makegoodaverage.php needs a better defualt value for avgjump and error checking

[leginonbot]

---

Author Name: Amber Herold (Amber Herold) Original Redmine Issue: 947, https://emg.nysbc.org/redmine/issues/947 Original Date: 2010-10-15

---

note from Arne:

I good this error message:

!!! WARNING: EMAN failed with subprocess error code 1 EMAN: proc3d /ami/data00/appion/10sep01b/recon/emanrecon4/eulers/refine4_10oct14o45/threed. ... /threed.mrc mask=100 norm Traceback (most recent call last): File "/opt/appion/bin/makegoodaverages.py", line 429, in ? makegood.start() File "/opt/appion/bin/makegoodaverages.py", line 380, in start apEMAN.make3d(self.params['oddstack'],"odd.mrc", sym=self.params['sym'],mode=self.params['mode'],hard=self.params['hard'])

KeyError: 'oddstack'

[leginonbot]

---

Original Redmine Comment Author Name: Amber Herold (Amber Herold) Original Date: 2010-10-19T17:34:54Z

---

From Dmitry:

I ran makegoodaverages.py today, and it ran fine for me.

I looked at the command that was issued in (/ami/data00/appion/ 10sep01b/recon/emanrecon4/eulers/refine4_10oct14o45) and noticed that --avgjump=0. What that means is that the program will not process any particles, since it's virtually impossible to come across a particle whose Euler angles stay exactly the same from iteration to iteration.

The webpage defaults to setting this value as 0. What we can do is remove the issuing of a default value, so that the user has to make sure to set it, and issue a warning on the following page if this value is 0. Additionally, we can look at the python and make sure that it doesn't fail if 'nprtcls<1' or provides adequate documentation on the issue.

[leginonbot]

---

Original Redmine Comment Author Name: Amber Herold (Amber Herold) Original Date: 2010-10-19T17:40:47Z

---

Thanks for the update. Do you think we could chose a "smart" default value that would work well for newbies who don't know what they are doing? I think we are missing documentation on this feature so far.

[leginonbot]

---

Original Redmine Comment Author Name: Dmitry Lyumkis (@LyumkisLab) Original Date: 2010-10-19T18:21:59Z

---

Yes, we can choose a "smart" value. A good initial guess would be the average median Euler jump. This is reported in the reconsummary.php page. The full statistics on the Euler jumpers are in eulergraph.php. See, for example, http://cronus3.scripps.edu/~dlyumkis/myamiweb/processing/reconsummary.php?expId=1680, and look at the last two reconstructions, where the average jumps are 22.3 and 17.0, respectively. We can set the value in the makegoodaverages.php page to the median value for the corresponding reconstruction. It might not be a bad idea to add checks / documentation to the python code though, if this error comes up again.

There is some explanation for this value in the pop-up menu, but you're right, we don't have full documentation on it.

[leginonbot]

---

Original Redmine Comment Author Name: Amber Herold (Amber Herold) Original Date: 2010-10-19T18:38:53Z

---

I'm moving this to a future version (don't think it is worth holding up 2.1 release), however the documentation needs to be done asap. Will create a separate issue for that #950.