Re: [weewx-user] Announcing weewx-skyfield (a modern almanac), weewx-loopdata (live report tags), and weewx-celestial (a live sky display)

['John Kline' via weewx-user]

Good news on that test failure: it's not related to your skin problem, and it's not your machine computing anything wrong.  That test compares my moon magnitude against the locally installed PyEphem, and PyEphem's lunar magnitude model was quite crude before version 4.2 — yours (likely Debian's python3-ephem 4.1.x) says -10.9, modern PyEphem 4.2 says -8.4, agreeing with my value.  I'll make that test tolerant of older PyEphem versions.    The 427 tests that *passed* are actually the interesting part: they show the Python environment you ran them in computes these durations correctly.  That makes the answers to the questions in my previous comment — especially `weectl extension list` and which Python environment weewxd itself runs in —  the most likely place the answer is hiding. I answered your first post here on the github issue you filed, but repeating here so others can follow.  The leading hypothesis is that you also have weewx-skyfield-almanac installed; but I’m awaiting your answers to these questions. Thanks for the report — and for the link to your live page, which helped a lot. Here is what I can see from your page: everything is being computed correctly except the durations. Rise, set and transit times, altitudes, azimuths, the constellations, even the widths of the bars in the day-timeline panel (which are drawn from the same rise/set times) are all right. Only the durations derived from the almanac's visible tag — the Daylight line, the "Up for" column, and the timeline tooltips — show 0h 00m. I have not been able to reproduce it. I ran the v1.13 code at your coordinates for July 28/29 under every Skyfield release the extension supports (1.47 through 1.54, including Debian trixie's Skyfield 1.53 + NumPy 2.2 combination), and visible comes out correct in every case (about 14h56m for the sun). So the difference must be something about your system, and I'd like to ask for a few details: WeeWX version and Python version, and how WeeWX and Skyfield were installed (Debian apt packages, or pip into a virtual environment)? The output of python3 -c "import skyfield, numpy; print(skyfield.VERSION, numpy.__version__)" using the same Python that runs WeeWX (if pip-installed, activate the venv first). The output of weectl extension list — in particular, is weewx-skyfield-almanac or weewx-celestial also installed? Another extension that registers its own almanac could be the one actually answering these tags. The weewxd startup log lines that mention Skyfield, e.g. sudo journalctl -u weewx | grep -i skyfield (the extension logs its own version and the Skyfield version at startup). Which files you changed or added for your French translation — did you add a lang/fr.conf, or edit the template or wxskyfield_sky.py directly? If you're up for one more test, temporarily add this line anywhere in skins/Skyfield/index.html.tmpl and report what it prints: <p>$almanac.sun.visible.raw</p> That is the raw number of seconds behind the Daylight line; today it should print about 53800. One request in return: the next release will ship a French translation of the Sky page and of the almanac's body and constellation names (marked Beta — the words await native-speaker review). Since you know the material and have already translated the page for yourself, would you be willing to look it over? I've attached the proposed lang/fr.conf (you will need to unzip it); corrections from a native speaker would be very welcome, and it should also spare you maintaining your own translation going forward. Thanks, John On Jul 29, 2026, at 9:40 AM, Jacques Terrettaz <jterret...@gmail.com> wrote: John, I don't know if it is related to my problems with the skin, , but I have a failed comparison when running your test scripts : ====================================================================================================== test session starts ======================================================================================================= platform linux -- Python 3.11.2, pytest-9.1.1, pluggy-1.6.0 rootdir: /home/meteo/weewx-skyfield-main collected 428 items                                                                                                                                                                                                               tests/test_almanac.py .........................................................F.......................................................................................................................................... [ 45%] ...........................................................................................................................................................................................                                [ 89%] tests/test_sky_page.py .............................................                                                                                                                                                       [100%] ============================================================================================================ FAILURES ============================================================================================================ ___________________________________________________________________________________ TestPhysicalAttributes.test_magnitudes_agree_with_pyephem ____________________________________________________________________________________ self = <test_almanac.TestPhysicalAttributes object at 0x7fff290fb0d0>, almanac = <weewx.almanac.Almanac object at 0x7fff28c7cd10>     def test_magnitudes_agree_with_pyephem(self, almanac):         ephem = pytest.importorskip('ephem')         observer = pyephem_observer()         for planet in ('mercury', 'venus', 'mars', 'jupiter', 'saturn', 'uranus', 'neptune',                        'sun', 'moon', 'pluto'):             body = getattr(ephem, planet.title())()             body.compute(observer)             # PyEphem uses older magnitude models, so agreement is loose. >           assert getattr(almanac, planet).mag == pytest.approx(body.mag, abs=0.6), planet E           AssertionError: moon E           assert np.float64(-8.420329067365575) == -10.9 ± 0.6 E              E             comparison failed E             Obtained: -8.420329067365575 E             Expected: -10.9 ± 0.6 tests/test_almanac.py:656: AssertionError ======================================================================================================== warnings summary ======================================================================================================== -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/weewx-user/C72B29DD-F726-4F21-87C8-7E55EF6A63B2%40gmail.com. -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/weewx-user/A3457398-1DFB-4604-8544-8DA1045326C8%40johnkline.com.