EMBOSS: Project Meeting (Fri 8th August 2003)
Alan Bleasby, Tim Carver, Jon Ison, Hugh Morgan, Damian Counsell
Minutes of the last meeting
Minutes of the meeting of 25th July 2003 are here
1. Action points
1.1 From this meetings:
1.2 From previous meetings:
ACD docs need updating (action PMR, 2003-02-21)
MEME documentation is required (action AJB, 2003-02-21)
Provide documentation for propehcy (AJB, 2003-03-21)
Inform GES guys of licensing issues wrt wordmatching algorithm. (action PMR, 2003-07-11)
Look at converting Phylip documentation to EMBOSS (action GW, 2003-07025)
2. Software Development
Matrix madness (JCI)
Jon mentioned he had reworked the code for substitution matrices so that you can now matrix labels are now treated as strings rather than characters. This allows for example residue environment substitutions with any number of rows/columns to be created. Hopefully, the user and programmer in fact will not notice any difference whether they give a matrix with character labels or with string labels. The only thing to watch out for is that when using a matrix with string labels, the conversion table pointed to from the matrix object can't be used like a normal conversion table (i.e. ajSeqCvtK should not be used with it - that's what ajSeqCvtKS is there for)
Changes to ajmatrices.c/h
1.1 'codes' element of AjOMatrix is now a AjPStr * rather than an AjPStr (labels of matrices are now kept in an array of strings, rather than a string of characters)
1.2 ajMatrixNew changed for new arguments (array of strings and size of array). This function now calls ajSeqCvtNewZeroS to allocate the conversion table.
1.3 ajMatrixfNew changed (as 1.2)
1.4 ajMatrixDel changed for new element
1.5 ajMatrixfDel changed (as 1.4)
1.6 ajMatrixChar changed. This function now writes a string passed in as an arg (for a matrix label) rather than returning a char.
1.7 ajMatrixfChar (as 1.6)
1.8 ajMatrixRead changed extensively. Notably it now calls ajSeqCvtKS to get integer codes for matrix labels that are strings (not characters as ajSeqCvtK was doing previously)
1.9 ajMatrixfRead (similar changes but not identical to 1.8)
2. Changes to ajfmt.c/h
3. Changes to ajstr.c/h
4. Changes to ajseq.c/h
4.1 ajSeqCvtNewZeroS added
4.2 ajSeqCvtKS added
4.3 'labels' element added to AjOSeqCvt
4.4 ajSeqCvtDel modified for new element
4.5 Modified ajSeqsetConStats for new calls to ajMatrixChar / ajDebug
5. Changes to ajalign.c
5.1 Modified alignConsStats for new calls to ajMatrixChar & ajMatrixfChar / ajDebug
Alan mentioned that he'd committed embIndex.c/h (b+ tree stuff). Library files alphabetically from ajax.c/h to ajmess.c/h have been tidied up to conform to the coding standard and have been commited.
Rodrigo mentioned he would assess the latest version of cpgreport (provided by Peter Robinson) to the version currently in EMBOSS to see if they should be merged or kept as two distinct programs.
3. Documentation & Training
Damian reported that various publishers had now made attractive offers. He will collate a list of the pros and cons for final discussion and decision at the next meeting.
6. Date Of Next Meeting
Next meeting to be held at 9.30 on August 22 2003, Pompeiian Room, Hinxton Hall