itu_predict

If you have ever used VOACAP (Very good, and very useful still today), then I hope you have also found the most excellent site called Proppy which lives at https://soundbytes.asia/proppy/planner.

If you want to have a look at / try / is this for me ... this is the

Easiest to use
Least confusing
Best looking
And least expensive (it is 100% free)

I really - really like Proppy, I think it is a very good example of Open-Source with a good front end.

But... I was curious as to how it was put together.....

How to run a Proppy like service at home

The Web service Proppy consists of 2 main objects

ITU Prediction Components
A Web Front End

I am only going to consider the ITU Prediction Components, as there are many places/ways to wrap 'back-end' components in a Web interface.

Basic Building blocks

The start point should be ITU R Github Source This is a git repo which has the main code. You will need to compile this (it is quite standard) to produce an executable called ITUHFProp.

As I was building this on a Mac (there are no default Mac build instructions) I created a Mac Makefile, which looks like this

prefix ?= /usr/local
export prefix

mandir = $(prefix)/share/man
man1dir = $(mandir)/man1

TOPTARGETS := all clean

SUBDIRS := ../P533/Mac/ \
        ../P372/Mac/ \
        ../ITURHFProp/Mac/

datarootdir = $(prefix)/share
datadir = $(datarootdir)

$(TOPTARGETS): $(SUBDIRS)
$(SUBDIRS):
        $(MAKE) -C $@ $(MAKECMDGOALS)

.PHONY: $(TOPTARGETS) $(SUBDIRS)

install: $(SUBDIRS)
        install -d $(DESTDIR)$(man1dir)
        install ITURHFProp.1 $(DESTDIR)$(man1dir)
        install -d $(DESTDIR)$(datadir)/iturhfprop/data
        install -m 744 -D ../P372/Data/*.txt $(DESTDIR)$(datadir)/iturhfprop/data
        install -m 744 -D ../P372/Data/*.bin $(DESTDIR)$(datadir)/iturhfprop/data

It should just be a case of

 make -j 4
 make install

Sadly at this point - you will need to start findind some reference data files, this actually had me confused for some time - until I carefully ready the following link ITURHFProp SW Projects .

After downloading the required files, and placing them in the correct directory.I was able to use run the ITUHFProp....

But now I need a config/plan for it too execute.

Make/Build/Copy a plan

The easiest way to get an example of a plan, is to go back to the https://soundbytes.asia/proppy/planner

- Select P2P 
- Source Text
  - Set to On 
- Pick your TX/RX etc
- Press Run Prediction.

In the lower half of the screen - Input File is the file format that you need.

So copy that and put it into a file say To_US.txt

Then to run your own local prediction

 /usr/local/bin/ituhfprop To_US.txt To_US.out

View the Output

The Output file (To_US.out) will look something like this

---------------------------------------------------------------------------
 International Telecommunications Union - Radiocommunication Sector (ITU-R)
     ITURHFProp         Ver Jun 11 2022
     HF Model (P533)    Ver 14.2
     Noise Model (P372) Ver 14.3
     Analysis Prepared  Sun Sep 11 01:18:42 2022

---------------------------------------------------------------------------

***************************** P533 Input Parameters ****************************

    Proppy Online HF Circuit Prediction: Point-to-Point
    Year          : 2022
    Month         : September
    Hour          : 1 (hour UTC)
    SSN (R12)     : 109
    Distance      : 13159.952929 (km)
    dmax          : 4000.000000 (km)
    Tx power      : 0.000000
    Tx Location     Transmitter
.
.
.
. Data Omitted
.
.
************************ Calculated Parameters ****************************

09, 01,    2.000,  16.63,-733.70,-616.43,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215, -10.5365
09, 01,    3.000,  16.63,-446.87,-325.36,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -5.4264
09, 01,    4.000,  16.63,-317.88,-193.56,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -3.1121
09, 01,    5.000,  16.63,-249.86,-123.27,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -1.8810
09, 01,    6.000,  16.63,-210.36, -81.61,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -1.1578
09, 01,    7.000,  16.63,-185.98, -55.14,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -0.7053
09, 01,    8.000,  16.63,-170.42, -37.56,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -0.4110
09, 01,    9.000,  16.63,-160.37, -25.66,   0.00,  56.8194,   3.7053,  -0.1700,  16.6319,  11.0935,   1.5109,   1.5931,   2.1600,   2.1600,   0.9215,  -0.2162
.
.
.Lots more Lines
.
.

If this is what you see - Great... If there is no output - check the Data File and how you ran the command.

Interpreting the ITUHFProp output

As unpleasant as it looks - this output file is easy to process. Here I will show you how you can accomplish this - but first let us understand what is in this output file.

The Input File (To_US.txt) was a Point to Point format data file, in which we asked to compute the Percentage Path Reliability from a TX Site, to an RX Site, using a specified Power, Antenna (more on this later), for ALL HF Frequencies in 1 Mhz steps, for a Specific Month, and Per Hour !! Phew ... sorry that was so long winded.

We will read the output file using Python.... Like this

import pandas as pd
fname='To_Us.out'
with open(fname, "rt") as ifp:
            data = ifp.read().split("\n")

        rec = []
        rec.append(
            [
                "Month",
                "Hour",
                "Freq",
                "Op_MUF",
                "rx_db",
                "snr_db",
                "rel",
                "fieldstrength_db",
                "focusstrength_db",
                "loss_db",
                "upper_ref_freq",
                "lower_ref_freq",
                "corr_tx",
                "corr_rx",
                "tx_gain",
                "rx_gain",
                "gyro_freq",
                "scale_factor",
            ]
        )
        for l in data:
            parts = l.split(",")
            if len(parts) == 18:
                rec.append(parts)

        df = pd.DataFrame.from_records(rec[1:], columns=rec[0])
        df = df.apply(pd.to_numeric)

We now have a DataFrame with just calculated data values.

Let us plot this and make something visual, rather than trying to look at tabular output.

Draw Heat Map of Point to Point output

Using the DataFrame (a variable called df) we can use a small method like this

import plotly.express as px
def draw_contest_heat(ctry: str, df: pd.DataFrame):
        df_heat = (
            df.pivot(index="Freq", columns="Hour", values="rel")
            .reset_index()
            .set_index("Freq")
        )
    // Limiting the Frequencies to Contest Bands 
    // WARC not included
        contest = [7, 14, 21, 28]
        df_contest = df_heat.filter(items=contest, axis=0)
        fig = px.imshow(
            df_contest,
            title=f"{ctry}",
            labels=dict(x="'Z' Hour", y="Freq (Mhz)", color="Reliabilty"),
            x=[f"{h}" for h in range(0, 24)],
            y=[f"{x}Mhz" for x in contest],
            zmin=0.0,
            zmax=100.0,
            aspect="auto",
            color_continuous_scale=["white", "grey", "blue", "green", "yellow",
"red"],
            text_auto=True,
        )
        fig.update_xaxes(side="top")
        return fig

f=draw_contest_heat('To_US',df)
f.write_image(To_US.png")

This will produce something like

/images/TO_VK.png

Which is a good starting point to know when is a good time to listen for VK stations - clearly showing that 10m/28Mhz is not looking promising.

Further Steps

There are several other types of output you can generate (and then load, and produce a graph for) each one is no more complex that the steps I have just outlined.

Many thanks to the most excellent Proppy site, and it's author James Watson M0DNS.