cigale issueshttps://gitlab.lam.fr/groups/cigale/-/issues2023-10-04T12:44:13Zhttps://gitlab.lam.fr/cigale/cigale/-/issues/24Too many open files2023-10-04T12:44:13ZLAMBERT Jean-charlesToo many open files### CIGALE version: 2022.0 — Python version: 3.10.3 — Platform: linux-x86_64
When running **pcigale** with many **blocks** and **cpu**, the program stop after several blocks (depend of pcigale.ini parameters) with the unix message "Too ...### CIGALE version: 2022.0 — Python version: 3.10.3 — Platform: linux-x86_64
When running **pcigale** with many **blocks** and **cpu**, the program stop after several blocks (depend of pcigale.ini parameters) with the unix message "Too many open files".
If I increase the limit (ex: ulimit -n 131000) in the terminal before running pcigale, it completes with no issue.
After investigating during a running pcigale run, I can see that **/proc/$PID/fd** folder is constantly increasing slightly. It looks like that some shared memory segments **/dev/shm/pym-xxxxxx** might not be deleted.
($PID is the process id of the running pcigale process)https://gitlab.lam.fr/cigale/cigale/-/issues/23Make nebular module ajustable to different SSP / IMF2023-03-20T14:46:22ZYannick Roehllyyannick.roehlly@lam.frMake nebular module ajustable to different SSP / IMFCurrently the nebular module uses a single “nebular model” that was build using a specific SSP (now BC03) and a given IMF (now Chabrier). The problem is that we may use this module with another IMF and / or another SSP. For instance, to ...Currently the nebular module uses a single “nebular model” that was build using a specific SSP (now BC03) and a given IMF (now Chabrier). The problem is that we may use this module with another IMF and / or another SSP. For instance, to use BPASS we need to replace the nebular models and rebuild the database.
The nebular module should be able to handle different models for different SSP and IMF and use the right one during the run by inspecting what was used by the stellar emission module.https://gitlab.lam.fr/cigale/cigale/-/issues/22pCIGALE incompatible with numpy==1.182020-04-07T12:09:56ZAlexandre BeelenpCIGALE incompatible with numpy==1.18Due to changes in the handling of `np.linspace` and `np.logspace`, https://github.com/numpy/numpy/commit/f4dfe833e3e037bb69113f7250fad3699f918cfc pcigale-2018.0.1 fails with
```python
self.wave = np.logspace(5., 9., 1000.)
File “<__arra...Due to changes in the handling of `np.linspace` and `np.logspace`, https://github.com/numpy/numpy/commit/f4dfe833e3e037bb69113f7250fad3699f918cfc pcigale-2018.0.1 fails with
```python
self.wave = np.logspace(5., 9., 1000.)
File “<__array_function__ internals>“, line 6, in logspace
[...]
TypeError: object of type <class ‘float’> cannot be safely interpreted as an integer.
````https://gitlab.lam.fr/cigale/cigale/-/issues/21Error when "emission = False" (nebular module) and data do not contain emissi...2019-02-09T16:33:28ZGuanggyang206265@gmail.comError when "emission = False" (nebular module) and data do not contain emission linesIf "emission" is set to False in the nebular module and the input data do not have emission lines, an error will occur.If "emission" is set to False in the nebular module and the input data do not have emission lines, an error will occur.https://gitlab.lam.fr/cigale/cigale/-/issues/20The shape of SPIRE filters depends on the extension of the source2018-02-19T14:39:27ZMédéric BoquienThe shape of SPIRE filters depends on the extension of the sourceThe actual effective shape of the SPIRE filters depends on the extension of the source. Currently cigale only has the point-source version. It would be better to include both versions.The actual effective shape of the SPIRE filters depends on the extension of the source. Currently cigale only has the point-source version. It would be better to include both versions.https://gitlab.lam.fr/cigale/cigale/-/issues/8Can't plot best SED without all error columns in the input file2017-07-11T21:17:19ZYannick Roehllyyannick.roehlly@lam.frCan't plot best SED without all error columns in the input fileWhen a flux column in the input table is not associated to an error column, CIGALE can proceed with the analysis using a default 10% error. Nevertheless, pcigale-plot does not work in that case an complain about a missing FOO_err column.When a flux column in the input table is not associated to an error column, CIGALE can proceed with the analysis using a default 10% error. Nevertheless, pcigale-plot does not work in that case an complain about a missing FOO_err column.https://gitlab.lam.fr/cigale/cigale/-/issues/9Can't use a configuration file other than pcigale.ini2017-07-11T21:17:19ZYannick Roehllyyannick.roehlly@lam.frCan't use a configuration file other than pcigale.iniThe pcigale script has an option to indicate which configuration file to use instead of the pcigale.ini file in the work directory: -c. Nevertheless, using this option is not possible because CIGALE expects a pcigale.ini file to be copie...The pcigale script has an option to indicate which configuration file to use instead of the pcigale.ini file in the work directory: -c. Nevertheless, using this option is not possible because CIGALE expects a pcigale.ini file to be copies in the out directory:
```python
FileNotFoundError: [Errno 2] No such file or directory: 'pcigale.ini'
```
https://gitlab.lam.fr/cigale/cigale/-/issues/10Negative fluxes should be treated at normal fluxes2017-07-11T21:17:19ZMédéric BoquienNegative fluxes should be treated at normal fluxesIn principle we should be able to treat negative fluxes as any other flux. Sawicki+12 points this out in footnote 2. A good option would be to treat such fluxes normally if the user does not request them to be treated as upper limits. Th...In principle we should be able to treat negative fluxes as any other flux. Sawicki+12 points this out in footnote 2. A good option would be to treat such fluxes normally if the user does not request them to be treated as upper limits. This should not incur any change in the pcigale.ini file format.
Request by Samir Salim.https://gitlab.lam.fr/cigale/cigale/-/issues/11Systematic uncertainties should be tunable2017-07-11T21:17:19ZMédéric BoquienSystematic uncertainties should be tunableFor now cigale adds a systematic 10% uncertainty to the errors to take into account the uncertainties on the models. This is not always appropriate. We should be able to set the level in pcigale.ini.
Request by Samir Salim.For now cigale adds a systematic 10% uncertainty to the errors to take into account the uncertainties on the models. This is not always appropriate. We should be able to set the level in pcigale.ini.
Request by Samir Salim.https://gitlab.lam.fr/cigale/cigale/-/issues/12Minimal errors should not be systematically set2017-07-11T21:17:19ZMédéric BoquienMinimal errors should not be systematically setCurrently cigale systematically set a minimal error if the uncertainty is lower this this minimal value. This should be optionally disabled as it can add hide problems in the fitting procedure.
Request by Samir Salim.Currently cigale systematically set a minimal error if the uncertainty is lower this this minimal value. This should be optionally disabled as it can add hide problems in the fitting procedure.
Request by Samir Salim.https://gitlab.lam.fr/cigale/cigale/-/issues/13numpy minimal version is 1.9.02017-07-11T21:17:19ZYannick Roehllyyannick.roehlly@lam.frnumpy minimal version is 1.9.0The validation of the ini file makes use of the return_counts parameter of numpy.unique that was introduced in numpy version 1.9.0.
Either set numpy minimal version to 1.9.0 or modify the code to get rid of the parameter.The validation of the ini file makes use of the return_counts parameter of numpy.unique that was introduced in numpy version 1.9.0.
Either set numpy minimal version to 1.9.0 or modify the code to get rid of the parameter.https://gitlab.lam.fr/cigale/cigale/-/issues/14Problem when error is set to 02017-07-11T21:17:19ZYannick Roehllyyannick.roehlly@lam.frProblem when error is set to 0Contrary to what is claimed, CIGALE does not replace errors lower than a tolerance (1e-12) with a 10% default error. This leads to problems in the fitting procedure because of division per 0 when the error column contains 0. Note that a...Contrary to what is claimed, CIGALE does not replace errors lower than a tolerance (1e-12) with a 10% default error. This leads to problems in the fitting procedure because of division per 0 when the error column contains 0. Note that a solution to this problem contradicts bug #12.https://gitlab.lam.fr/cigale/cigale/-/issues/15Filters with decreasing wavelengths lead to wrong results2017-07-11T21:17:19ZYannick Roehllyyannick.roehlly@lam.frFilters with decreasing wavelengths lead to wrong resultsSometime filter response curves are provided with decreasing wavelengths. Such
filters can be imported into CIGALE database without error but when they are
used the produce negative fluxes.
The problem comes from the filter normali...Sometime filter response curves are provided with decreasing wavelengths. Such
filters can be imported into CIGALE database without error but when they are
used the produce negative fluxes.
The problem comes from the filter normalisation and from the integration with
trapz which produces negative values because of the order of the wavelengths.https://gitlab.lam.fr/cigale/cigale/-/issues/16Too few estimates of the IR → issue computing the synchrotron emission2017-07-11T21:17:19ZMédéric BoquienToo few estimates of the IR → issue computing the synchrotron emissionDifferent papers make different assumptions as to what is the infrared exactly. Some assume it is all the dust emission (as we do), some think of it as 8-1000 micron. Some just consider the FIR. We could compute it using different defini...Different papers make different assumptions as to what is the infrared exactly. Some assume it is all the dust emission (as we do), some think of it as 8-1000 micron. Some just consider the FIR. We could compute it using different definitions. This would also allow us to better estimate the synchrotron emission as we rely on the FIR-radio correlation, with the FIR being 42.5-122.5 micron. So in effect we overestimate the synchrotron emission.https://gitlab.lam.fr/cigale/cigale/-/issues/17Add mass weighted and luminosity weighted ages2017-07-11T21:17:19ZMédéric BoquienAdd mass weighted and luminosity weighted agesThe definition of the age in cigale is uncommon and it causes a lot of confusion. We should add explicitly the mass-weighted and/or luminosity-weighted ages.The definition of the age in cigale is uncommon and it causes a lot of confusion. We should add explicitly the mass-weighted and/or luminosity-weighted ages.Médéric BoquienMédéric Boquienhttps://gitlab.lam.fr/cigale/cigale/-/issues/1The redshifting module should not be mandatory to compute the Fnu2017-05-18T11:24:59ZYannick Roehllyyannick.roehlly@lam.frThe redshifting module should not be mandatory to compute the FnuIf one tries to get the fnu property of a SED that did not pass through the redshifting module, it fails because the SED info dictionary does not contain 'universe.luminosity_distance' which is added by the redshifting module. This modul...If one tries to get the fnu property of a SED that did not pass through the redshifting module, it fails because the SED info dictionary does not contain 'universe.luminosity_distance' which is added by the redshifting module. This module should not be mandatory.https://gitlab.lam.fr/cigale/cigale/-/issues/2SFH quenching fails if quenching age is greater than galaxy age2017-05-18T11:24:59ZYannick Roehllyyannick.roehlly@lam.frSFH quenching fails if quenching age is greater than galaxy ageWhen one uses the SFH quenching module if the quenching age is superior to the galaxy age there is an error and the computing stops. Of course, the quenching age must be smaller than the galaxy age but we may need to find a way to deal w...When one uses the SFH quenching module if the quenching age is superior to the galaxy age there is an error and the computing stops. Of course, the quenching age must be smaller than the galaxy age but we may need to find a way to deal with this case.Yannick Roehllyyannick.roehlly@lam.frYannick Roehllyyannick.roehlly@lam.frhttps://gitlab.lam.fr/cigale/cigale/-/issues/3Attenuation below Lyman limit2017-05-18T11:24:59ZYannick Roehllyyannick.roehlly@lam.frAttenuation below Lyman limitThe attenuation modules attenuate the SED below the Lyman limit (for instance the dustatt_calzleit module use the Leitherer formula even if it is not defined below 912 Å). But this attenuation should be treated differently, like it's don...The attenuation modules attenuate the SED below the Lyman limit (for instance the dustatt_calzleit module use the Leitherer formula even if it is not defined below 912 Å). But this attenuation should be treated differently, like it's done in the nebular module with a parameter giving the fraction of Lyman continuum photos absorbed by dust.
First the attenuation modules should exclude the Lyman continuum from the attenuation. Then we should decide either between:
- We assume that no Lyman continuum photon is absorbed by dust. Users needing a more precise attenuation should use the nebular module.
- We take the “fraction of Lyman continuum photons absorbed by dust” out of the nebular module and add it to the attenuation modules.https://gitlab.lam.fr/cigale/cigale/-/issues/4Wrong names for SED components in VO-table export2017-05-18T11:24:59ZYannick Roehllyyannick.roehlly@lam.frWrong names for SED components in VO-table exportWhen exporting a SED to VO-table, one table in named “Flambda” but actually contain luminosities. It should either be renamed or contain the fluxes.When exporting a SED to VO-table, one table in named “Flambda” but actually contain luminosities. It should either be renamed or contain the fluxes.https://gitlab.lam.fr/cigale/cigale/-/issues/5Make the filter list optional in the param module2017-05-18T11:24:59ZYannick Roehllyyannick.roehlly@lam.frMake the filter list optional in the param moduleIf the filter list is empty or is not given when the param module is used CIGALE won't run. This list must not be mandatory.If the filter list is empty or is not given when the param module is used CIGALE won't run. This list must not be mandatory.Yannick Roehllyyannick.roehlly@lam.frYannick Roehllyyannick.roehlly@lam.fr