search

PiRSquared17 / cing

Automatically exported from code.google.com/p/cing
0 stars 0 forks source link

Various bugs/enhancements needed after last major rehaul by Geerten #98

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Need to be split to separate issues of course:

- reformat: whatif.RAMCHK.valueList: 0.17999999999999999 (+- .)
- link to residues should not have #top in url because it obscures the header 
which is needed for 
fast navigation.
- Don't include an orange critique when a red critique is present: e.g. 1brv 
omega 176: ORANGE: 
violMax: 5.774, RED: fractionAbove: 1.000
- For the above residue why is CHI1 red? No critique present or so...
- for 2hgh rearrange the restraint list to bubble up the red restraints.
- split the chain A for 2hgh; it looks stupid.
- correct nucleic acid data (e.g. for wi) because all get's flagged.
- When no atom list (cs) critiques; show no critiques instead of an empty list.
- remove from atom list the link from the critique because it links to 
nowhere...

- fails to read 1y4o_1model test case
- speed up calculation of dihedrals again.
- skip dssp and shiftx when there are only nucleic acids.
- enable the cys checks again by refining the cys residue selection in 2 places.
- Unlinked restraints will in the futrue come thru from ccpn.

Original issue reported on code.google.com by jurge...@gmail.com on 2 Oct 2008 at 2:35

GoogleCodeExporter commented 9 years ago 
- reformat: whatif.RAMCHK.valueList: 0.17999999999999999 (+- .)
Happens for 1brv_1model pro 172
Fixed by adding another formatting to NTvalue() called self.fmt2 and updated 
code through out with revision 
335.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 7:37

GoogleCodeExporter commented 9 years ago 
- link to residues should not have #top in url because it obscures the header 
which is needed for 
the SUPER fast navigation.
Finding bug in html.py#_navigateHtml->HTMLfile#insertHtmlLink->findHtmlLocation

Fixing by reenabling test_classes.py#test_HTMLfile and looking at results.
Looks like we need to point to the header instead of _top tag:
<body>
<div id="container">
    <div id="header">
        <h1>1brv_1model.A.PRO172</h1>
        <a href="../../../index.html#_top">Home</a>
        <a href="../VAL171/index.html#_top">1brv_1model.A.VAL171</a>
        <a href="../index.html#_top">UP</a>
        <a href="../CYS173/index.html#_top">1brv_1model.A.CYS173</a>
        <a href="../../../help.html">Help</a>
    <!-- end header --> </div>
    <div id="main">
    <a id="_top" name="_top"></a>
        <div id="left">
        <h2 id="PHI">PHI</h2>
        <img src="PHI.png" alt=""/>
        <p>PHI   : average:  -47.1   cv:  0.000  ||  outliers:   0 (models [])</p>
        <h2 id="PSI">PSI</h2>
        <img src="PSI.png" alt=""/>
...
        <!-- end left -->       </div>
    <!-- end main -->   </div>
    <br style="clear: both;"/>
<!-- end container --></div>

Fixed by changing the navigation to the header in revision 336.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 8:30

GoogleCodeExporter commented 9 years ago 
- link to residues should not have #top in url because it obscures the header 
which is needed for 
the SUPER fast navigation.
Finding bug in html.py#_navigateHtml->HTMLfile#insertHtmlLink->findHtmlLocation

Fixing by reenabling test_classes.py#test_HTMLfile and looking at results.
Looks like we need to point to the header instead of _top tag:
<body>
<div id="container">
    <div id="header">
        <h1>1brv_1model.A.PRO172</h1>
        <a href="../../../index.html#_top">Home</a>
        <a href="../VAL171/index.html#_top">1brv_1model.A.VAL171</a>
        <a href="../index.html#_top">UP</a>
        <a href="../CYS173/index.html#_top">1brv_1model.A.CYS173</a>
        <a href="../../../help.html">Help</a>
    <!-- end header --> </div>
    <div id="main">
    <a id="_top" name="_top"></a>
        <div id="left">
        <h2 id="PHI">PHI</h2>
        <img src="PHI.png" alt=""/>
        <p>PHI   : average:  -47.1   cv:  0.000  ||  outliers:   0 (models [])</p>
        <h2 id="PSI">PSI</h2>
        <img src="PSI.png" alt=""/>
...
        <!-- end left -->       </div>
    <!-- end main -->   </div>
    <br style="clear: both;"/>
<!-- end container --></div>

Fixed by changing the navigation to the header in revision 336.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 8:30

GoogleCodeExporter commented 9 years ago 
- Don't include an orange critique when a red critique is present: e.g. 1brv 
omega 176: ORANGE: 
violMax: 5.774, RED: fractionAbove: 1.000

Simplified the setMaxColor code but that wasn't the problem.
Fixed bugs in #criticize of DistanceRestraint and DihedralRestraint in revision 
337.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 8:59

GoogleCodeExporter commented 9 years ago 
- Don't include an orange critique when a red critique is present: e.g. 1brv 
omega 176: ORANGE: 
violMax: 5.774, RED: fractionAbove: 1.000

Simplified the setMaxColor code but that wasn't the problem.
Fixed bugs in #criticize of DistanceRestraint and DihedralRestraint in revision 
337.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 8:59

GoogleCodeExporter commented 9 years ago 
- For the above residue why is CHI1 red? No critique present or so...
Not reproduced.
- for 2hgh rearrange the restraint list to bubble up the red restraints.
In 1brv the red one is on top but the orange are at the bottom.
Solution: use same sorting algorithm as Geerten coded in: 
ResidueHTMLfile#_generateRestraintsHtml and use 
an extra method for it.
sortListByRogAndKey

While fixing this; also fixed the annotation of the dihedrals in the dihedrals 
list in revision 338.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 9:36

GoogleCodeExporter commented 9 years ago 
- split the chain A for 2hgh; it looks stupid.
40-49   RGUA40  RGUA41  RADE42  RADE43  URA44   RADE45  RCYT46  RCYT47  RADE48  RGUA49
50-59   RGUA50  URA51   RGUA52  RCYT53  RCYT54  RCYT55  MET104  TYR105  VAL106  CYS107  HIS10
8   PHE109
110-119 GLUH110 ASN111  CYS112  GLY113  LYS114  ALA115  PHE116  LYS117  LYS118  HIS119

There's a cut needed between RCYT55 and MET104. The data came from CYANA 
dealing only with single chains per molecule.
Fixed in revision 339.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 9:54

GoogleCodeExporter commented 9 years ago 
- When no atom list (cs) critiques; show no critiques instead of an empty list.
Fixed by adding check:
        tableSelection = table.rows(critiqued)
        if len(tableSelection.getRows()):

- remove from atom list the link from the critique because it links to 
nowhere...
        atom.rogScore.createHtmlForComments(self.main)
Recoded from:
<a href="" class="orange">xxx</a>
to:
<font color="orange">xxx</font>
for where there are no refs which is always for now.

While we are here: 
A) The columns have been narrowed to all cs values be 50px. 
B) I've abbreviated the messages where possible. Geerten please check these 
mods; you might not find them 
descriptive enough.

Fixed in revision 340.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 11:19

GoogleCodeExporter commented 9 years ago 
- When no atom list (cs) critiques; show no critiques instead of an empty list.
Fixed by adding check:
        tableSelection = table.rows(critiqued)
        if len(tableSelection.getRows()):

- remove from atom list the link from the critique because it links to 
nowhere...
        atom.rogScore.createHtmlForComments(self.main)
Recoded from:
<a href="" class="orange">xxx</a>
to:
<font color="orange">xxx</font>
for where there are no refs which is always for now.

While we are here: 
A) The columns have been narrowed to all cs values be 50px. 
B) I've abbreviated the messages where possible. Geerten please check these 
mods; you might not find them 
descriptive enough.

Fixed in revision 340.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 11:19

GoogleCodeExporter commented 9 years ago 
- enable the cys checks again by refining the cys residue selection in 2 places.
Fixed but left duplication in code in revision 341.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 11:36

GoogleCodeExporter commented 9 years ago 
- enable the cys checks again by refining the cys residue selection in 2 places.
Fixed but left duplication in code in revision 341.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 11:36

GoogleCodeExporter commented 9 years ago 
- fails to read 1y4o_1model test case
No model read from pdb:

WARNING: in #PDB2Molecule: CYANA, model 1 incompatible record (ATOM   3049  HZ3 
LYS B 283       0.440  20.753  -0.381  1.00  0.00           H  )
WARNING: in #PDB2Molecule: CYANA, model 1 incompatible record (ATOM   3188  HZ3 
LYS B 292     -13.442  -7.198   0.819  1.00  0.00           H  )
WARNING: Total number of warnings: 56
==> PDB2Molecule: new Molecule <Molecule "1y4o_1model" (C:2,R:208,A:3904,M:0)> 
from /Users/jd/workspace34/cing/Tests/data/cyana/1y4o_1model/1y4o_1model.pdb
DEBUG: Parsed PDB file 
"/Users/jd/workspace34/cing/Tests/data/cyana/1y4o_1model/1y4o_1model.pdb", 
molecule <Molecule "1y4o_1model" (C:2,R:208,A:3904,M:0)>

Bug was in example data's lack of ENDMDL record. 
Added in revision 342.

 WATCH OUT THIS HAPPENS OFTEN WITH HAND EDITED FILES.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 12:15

GoogleCodeExporter commented 9 years ago 
- skip dssp and shiftx when there are only nucleic acids.
Testing for entry 1a4d. Fixed by adding a check on method 
molecule.hasAminoAcid()

- correct nucleic acid data (e.g. for wi) because all get's flagged.
E.g. there is ANGCHK data for this piece of RNA. O wait, that data present is 
shown. Voiding this bug item.

Fixed the above in revision 343.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 12:47

GoogleCodeExporter commented 9 years ago 
That just leaves:

- speed up calculation of dihedrals again.

which I'll create a separate less important issue for.

Original comment by jurge...@gmail.com on 17 Oct 2008 at 12:53