg
Extracting pymtl-0.1-py3.10-linux-x86_64.egg to /usr/local/lib/python3.10/dist-packages

Note: remember to rebuild the python binding if any change to the public API of MTL.

3. Run python example code#

Install opencv-python dependency:

# for yuv display
sudo pip3 install opencv-python
# PyAv for video decode/encode
sudo pip3 install av
# for rx_timing_parser.py
sudo pip3 install matplotlib

3.1. st20p_tx.py#

Run the st20p_tx.py, which reads YUV video data from a file and transmits it over the network as a ST2110 ST_FRAME_FMT_YUV422RFC4175PG2BE10 stream.

python3 python/example/st20p_tx.py --p_port 0000:af:01.0 --p_sip 192.168.108.101 --p_tx_ip 239.168.85.20 --tx_url yuv422p10le_1080p.yuv --pipeline_fmt YUV422PLANAR10LE --width 1920 --height 1080 --udp_port 20000 --payload_type 112

3.2. st20p_rx.py#

Execute the st20p_rx.py to receive a ST2110 ST_FRAME_FMT_YUV422RFC4175PG2BE10 stream and display it.

python3 python/example/st20p_rx.py --p_port 0000:af:01.1 --p_sip 192.168.108.102 --p_rx_ip 239.168.85.20 --pipeline_fmt YUV422PLANAR10LE --width 1920 --height 1080 --udp_port 20000 --payload_type 112 --display

3.3. st20p_tx_decode.py#

Use st20p_tx_decode.py to decode YUV video from an encoded file jellyfish-3-mbps-hd-hevc-10bit.mkv and transmit it as a ST2110 ST_FRAME_FMT_YUV422RFC4175PG2BE10 stream across the network.

python3 python/example/st20p_tx_decode.py --p_port 0000:af:01.0 --p_sip 192.168.108.101 --p_tx_ip 239.168.85.20 --tx_url jellyfish-3-mbps-hd-hevc-10bit.mkv --udp_port 20000 --payload_type 112

3.4. st20p_rx_encode.py#

Run the st20p_rx_encode.py to receive a ST2110 ST_FRAME_FMT_YUV422RFC4175PG2BE10 stream and encode it to a .mp4 encoder file.

python3 python/example/st20p_rx_encode.py --p_port 0000:af:01.1 --p_sip 192.168.108.102 --p_rx_ip 239.168.85.20 --rx_url test.mp4 --width 1920 --height 1080 --udp_port 20000 --payload_type 112

3.5. st22p_tx.py#

Run the st22p_tx.py, which reads YUV video data from a file and transmits it over the network as a compressed ST2110-22 stream.

python3 python/example/st22p_tx.py --p_port 0000:af:01.0 --p_sip 192.168.108.101 --p_tx_ip 239.168.85.20 --tx_url yuv422p10le_1080p.yuv --pipeline_fmt YUV422PLANAR10LE --st22_codec jpegxs --width 1920 --height 1080 --udp_port 20000 --payload_type 112

3.6. st22p_rx.py#

Execute the st22p_rx.py to receive a compressed ST2110-22 stream stream and display it.

python3 python/example/st22p_rx.py --p_port 0000:af:01.1 --p_sip 192.168.108.102 --p_rx_ip 239.168.85.20 --pipeline_fmt YUV422PLANAR10LE --st22_codec jpegxs --width 1920 --height 1080 --udp_port 20000 --payload_type 112 --display

3.7. Interlaced#

For TX, interlaced yuv file is used and --interlaced is enabled.

st20p_tx:

python3 python/example/st20p_tx.py --p_port 0000:af:01.0 --p_sip 192.168.108.101 --p_tx_ip 239.168.85.20 --tx_url yuv422p10le_1080i.yuv --pipeline_fmt YUV422PLANAR10LE --width 1920 --height 1080 --udp_port 20000 --payload_type 112 --interlaced

st22p_tx:

python3 python/example/st22p_tx.py --p_port 0000:af:01.0 --p_sip 192.168.108.101 --p_tx_ip 239.168.85.20 --tx_url yuv422p10le_1080i.yuv --pipeline_fmt YUV422PLANAR10LE --st22_codec jpegxs --width 1920 --height 1080 --udp_port 20000 --payload_type 112 --interlaced

For RX, --interlaced is enabled.

st20p_rx:

python3 python/example/st20p_rx.py --p_port 0000:af:01.1 --p_sip 192.168.108.102 --p_rx_ip 239.168.85.20 --pipeline_fmt YUV422PLANAR10LE --width 1920 --height 1080 --udp_port 20000 --payload_type 112 --interlaced --display

st22p_rx:

python3 python/example/st22p_rx.py --p_port 0000:af:01.1 --p_sip 192.168.108.102 --p_rx_ip 239.168.85.20 --pipeline_fmt YUV422PLANAR10LE --st22_codec jpegxs --width 1920 --height 1080 --udp_port 20000 --payload_type 112 --interlaced --display

3.8. rx_timing_parser#

python3 python/example/rx_timing_parser.py --p_port 0000:af:00.1 --ptp --p_sip 192.168.108.102 --p_rx_ip 239.168.85.20 --pipeline_fmt YUV422RFC4175PG2BE10 --width 3840 --height 2160 --udp_port 20000 --payload_type 112

tx:

python3 python/example/st20p_tx.py --p_port 0000:af:00.0 --ptp --width 3840 --height 2160 --pipeline_fmt YUV422RFC4175PG2BE10 --tx_url test_4k.yuv --nb_tx_desc 4096 --udp_port 20000 --payload_type 112 --lcores 91,92