19 August 2020 at 4:39 pm #574704
Having just found this information I decided to look into having a go for the Ariel Project and having signed up and read the the manual – decided to download the software. First question – the Anaconda website is now presenting only a 3.8 version of Python is this compatible with the HOPS sofware? As we have still 8 years to launch this will present an ongoing issue. This also means a new version of the HOPS manual is also needed.
Second question is how would I be able to use AstroImageJ to do the same thing – this is based on a fixed platform version of ImageJ and not as yet based on the commercial considerations of a third party? Have got AstroimageJ down to a fine art for variable star processing. (It was originally designed for exoplanet transits!)
George20 August 2020 at 2:18 am #583025Dominic FordKeymaster
Welcome to the wonderful world of Python, where specific versions are only supported for a couple of years before programmers are expected to update all the code they’ve ever written to run in a newer language version.
However, the good news is that Python 3.7 isn’t dead yet (it’s supported until 2023), and I believe Anaconda still support it too.
It’s just that the Anaconda website is entirely unhelpful. If you download the current installer — the one Anaconda tell you is Python 3.8 — and then you follow the completely incomprehensible instructions here: <https://conda.io/projects/conda/en/latest/user-guide/concepts/environments.html> then you can use it to create a Python 3.7 installation.
In case you are not a fluent speaker of gobbledygook, and need somebody to translate the Anaconda website into English for you… there’s a StackOverflow thread here where somebody has asked a question very similar to yours, and the comments explain exactly what commands you need to type to get a Python 3.7 installation… https://stackoverflow.com/questions/42978349/anaconda-version-with-python-3-520 August 2020 at 8:09 am #583026Grant PrivettParticipant
I discovered the Python Wikipedia entry this week. Highly entertaining reading. About the only thing it doesn’t do, apparently, is turn water into wine – well, thats what the zealots appear to think it seems.20 August 2020 at 9:34 am #583027William BristowParticipant
I have re-run and compared three sets of transit data through HOPS using the latest Anaconda – Python 3.8.3 distribution against the previous Anaconda – Python 3.7.6 distribution and the output from HOPS is identical with no issues found on Windows 10 build 2004 or macOS Catalina. I’m not currently able to test on a Linux build or earlier builds of Windows 10 or macOS.
Based on this (limited) test I would say it is ok to use the latest Anaconda – Python 3.8.3 distribution as a platform for HOPS using Windows 10 build 2004 or macOS Catalina.
The installation guide for Anaconda – Python – HOPS contained in appendix A of Roger Dymock’s Ariel Space Mission document remains valid with only the need to account for the latest version numbers of Anaconda and Python in the text and figures:
Ultimately though it is the responsibility of the HOPS software author to test against a new release of Anaconda – Python, which historically occur two or three times per year, and revise the HOPS release build and instructions accordingly. All the previous versions of Anaconda – Python are available should you wish to install and run HOPS in it’s original build environment:
Regarding using AstroImageJ for exoplanet detection you will find Roger’s exoplanet section here on the BAA website contains links to all the information you require:
Roger’s page provides web links out to the most relevant sites, I would suggest beginning with the Grady Boyce – Dennis Conti Cookbook and guides and the latest release of AstroImageJ:
William.20 August 2020 at 9:54 am #583028Paul LeylandParticipant
To be fair to Python (which is hard for me because I have a profound dislike of the language) the various point releases (e.g. 3.7 versus 3.8) appear to be very backward compatible. Something written for 3.7 will almost certainly run without any issues on 3.8 and, when it arrives, 3.9. Problems can arise in the other direction but old code tends to run well.
The real problems arose when Python 2.x reached end of life earlier this year. Vast amounts of code is now on emergency life support with kludged-together installations of Python 2. Much more is unavailable. To give just one example, my favourite GUI network scanner, zenmap, is no longer part of the Linux systems I run.20 August 2020 at 2:10 pm #583031
Thank you that is most helpful about the HOPS software and Python. LIke previous replies I have reservations about using such a fluid language which has a had at least one really big change and over the next 8 years could cause a similar break in support.
As I mentioned AstroImageJ would be my first choice and I have used it a lot on variable stars – hadn’t realised there was quite so much material for exoplanets. I have a lot of reading and cross checking to do. Am I right in presuming that the resulting lightcurve can be uploaded to the exoclock website or should I be thinking about processing via AstroImageJ and then uploading from HOPS software? Or I just need to do more reading!?
George20 August 2020 at 2:55 pm #583032William BristowParticipant
For the Ariel Space Mission, ExoClock project, the preferred data input method is specifically the fitted plot straight out of HOPS.
You can see some data that has already submitted by following the link below, click the “View” button(s) to inspect the plots:
For newly announced exoplanet discoveries that are currently unknown to the ExoClock database and your own survey projects then AstroImageJ would be first choice since HOPS will be unable to process the data until a minimum set of object properties are known and the HOPS database is updated.
William.20 August 2020 at 3:28 pm #583033
Sounds like my next mission is to get the HOPS Software up and running for the upload capabilty. I think I will use the AstroImageJ to process the initial images,flats, darks etc and then load the resulting scientific images to HOPS.
Looks like I have even more reading to while away the cloud covered evenings – are we ever going to see stars again in North East England?
Thank you for your helpful comments
- You must be logged in to reply to this topic.