ocesou / hooke

Automatically exported from code.google.com/p/hooke
GNU Lesser General Public License v3.0
0 stars 0 forks source link

FJC fails when choosing contact point manually #28

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago
What steps will reproduce the problem?
1. Open Hooke
2. Load the test.hkp playlist
3. run 'autopeak noauto'
4. Select a contact point in the middle of the second peak (see attached 
screenshot)

What is the expected output? What do you see instead?
Either a fit or a message 'no peaks found'.

The steps above are just meant to illustrate the issue, the selection of 
the contact point is obviously not correct.

However, for some of our curves, this problem occurs even when we select a 
reasonable contact point.

Original issue reported on code.google.com by illy...@gmail.com on 18 Nov 2009 at 2:24

Attachments:

GoogleCodeExporter commented 8 years ago
Uh, thanks. I cannot reproduce it, but it confirms something is wrong with the 
FJC
and contact point. I'm looking into it now.

Original comment by devicera...@gmail.com on 19 Nov 2009 at 3:45

GoogleCodeExporter commented 8 years ago
Ok, I think I have a clue. The problem is: The clicked indexes it uses
are indexes *on the Y*. Since it is all reversed, when it evaluates the
fit back etc. it does it on a *Y* range where it calculates the
corresponding X. So the extension of the fitted plot on the X can be
very long or very short, depending on the deltaY between contact point
and measured peak. This explains the plot weirdness and the error.

I'd say that the fits are apparently correct (or at least, if they're
uncorrect is not directly a consequence of these bugs/quirks, that are
mostly cosmetic) -in fact, even when they're weirdly extended a lot
beyond the contact point, autopeak curves all cross at the same point,
which is the clicked contact point, indicating that it indeed does the
fit with the right contact point.

Hope it helps. I am now trying to fix the plotting anyway and to fix the
 crash.

Original comment by devicera...@gmail.com on 19 Nov 2009 at 5:39

GoogleCodeExporter commented 8 years ago
This specific crash should now be solved in revision 212.

Original comment by devicera...@gmail.com on 19 Nov 2009 at 7:16