================= PyEphem CHANGELOG ================= Version 4.1.3 (2021 December 13) -------------------------------- - Fixed an inadvertent loss of precision in the routine that computes a date’s hours, minutes, and seconds. It was sometimes returning a small negative number of seconds, which caused Python’s ``datetime`` type to complain ``ValueError: second must be in 0..59``. - Users installing from source on Windows are now protected against a possible encoding error as ``setup.py`` reads in two text files. Version 4.1.2 (2021 December 5) ------------------------------- - Fixed the new rising and setting routines so they properly detect if a body is always below the horizon and raise a ``NeverUpError`` instead of a plain ``ValueError``. - Gave bodies a new ``ha`` Hour Angle attribute, since the quantity was computed internally but then discarded. - Renamed the observer attribute ``temp`` to ``temperature``, leaving an alias behind to support the old spelling. Version 4.1.1 (2021 November 27) -------------------------------- - When you provide PyEphem with a Python ``datetime`` that has a time zone attached, PyEphem now detects the time zone and converts the date and time to UTC automatically. - A new search routine had been written and tested to power the :ref:`transit-rising-setting` methods ``previous_rising()``, ``previous_setting()``, ``next_rising()``, and ``next_setting()``. They should no longer be susceptible to getting hung up in a loop. You should also find them substantially faster. - Fixed the ``constellation()`` routine so that it uses astrometric, rather than apparent, right ascension and declination. This should make it more accurate along the borders of each constellation. - Fixed how the underlying “libastro” library computes whether a body’s image is deflected by gravity when its light passes close to the Sun. Previously, users would see coordinates jump unexpectedly as the deflection formula turned on and off haphazardly. - Fixed the star positions in the ``ephem.stars`` star catalog by adding 8.75 years of proper motion to each star. Previously, each 1991.25 position from the Hipparcos catalog was being incorrectly treated as a 2000.0 position. - A new routine ``unrefract()`` lets you compute the real altitude of a body that you observe in the sky at a given apparent altitude. - The old ``cities.lookup()`` function is now officially deprecated. Because of a Google API restriction, it stopped working in 2018. Version 4.1 (2021 September 29) ------------------------------- - Planetary moon positions are now available through the year 2040 (previously, asking for a position past 2020 returned zeroes). - The ``Date`` object is improved such that the return values of ``str(d)``, ``d.datetime()``, and ``d.tuple()`` always agree and are always rounded to the nearest microsecond. - Earth satellites offer new orbit parameter attributes. The old names (which are still present, but no longer documented) started awkwardly with underscores, and had inconsistent getters and setters that would change the satellite if you attempted a round trip like ``sat._inc = sat._inc``. The new attributes (see the list in the Quick Reference) have simple names and use the same units when getting and setting. - Updated Pluto’s long-term orbital elements to match the Astronomical Almanac 2020. - Expanded the ∆T table so that it now runs through 2018, with data from the Astronomical Almanac 2020. - The ``Observer.copy()`` method is now documented, and (after a user requested it) the class also now works with Python’s ``copy`` module. - PyEphem should now be able to compile for pypy3. Version 4.0.0.2 (2020 June 14) ------------------------------ - Restore PyEphem’s undocumented ability to parse angle strings like ``'12 34 56'`` that are only separated with spaces, instead of insisting on ``'12:34:56'`` for 12 degrees, 34 arcminutes, and 56 arcseconds. - Fix a compile error `‘for’ loop initial declarations are only allowed in C99 mode` reported from a user on Oracle Linux. Version 4.0.0.1 (2020 June 12) ------------------------------ - Fix ``MANIFEST.in`` so the ``.tar.gz`` source distribution includes all the header files necessary for compilation. Version 4.0.0 (2020 June 12) ---------------------------- - Upgraded to the MIT license following Elwood Downey’s generous decision to open-source XEphem’s code. - Fix a bug where supplying a string with a decimal degree measurement could send the parser into an infinite loop. - The ``FixedBody`` constructor, which accepts no arguments, now correctly raises a ``TypeError`` if any are supplied. Version 3.7.7.1 (2020 February 22) ---------------------------------- - Fixed the body ``copy()`` method to correctly copy the extra attributes that some bodies have beyond those of a normal body, like the catalog number of an Earth satellite. This bug had in some cases caused segmentation faults. - GitHub issue #166: Fixed a memory leak in ``readdb()``. - GitHub issue #119: Fixed the ``Body.copy()`` method to correctly copy object-specific fields across to the new object, like Saturn ring tilt and Earth satellite catalog number. Version 3.7.7.0 (2019 August 18) -------------------------------- - Upgraded libastro to 3.7.7. - The internal star catalog now includes all 57 navigational stars. - GitHub issue #63: The rise, culminate, and set returned by ``next_pass()`` are now consecutive values for a single pass. Pass ``singlepass=False`` to return the original next_rise, next_culminate, next_set even if next_set < next_rise (the satellite is already up). - GitHub issue #141: ``ephem.delta_t('0')`` now returns the correct value, instead of misbehaving for that particular input. Version 3.7.6.0 (2015 August 19) -------------------------------- - The new ``ephem.cities.lookup()`` function runs a Google geocoding search and returns an ``Observer`` object for the top result. - When an Earth satellite position cannot be computed, PyEphem now raises an exception instead of freezing and locking up Python. - Upgraded to the ``libastro`` from XEphem 3.7.6, bringing improvements to Earth satellite transit calculations. - GitHub issue #76: Earth satellite velocity is now calculated with greater accuracy. - GitHub issue #64: rising and setting routines are now careful to restore your ``Observer.date`` even if they die with an exception. - GitHub issue #56: Earth satellites now raise an exception for dates a year or more from their TLE epoch, because ``libastro`` refuses to process old elements and would return nonsense coordinates. - GitHub issue #44: a segmentation fault would eventually kill Python 3 if a script called ``Body.copy()`` either directly or via the Standard Library ``copy.copy()`` function. Version 3.7.5.3 (2014 May 29) ----------------------------- - Gave all bodies a ``.parallactic_angle()`` method that computes the same angle as the ``PA`` column in XEphem itself (GitHub #24). - Added a ``.long`` alias for the ``.lon`` longitude attribute on ecliptic and galactic coordinates (GitHub #41). - Combined the Python 2 and Python 3 code bases using the magic of ``#ifdef`` and a barrel full of ``print()`` parentheses, which should prevent either version from ever falling behind the other again. - Fixed GitHub issues #35, #37, #40. Version 3.7.5.2 (2013 December 21) ---------------------------------- - The ``separation()`` function will no longer allow hardware floating point rounding errors to produce a non-zero result when a position is compared to itself, nor return a ``NaN`` result (which one user reports seeing as the angle ``1389660529:33:00.8`` degrees). `(GitHub #31) `_ - PyEphem routines no longer ignore the microseconds of ``datetime`` objects provided as input. `(GitHub #29) `_ - PyEphem is now more careful to raise an exception if angles are specified using strings that contain invalid characters. - The Earth-satellite attributes ``ra`` and ``dec`` are now correctly referenced to the epoch-of-date, instead of being expressed in J2000 like the astrometric attributes. Version 3.7.5.1 (2011 November 24) ---------------------------------- - Upgraded the underlying astronomy library to 3.7.5. - **Incompatible Change**: the transit functions are now symmetric with the rising and setting functions: while they still return the date and time of the event, they do *not* alter the ``.date`` attribute of the Observer which gets passed to them. This brings their behavior into line with the documentation. `(Launchpad #861526) `_ - ``Date('1986-2-9')`` now means February 9th instead of meaning “the beginning of 1986, minus two months, minus nine days.” `(Launchpad #792321) `_ - Earth satellite positions are now computed to six additional digits, in an attempt to eliminate small jumps in position that some users were observing in their figures. `(Launchpad #812906) `_ - Coordinate pair creation no longer leaks memory. `(Launchpad #798155) `_ Version 3.7.4.1 (2011 January 5) --------------------------------- - Renamed the ``Observer.long`` attribute to ``lon`` after realizing that the official syllabification of “longitude” is “lon·gi·tude.” Also changed ``Body`` objects so that ``hlong`` is ``hlon`` instead. The old names will always be supported for compatibility with older programs. - Upgraded the underlying astronomy library to 3.7.4. - **Bugfix:** repaired the ``separation()`` function so that it no longer leaks memory; thanks to Enno Middelburg for the bug report! - **Bugfix:** completely rebuilt the geographic data used by ``city()`` after Giacomo Boffi pointed out several errors. Version 3.7.3.4 (2009 April 30) ------------------------------- - Added a new ``next_pass()`` method to ``Observer`` that searches for when a satellite next rises, culminates, and sets. - Added a ``compute_pressure()`` method to ``Observer`` which computes the standard atmospheric pressure at the observer's current elevation. This function now gets called automatically on new ``city()`` objects before they are returned to the user. - Corrected the altitude of San Francisco as returned by ``city()``. - Improved the copyright message so that two more authors are credited. Version 3.7.3.3 (2008 October 3) -------------------------------- - Added ``cmsI`` and ``cmsII`` attributes to ``Jupiter`` to provide the central meridian longitude in both System I and System II. - **Bugfix**: Saturn was returning the wrong values for its earthward and sunward angle tilt. Version 3.7.3.2 (2008 July 2) ----------------------------- - **Bugfix**: the rising and setting functions, if called repeatedly, would sometimes get hung up on a single answer which they would return over and over again instead of progressing to the next rising or setting. They should now always progress instead of getting stuck. Version 3.7.3.1 (2008 July 1) ----------------------------- - **Bugfix**: the rising and setting functions were attempting to achieve such high precision that users sometimes found circumstances under which they would not complete at all! They now stop and return an answer once they are withing a half-second of the real time of rising, transit, or setting, which solves the problem without damaging the quality of the results when tested against the Naval Observatory. - Upgraded to the libastro from XEphem 3.7.3. Version 3.7.2.4 (2008 June 12) ------------------------------ - **Incompatible Change**: After feedback from users, I have changed the ``Observer`` methods which find risings, settings, and transits, so that they do not change their Observer's ``.date`` attribute. So the sequence: .. code-block:: python r1 = boston.next_rising(mars) r2 = boston.next_rising(mars) now computes the same value twice! If you want a series of calls to each begin when the other left off, you can use the ``start=`` parameter described in the next item: .. code-block:: python r1 = boston.next_rising(mars) r2 = boston.next_rising(mars, start=r1) - Added an optional ``start=`` argument to the rising, setting, and transit ``Observer`` functions, that tells them from which date and time to begin their search. - **Bugfix**: Rewrote planetary moon routines so that moons of Mars, Jupiter, Saturn, and Uranus now return appropriate data for years 1999-2020. (Each moon had been returning the unmodified position of its planet, because I was unsure whether I could distribute the moon data with PyEphem.) - You can no longer create arbitrary attributes on an ``Observer``, to prevent users from accidentially saying things like ``here.longitude`` or ``here.lon`` when they mean ``here.long``. Create your own subclass of ``Observer`` if you need the power to set your own attributes. - The ephem module now provides a ``__version__`` symbol. - Added test suite that tests planet and planet moon positions against JPL ephemeris data (needs more work). Version 3.7.2.3 (2008 January 8) -------------------------------- - Three new classes ``Equatorial``, ``Ecliptic``, and ``Galactic`` allow coordinates to be transformed between the three systems (ability to transform coordinates was requested by Aaron Parsons). - Added constants for popular epochs ``B1900``, ``B1950``, and ``J2000``. - Added named functions for every solstice and equinox (before, only the vernal equinox could be asked for specifically). - Product tests have been moved inside of the ``ephem`` module itself. - **Bugfix**: ``Angle()`` can no longer be directly instantiated. - **Bugfix**: San Francisco had the wrong coordinates in the cities database (pointed out by Randolph Bentson). Version 3.7.2.2 (2007 December 9) --------------------------------- - The phases of the moon can now be determined through the functions ``next_new_moon()``, ``next_full_moon()``, ``previous_new_moon()``, et cetera. - Added a modest database of world cities; the ``city()`` function returns a new Observer on each call: .. code-block:: python observer = ephem.city('Boston') - Using the old ``rise``, ``set``, and ``transit`` attributes on ``Body`` objects now causes a deprecation warning. - **Bugfix**: the last release of PyEphem omitted the constants ``meters_per_au``, ``earth_radius``, ``moon_radius``, and ``sun_radius``; the constant ``c`` (the speed of light) is also now available. Version 3.7.2.1 (2007 October 1) -------------------------------- - Functions now exist to find equinoxes and solstices. - Bodies now cleanly offer three different versions of their position, rather than making the user remember obscure rules for having each of these three values computed: * Astrometric geocetric right ascension and declination * Apparent geocentric right ascension and declination * Apparent topocentric right ascension and declination - Bodies can now find their next or previous times of transit, anti-transit, rising, and setting. - A ``localtime()`` function can convert PyEphem ``Date`` objects to local time. - Now ``ephem.angle`` instances can survive unary ``+`` and ``-`` without getting changed into plain floats. - The ``elev`` Observer attribute has been renamed to ``elevation``. - Observers now display useful information when printed. - Added a much more extensive test suite, which, among other things, now compares results with the United States Naval Observatory, insisting upon arcsecond agreement. - **Bugfix**: When a fixed body is repeatedly precessed to different dates, its original position will no longer accumulate error. Version 3.7.2a (2007 June) -------------------------- - Upgraded to the libastro from XEphem 3.7.2. - Should now compile under Windows! - **Bugfix**: rewrote date-and-time parsing to avoid the use of ``sscanf()``, which was breaking under Windows and requiring the insertion of a leading space to succeed. - Improved the error returned when a date string cannot be parsed, so that it now quotes the objectionable string (so you can tell which of several date strings on the same line gave an error!). Version 3.7b (2005 August 25) ------------------------------ - **Bugfix**: in the underlying library, earth satellite objects do not support ``SOLSYS`` attributes like ``sun_distance``; so ``EarthSatellite`` must inherit from ``Body`` rather than ``Planet`` (and lose several attributes, which were returning nonsense values). Version 3.7a (2005 August 22) ------------------------------ - Upgraded to the libastro from XEphem 3.7. - **Bugfix**: after creating an earth satellite and calling ``compute()``, some attributes (including ``sublat`` and ``sublong``) would always equal zero until you had accessed a more mainstream attribute (like ``ra`` or ``dec``); now, all attributes should return correct values on their first access. - **Bugfix**: the ``sidereal_time()`` function of an ``Observer`` now returns a correct floating-point number that measures in radians, rather than a number in the range [0,1). - The ``Observer`` now has an ``radec_of(az=, alt=)`` function that returns the right ascension and declination of a point in the sky. - You can normalize an ``Angle`` into the range [0,2pi) by requesting the attribute ``.norm``. - Earth satellite objects read in from TLE files now retain their TLE catalog number as an attribute ``catalog_number``. - Uninitialized bodies now start off with ``None`` for their name, rather than the string ``"unnamed"``. Version 3.6.4a (2005 July 18) ------------------------------ - Upgraded to the libastro from XEphem 3.6.4, which: * No longer incorrectly applies relativistic deflection to objects on this side of the Sun, whose light will obviously not go past the sun and be deflected. * Now correctly handles earth satellites with a negative ``es_decay`` parameter. - Added several functions to the module: * ``moon_phases()`` computes a new and full moon following a date. * ``delta_t()`` computes the difference between Terrestrial Time and Universal Time. * ``julian_date()`` computes the Julian Date for a ``date`` or ``Observer``. * ``millennium_atlas()`` and ``uranometria()`` and ``uranometria2000()`` determine the star atlas page on which a given location falls, given as right ascension and declination. - Added a function to the Observer class, which takes no arguments: ``sidereal_time()`` computes the sidereal time for the Observer - Each ``Observer`` now has a ``horizon`` attribute, with which you can specify the degrees altitude at which you define an object to be rising or setting. Normally, all rising and setting times are computed for when the object appears to be exactly at the horizon (at zero degrees altitude). Version 3.6.1a (2004 November 25) ---------------------------------- - All major moons in the solar system are now supported. - Added ``copy()`` method to bodies, that returns a new instance of the body which should be identical in all properties. - Improved the definitions of body attributes, both in their docstrings and in the PyEphem Manual. - Improved access to the orbital parameters by which the user defines bodies in ellipical, parabolic, and hyperbolic orbits, as well as artificial Earth satellites; users can now create such objects entirely through setting their parameters, without having to use the ``readdb()`` function to parse a definition of the object in Ephem database format. - Source distribution now includes test suites, one of which actually checks to see whether your version of PyEphem produces the same output as the examples from the PyEphem Manual (two of which will fail). - Following the same adjustment in the XEphem application, PyEphem now uses a default atmospheric pressure of 1010 millibar, rather than the old value of 1013, when computing the altitude of a body near the horizon. - The ``constellation()`` function now correctly forces the computation of a body's ``ra`` and ``dec`` before determining the constellation in which the body lies. - Code should produce cleaner compiles on many platforms. Early History ------------- - **27 Jul 1998**: the original PyEphem, an awkward SWIG wrapper around the raw ``libastro`` C structures and functions, was ready to appear on the “Contributed.html” page on the Python web site. (With an apology from the python.org webmaster, it was not actually posted until 1998 August 26.)