Using OpenVLBI

You can write an application using libopenvlbi by linking against libopenvlbi.so in your gcc command line:

gcc yourapp.c -lopenvlbi -o yourapp

Each instance of vlbi is initiated by vlbi_init():

vlbi_context context = vlbi_init();
 
...
 
// you must be familiar with the dsp_stream_p type, see "DSP API Documentation"
dsp_stream_p stream1 = dsp_stream_new();
dsp_stream_add_dim(stream1, elements_n);
dsp_stream_alloc_buffer(stream1);
 
//dsp_buffer_copy converts numeric types into dsp_t (aka double)
dsp_buffer_copy(input_arr, stream1->in, stream1->len);
 
//Latitude and longitude in degrees, elevation in meters OSL
stream1->location.geographic.lat = Latitude;
stream1->location.geographic.lon = Longitude;
stream1->location.geographic.el = Elevation;
 
stream1->starttimeutc = vlbi_time_string_to_timespec("2018-06-22T02:12.000154874");
 
//assign to this node a friendly, unique name and add it to the context, tell to OpenVLBI to use geographic conversion
vlbi_add_node(context, stream1, "location1_capture", 1);
 
dsp_stream_p stream2 = dsp_stream_new();
 
...
 
vlbi_add_node(context, stream2, "location2_capture", 1);
 
dsp_stream_p stream3 = dsp_stream_new();
...
 
vlbi_add_stream(context, stream3);
...
 
double* target = calloc(sizeof(double), 3);
//RA must be in 24H format
target[0] = RightAscension * 360.0 / 24.0;
target[1] = Declination;
 
double frequency = 60.0e+6;
double samplerate = 100.0e+6;
 
//since we're using geographic coordinates we'd calculate the delay of each baseline
int no_delay_calculation = 0;
 
//telescopes do not change their location during capture, so no location companion stream
int moving_baseline = 0;
 
//obtain the UV plot of the observation and save it as model, 
vlbi_get_uv_plot(context, "obs1_plot_model", 1024, 1024, target, frequency, samplerate, no_delay_calculation, moving_baseline, vlbi_default_delegate);
 
//the default delegate just multiplies each element of the stream by indexing them according to their delay from the farest node to the target
//after this operation a new model with size 1024x1024 containing the plot of the correlation degrees of all baselines will be added into the context
 
//you should add a model with an estimation of the phase to get an inverse fourier transform
vlbi_add_model_from_png(context, "my_phase_estimation.png", "my_phase_estimation");
 
 //Create a new model containing the inverse Fourier transform of your observation as magnitude component and your estimation of the phase
vlbi_get_ifft(context, "ifft_estimation", "obs1_plot_model", "my_phase_estimation");
 
dsp_stream_p image_estimation = vlbi_get_model(context, "ifft_estimation");
 
...
 
dsp_stream_free_buffer(image_estimation);
dsp_stream_free(image_estimation);
vlbi_exit(context);

You can read the API documentation at https://iliaplatone.github.io/OpenVLBI/