Releases: radis/radis
0.9.29
(see Pull Request and code changes here)
0.9.29
3rd release of RADIS in 2021. Mainly many small bug/doc fixes. Among the new stuff: new configuration file format, lazy-loading for SpecDatabase, fast apply-slit, and Sphinx Gallery Examples. 🚀
Changes since 0.9.28 (Feb 2021) :
🔬 (Physics) Features
Nothing new in the code. However, we released the first Spectroscopy Tutorials : https://github.com/radis/spectro101 . Have a look, feedbacks are welcome ! Future tutorials are discussed in the Wiki and on 💬 Slack #spectroscopy-tutorials .
⚙️User-interface
- #212 new configuration file format :
~radis.json
replaces~/.radis
. Will allow to use global parameters (for plots, etc.). Auto-conversion of your existing files by @anandxkumar ! - #229 reduce warning/error threshold of too-few-gridpoints-per-line from 5/2 to 3/1
- #201 RADIS works for non-HITRAN molecules (they're not implemented, but you can define your own)
🏃 Performance improvement
- #227 CO2 3-T Treanor calculations up to 5x faster when calling non_eq_spectrum a 2nd time
- #193 #227 fetching of Evib and Erot about ~30% faster for CO2 Treanor, and CO.
- #206 lazy-loading feature by @gugzy when working with a SpecDatabase (i.e., a folder of spectra) : only load them on demand (when using
SpecDatabase.get(Tgas=.., etc.)
). Setting up the database is instantaneous, and generate a summary csv file showing all calculation/experimental conditions ! - #220 Apply_slit is now 10x faster thanks to @CorentinGrimaldi
🐛 Bug fixes
- #81 #214 #217 fix a wavelength range mismatch
- #203 fix direct HITEMP download for nonequilibrium calculations (CO)
- #202 fixed export_lines parameter not working in calc_spectrum
- #111 #191 improve error message with unlabelled CO2 lines of HITEMP, and suggest how to fix it, by @mverleg
- #78 #80 @gagan-aryan fixed errors with missing configuration file parameters,
📝 Documentation
- #225 add more equations in function docstrings, automatically generated from Python code with pytexit
- #221 docs improvements
- #216 fix rendering of radis flow chart by @anandxkumar
- #209 mention HAPI when using Astroquery
- #205 add Sphinx gallery examples (we still need more of them : #136 is open!)
- #207 improve Contributing guide
🏁 Reliability / Tests
nothing here
🏗️ Refactor / change in architecture
nothing neither
0.9.28
0.9.28
2nd release of RADIS in 2021
Changes since 0.9.27 (Jan 2021) :
🔬 (Physics) Features
- #194 automatic download of HITEMP database 🎉 (except CO2 and H2O for the moment)
from astropy import units as u
from radis import calc_spectrum
calc_spectrum(
wavenum_min=1500 / u.cm,
wavenum_max=2400 / u.cm,
molecule="CO",
Tgas=3000,
pressure=1.01325*u.bar,
databank="hitemp", # fetch directly
wstep=0.003, # accuracy
).plot()
⚙️User-interface
- #194 creates a
~/.radisdb
folder for local HDF5 databases automatically generated by RADIS. These files are registered in the~/.radis
config file. - #194 also includes direct loading of HDF5 files (for instance for future ExoMol support)
- #183 add cutoff as calc_spectrum option
Spectrum.get()
returns with the Spectrum calculationwunit
(cm-1 or nm), not always nm by default as before (will be more straightforward wavenumber people !)- add
radis.spectrum.utils.has_nan
util function to test spectra export_lines=False
by default. Saves space, but no LineSurvey by default.
🏃 Performance improvement
- #185 faster rejection of irrelevant files (with metadata implementation), improves loading time for databases with large number of small files (typically CDSD-4000)
- #164 prevent Memory Errors
- #192 Improve computation performance by orders of magnitude when computing line-of-sight with hundreds of slabs
🐛 Bug fixes
📝 Documentation
- lots of tiny doc improvements #178
- reference https://radis.app in the GitHub readme.
🏁 Reliability / Tests
- #194 downloads the HITEMP OH database with fetch_hitemp (small enough - 900 kb - so we don't burn trees by doing that!). Can be used by other tests later.
🏗️ Refactor / change in architecture
Cleaner code :
Not fixed yet in 0.9.28 :
0.9.27
0.9.27
(see pull Request here : #168)
Happy new year ! A new minor version of RADIS was released !
Changes since 0.9.26 (Dec 2020) :
🔬 (Physics) Features
- #160 : HITRAN CO coefficients changed. Validation tests were updated in consequence.
- #152 : update to TIPS-2017 for partition function calculations (no more 3000 K limit for equilibrium calculations)
⚙️User-interface
- #165 add plot_by_parts option to Spectrum.plot()
- 887c3fe replace latex superscripts with unicode (nicer output)
- 9b15768 add legendargs** option for plot_diff
- #170 make compress=True the default (also check compress=2 in store() )
- #173 Line survey appear in Jupyter Notebooks
Not directly related, but the radis-lab online-environment is in beta ! Discussions on 💬 Slack #radis-lab.
🏃 Performance improvement
Not directly related, but radis-benchmark are now available. Contribute on https://github.com/radis/radis-benchmark and on 💬 Slack #benchmark.
🐛 Bug fixes
- #155 fix wildcards not working if used in a list of paths
- #163 fix loading of Spectra generated before 0.9.26
- bebde46 fix SpecDatabase.get_closest() failing
- #169
📝 Documentation
🏁 Reliability / Tests
- Change in pre-commits / isort / black / general improvement of the code : #137, #139, #140, #141 , #142 , #147 , #149
🏗️ Refactor / change in architecture
- #140 moved definitions of HITRAN groups and list of supported molecules in radis/db ( @suzil this will break radis-ui, we will finally need the changes of arunavabasucom/radis-app#112 ). I'll submit a PR after 0.9.27 is released.
Not fixed yet in 0.9.27 :
0.9.22b
0.9.21b
Changes: #40
Minor patch of 0.9.20
Essentially equal to 0.9.20 #39
Major changes:
- Implemented Density Line Mapping (DLM) resulting in ~x100 performance improvement. See detail in #37 , #38
Minor changes:
- possible to use an HITRAN .par file directly in
calc_spectrum(databank='blabla.par')
- updated docs
0.9.23
0.9.26
0.9.26
Featuring new major functionalities such as GPU support for equilibrium calculations (GSoC 2020 project with OpenAstronomy, congrats @pkj-m ), many bug fixes and perf improvement (new DLM model by @dcmvdbekerom , paper soon !), and some neat new interface features (multi-molecules @ulyssed )
👏 7 Contributors : see the pull Request
New to RADIS ?
Welcome ! Let's get started :
1️⃣ Join the community : 🌱 https://radis.github.io/
2️⃣ Read the docs and see the examples
3️⃣ Tweet your best spectra 🐦🐦 https://twitter.com/radis_radiation/status/1333686549083533314 😉
Already a RADIS user ?
👉 Update (public version) :
pip install radis --upgrade
👉 Update (developer version) :
Just pull. See more on the docs.
Main changes since 0.9.25 (Feb 7) :
For file details see the PR at #119
(Physical) Features
- adding initial effective Hamiltonian support directly into Radis : #91
User-interface
- read all datafiles in a folder directly : #107
- dimensioned wmin, wmax parameters directly in calc_spectrum / SpectrumFactory : #96
- added dimensioned units support : #106
- calculate multiple molecules directly #108
- new .min(), .max(), .normalize() methods #122
- new .take() method to return a Spectrum with a single a spectral quantity , useful for chaining, ex :
Spectrum.take('radiance').max()
#133
Performance improvement
- GPU calculations following the Google Summer of Code Project ! Congrats @pkj-m 🎉 #117
- improve .spec loading/writing times #86
- improve DLM accuracy : #115 implemented by #116 #130
Bug fixes
Reliability / Tests
- more automatic tests : #99
- #90 : "Regenerate levels file automatically if manually changed " safety feature to prevent potential bug
- #100
- Test coverage improved from 70% to ~77% but still not the 80% required for #62
Documentation
- new contribution guides : #71, #75
- fix link #132
- new Slack invite
Not fixed yet in 0.9.26 :
For details see #119
0.9.25
0.9.21
0.1.11: Merge pull request #5 from radis/develop
Develop