Merge pull request #623 from xpsi-group/disk_obscuration

Disk obscuration

Author: YinghanMao

CHANGELOG.rst

@@ -1,3 +1,59 @@
+[v3.2] - 2025-12-01
+~~~~~~~~~~~~~~~~~~~
+
+Summary
+^^^^^^^
+
+- The main addition is an optional ray-blocking feature due to an accretion disk, following Ibragimov & Poutanen (2009)(`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__). Additional minor updates are listed below.
+
+
+Fixed
+^^^^^
+
+- The Poisson likelihood with a fixed background was modified so that it can have the correct normalization if the user provides the precomputed ln-data-factorial term (as already done for the default background marginalized likelihoods). In addition, the normalizations of the Gaussian ln-likelihood functions (in the Gaussian extensions) were fixed to have the correct sign. (`#614 <https://github.com/xpsi-group/xpsi/pull/614>`__)
+- Fix issue (`#620 <https://github.com/xpsi-group/xpsi/issues/620>`__) where the interstellar attenuation could be greater than one, although it should not. (`#622 <https://github.com/xpsi-group/xpsi/pull/622>`__)
+- Fix issues (`#534 <https://github.com/xpsi-group/xpsi/issues/534>`__) and (`#574 <https://github.com/xpsi-group/xpsi/issues/574>`__). (`#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+
+
+Added
+^^^^^
+
+- Added details of 3 new published papers and 1 new submitted paper. (`#612 <https://github.com/xpsi-group/xpsi/pull/612>`__)
+- Added the possibility to add custom legend labels in the corner plot. This has become necessary as the multi-modal cornerplot gives default name to new modes. This can be done with a new parameter in the ``CornerPlotter.plot()`` function such as ``legend_labels=['list','of','my','labels']``. (`#619 <https://github.com/xpsi-group/xpsi/pull/619>`__)
+- Now, if the priors are identical (i.e. ``priors_identical=True`` in ``CornerPlotter.plot()``), plot them in black. (`#619 <https://github.com/xpsi-group/xpsi/pull/619>`__)
+- New ``TestRun_Disk_Occultation.py`` and ``TestRun_Disk_Occultation_IQU.py`` have been added which demonstrate the usage of the disk blocking. (`#622 <https://github.com/xpsi-group/xpsi/pull/622>`__)
+
+
+Changed
+^^^^^^^
+
+- In the example modules,``CustomHotregion_Accreting.py`` has been modified to accomodate the linear model from Molkov et al. 2024 as a "user" option, alongside the Bobrikova atmosphere as the "Num5D" option. (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- In the example modules, ``CustomPhotosphere.py`` can now accomodate disk obscuration. (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- The azimuthal integrators (regular, IQU, and split) now feature the ``R_in`` keyword argument for disk blocking of rays due to being obscured by an accretion disk. The math used is taken from Ibragimov & Poutanen (2009), appendix B. `R_in` has a default value of `1e6` (corresponding to 1000 km), at which point the disk is so large that practically speaking no blocking would occur. In that case however, the blocking calculation is skipped, ensuring no extra computation time is introduced here, e.g. for RMPs. (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- ``HotRegion.py`` and ``HotRegions.py`` now feature the ``R_in`` keyword arguments, which are passed to the integrators. Default values are ``1e6``, in which case blocking calculations are skipped. (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- The ``integrator_for_azimuthal_invariance.pyx`` now features new comments that detail how the disk obscuration works as well as other comments that explain in more detail how the intensity calculation works. (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- Updated instrument trimming: replaced the threshold-based method with a tolerance-based one. Each channel now uses a cumulative response computed over the full energy range. Trimming occurs at the first energy where the cumulative responses of all channels exceed the (1 − tolerance) quantile. (`#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+- Changed the structure of ``InstrumentPileup`` in order to simplify it. (`#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+
+
+Removed
+^^^^^^^
+
+- An unused and incorrect Poisson IQU likelihood extension was removed. (`#614 <https://github.com/xpsi-group/xpsi/pull/614>`__)
+- The "threshold" argument for Instrument trimming has been removed and replaced by "tolerance" as the method used for trimming has been changed. (`#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+
+Attribution
+^^^^^^^^^^^
+
+- Yinghan Mao (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- Bas Dorsman (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__)
+- Tuomo Salmi (`#623 <https://github.com/xpsi-group/xpsi/pull/623>`__, `#614 <https://github.com/xpsi-group/xpsi/pull/614>`__, `#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+- Lucien Mauviard (`#622 <https://github.com/xpsi-group/xpsi/pull/622>`__, `#619 <https://github.com/xpsi-group/xpsi/pull/619>`__, `#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+- Anna Watts (`#612 <https://github.com/xpsi-group/xpsi/pull/612>`__)
+- Denis Gonzalez-Caniulef (`#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+- Pierre Stammler (`#588 <https://github.com/xpsi-group/xpsi/pull/588>`__)
+
+
 [v3.1] - 2025-05-23
 ~~~~~~~~~~~~~~~~~~~