Re: Install question
Sebastian Haase
1. After I recently also installed Ubuntu, I found the same problem
you found: /bin/sh is not a link to /bin/bash -- but I fixed the
script so that /bin/dash (i.e. Ubuntu's /bin/sh) now also works ---
this version should be on google code ... but your fix is also just
fine !!
2. The Changelog_SF_old is really "old" !! By now the 'P' module
refers to 'Plot' (not 'Priidsm') -- I reused the letter, after it had
been freed up for years ...
BUT: the error you got is not fatal for AIDA -- in fact the 'P'
module is not used anywhere so far (because it is not installed on all
platforms, as you found ;-) )
So, all should / could be fine for you now.....
again:
you can safely ignore the "* couldn't load module(s): P"
What else did your AIDA run produce ? More error messages, or did it
really deconvolve successfully and you just did not expect it to ;-)
Regards,
Sebastian
On Mon, Aug 25, 2008 at 6:19 PM, Erik F. Y. Hom <er...@mcb.harvard.edu> wrote:
> Dear Bart,
>
> Thanks for your question and for informing me of the problem. And sorry for
> the delayed reply - my wife will be giving birth anytime soon, so have been
> a bit preoccupied....
>
> Anyhow, I think this may be a Priithon upgrade-related issue and I've cc:ed
> the developers of that of this. We'll try to get to the bottom of this and
> let you know. Sorry for the bugs!
>
> Best,
> erik
>
>
>
> On Aug 23, 2008, at 2:46 PM, Bart Hazes wrote:
>
>> Dear AIDA team,
>>
>> I just installed the AIDA package on an UBUNTU linux PC. The three files I
>> downloaded were
>>
>> Priithon_25_lin.tbz
>> <http://msg.ucsf.edu/sedat//Priithon/Priithon_25_lin.tbz> from
>> http://msg.ucsf.edu/sedat//Priithon/
>> (note: the Linux:32-bi
>> <http://www.ucsf.edu/sedat/Priithon/PrLinN_20071031.tbz>link on
>> http://msg.ucsf.edu/AIDA/index.html is dead)
>>
>> AIDA1p2p1.tbz from the AIDA home page
>>
>> testdataset.tar.bz2 from the AIDA home page
>>
>> All compressed tar files were unpacked as indicated, generating
>> directories
>>
>> AIDA1p2p1
>> Priithon_25_lin
>> testdataset
>>
>> In the AIDA1p2p1 directory I created "priithon" as a symbolic link to
>> ../Priithon_25_lin/priithon_script
>>
>> To run the test case I went to the testdataset directory and issued:
>>
>> nohup ../AIDA1p2p1/priithon ../AIDA1p2p1/AIDA.py -S 2Dtest_Settings.py
>>
>> This resulted in thousands of "[: ###: ==: unexpected operator" lines in
>> nohup.out, with ### corresponding to "fi" lines in scripts.
>>
>> Changing #!/bin/sh to #!/bin/bash in the first line of
>> Priithon_25_lin/priithon_script solves this (on Ubuntu /bin/sh is a symbolic
>> link to /bin/dash), but now nohup.out says:
>>
>> * couldn't load module(s): P
>>
>> On "http://code.google.com/p/priithon/wiki/ChangeLog_old_SF" I found the
>> following:
>>
>> 2006-07-19
>>
>> * removed P and S from being loaded automatically with from
>> Priithon.all import * (P=priism: use F and Mrc instead -- S=seb:
>> use U instead)
>>
>> Suggesting that this module should no longer autoload. As I could not find
>> the location of the module I am at a loss on how to proceed.
>>
>> I would very much appreciate your help, either pointing me to the
>> appropriate documentation or explaining what needs to be done to run the
>> test example.
>>
>> Sincerely, Bart Hazes
>>
>>
>>
>>
>
>
Sebastian Haase
I'm very sorry about the current mess -- the problem results from the
fact that I personally don't use AIDA (yet).
Some general points:
1) I don't know about the "official" state of AIDA using Python 2.5
-- in the past AIDA was known to work with Python 2.4 -- definitely
there are know problems on Intel Mac running 2.5
1b) so if you can at all install python 2.4 try this first please ---
just to lower your own pain level .....
2) The "current" Priithon is definitely based on Py 2.5 and to be
downloaded from the google code site. But there are absolutely no
changes affecting AIDA -- except one might work, the other doesn't
;-) :-(
3) generally it makes only sense to a certain point to try "hacking
into unknown code" --- e.g. while
CCG_d (without leading underscore) refers to a python thin wrapper
module, _CCG_d (with leading underscore) referes to the C/C++ compiled
binary module -- there are two different things and are not
interchangeable
3b) maybe the ccg problem is solved in the package downloadable from
the google-code site -- I will look into this.
3c) I really need to refresh my memory on using / testing the CCG
module directly -- I think this seems to be the central
trouble-generator if AIDA is not behaving right ..
What do you say ? -- keep me posted.
-Sebastian
PS: I'm cc'ing the Priithon mailing list, because these are some
general points - and AIDA is one of the "main players" for Priithon
On Fri, Aug 29, 2008 at 12:02 AM, Bart Hazes <bart....@ualberta.ca> wrote:
>
> Hi Sebastian,
>
> Thanks for the info but I am not getting any results yet ;(
> I thought the "P" module was the problem but apparently it must be something
> else.
>
> To be sure, I started again from the original tarbals to reinstall the
> package and discovered that I forgot to mention one other change I needed to
> make to get the program to "work". After sorting out the bash issue the
> progam actually reported errors relating to the CCG_d and CCG_s modules:
>
> nohup ../Priithon_25_lin/priithon ../AIDA1p2p1/AIDA.py -S 2Dtest_Settings.py
> nohup: appending output to `nohup.out'
>
> cat nohup.out
> * couldn't load module(s): P
> Traceback (most recent call last):
> File "../AIDA1p2p1/AIDA.py", line 20, in <module>
> import AIDA_Functions as Fun
> File "/home/bart/Imaging/Aida/AIDA1p2p1/AIDA_Functions.py", line 19, in
> <module>
> from Priithon import ccg
> File "/home/bart/Imaging/Aida/Priithon_25_lin/Priithon/ccg.py", line 4, in
> <module>
> import CCG_d as _CCG_d
> File "/home/bart/Imaging/Aida/Priithon_25_lin/Priithon/CCG_d.py", line 7,
> in <module>
> import _CCG_d
> ImportError: No module named _CCG_d
>
> Since the ccg.py file used "import CCG_d as _CCG_d" instead of
> "import _CCG_d" I replaced the latter with the former in CCG_d.py (and had
> to make a similar change for CCG_s.py).
> In hindsight it doesn't make much sense to have a python file import itself
> so perhaps that's were I went wrong. If so, where do I find the _CCG_d and
> _CCG_s modules?
>
> However, after the change running the program takes about 10 seconds or so
> to complete, so something is finally happening. In addition, a directory
> _Results_2D/2D_myopic_080828_134555/ is created with a single log file of
> length zero bytes (AIDA_080828_134555.log). The nohup file now contains:
>
> ImportError: No module named _CCG_d
> ../Priithon_25_lin/priithon: line 41: 5221 Illegal instruction $PY $*
>
> I checked and $PY, $PRI and $* are set properly to
> python2.5 -Qnew
> /home/bart/Imaging/Aida/Priithon_25_lin
> ../AIDA1p2p1/AIDA.py -S 2Dtest_Settings.py
>
> PYTHONPATH and LD_LIBRARY_PATH also seem to be set properly.
>
> So it looks like the program gets through the priithon startup script and
> then bails out after it has started the main python code. I hope these
> symptoms give you a clue to the diagnosis as I'm at a loss.
>
> Bart
> --
>
> Bart Hazes (Associate Professor)
> Dept. of Medical Microbiology & Immunology
> University of Alberta
> 1-15 Medical Sciences Building
> Edmonton, Alberta
> Canada, T6G 2H7
> phone: 1-780-492-0042
> fax: 1-780-492-7521
>