New example
Issue #13
resolved
There are still Rd files that have missing data or run errors in the example codes (especially in the part of "\dontrun{}". We may consider sticking all codes to one consistent example, like HIV protease.
Comments (8)
-
-
Lets fix the offending Rd files for now and then worry about the new example later.
-
I've included downloading the PDB for each specific example - which I know should be avoided (but it's very convenient). I'm also not consistent in which protein I use. One example protein would be nice, but different functionality might also require different example proteins to make the point you want to show. e.g. NMA: small protein with no fuzz; binding.site: protein with a ligand (and/or other binding partner); structural alignment: missing residues / different sequence, and conformational variability; core.find: large conformational changes and gaps and stuff.
-
reporter
Here is a list of the offending Rd files:
- rmsip.Rd
- pca.xyz.Rd
- dssp.trj.Rd
- read.mol2.Rd
- binding.site.Rd
Note that most problems occurred in the "not run" part of the example codes, which is considered in generating the HTML version documentation. Maybe we could try to fix these files with current examples first.
-
Hmm. mostly my stuff.. shame on me. I though that I could add anything I wanted in the "/dontrun".. thats not the case?
-
The nice new html docs that Xinqiu has been working on include all the output and plots of all the Rd files including the /dontrun sections. We thought this would help folks get a quick feel for the functions usage and output. I really like these new pages. E.g.
see: < http://thegrantlab.org/bio3d/html/blast.pdb.html > and others: < http://thegrantlab.org/bio3d/html/index.html >
At the moment having the /dontrun means that both example() and Rcheck will skip them but not the html doc generation functions. Therefore, ideally all example code should be standalone working snippets with some nice graphical output if appropriate.
Basically, I think example code that takes a long time or download lots of data should be wrapped in a /dontrun
-
reporter
Don't worry, Lars. Actually there are a lot other Rd files that have the similar problems. I have fixed most of them. The above examples couldn't be fixed easily because they will read specific files, like a MD trajectory. Usually we don't care too much about the "dontrun" codes, but we want them be executable anyway. Especially, when we are generating documentations, we will run all the example codes and put the outputs online.
-
reporter
- changed status to resolved
Finally we resolved this issue with new transducin example
- Log in to comment
Do we have a list of the offending Rd files? I agree that HIV-P would be a nice example. However, its a two chain dimer with 300+ structures, which will slow down and complicate the examples, requiring folks to download lots of PDBs and then join appropriate pairs of chains from the same structure.
Ideally we want a small but dynamic single domain structure with not too many PDB structures available. Another option is to limit the number of HIV-P structures downloaded...