Here https://github.com/bioinf/Sibelia/blob/master/src/outputgenerator.cpp#L226

FASTA headers instead of 1, 2, 3...

Write reliable FASTA format (http://en.wikipedia.org/wiki/FASTA_format) parser.
Specification of the format: http://blast.ncbi.nlm.nih.gov/blastcgihelp.shtml.
Some notes:

1. Parser must allow only characters {A, C, G, T, U, N, a, c, g, t, n, u} in sequences. Other characters in sequences must be considered as errors
2. Allow spaces in headers. I.e. header "> Header header" must be read as "Header header"

Requirements:

1. There should be a class "FASTAReader" that takes a file name in constructor
2. FASTAReader has method "GetSequences", it returns (by reference) a vector of structs, each struct has two fields ("sequence" and "header")
3. I case of any errors it must throw an exception that contains detailed information about the error, like "Wrong character '$' at line 97, position 33".
4. It can be easily modified to be able to work with gzipped FASTA (or do it now).

Probably find an efficient library that computes LCP directly.

While inserting at position Y use possible free space before Y. For example:
0 1 2 3 4 5
X X X A C G

If we insert before position 3, then elements 0, 1, 2 must be used for insertion.

Linking CXX executable Sibelia
[100%] Built target Sibelia
Install the project...
-- Install configuration: "Release"
-- Installing: /usr/local/bin/Sibelia
CMake Error at cmake_install.cmake:42 (FILE):
file INSTALL cannot copy file
"/home/vyahhi/bioinf-Sibelia-139af21/build/Sibelia" to
"/usr/local/bin/Sibelia".

make: *** [install] Error 1

Incapsulate output into a stand-alone class

50 push_back calls create 50 nodes with only one value in each array

make

[  5%] Building C object libdivsufsort-2.0.1/lib/CMakeFiles/divsufsort.dir/divsufsort.o
[ 11%] Building C object libdivsufsort-2.0.1/lib/CMakeFiles/divsufsort.dir/sssort.o
[ 16%] Building C object libdivsufsort-2.0.1/lib/CMakeFiles/divsufsort.dir/trsort.o
[ 22%] Building C object libdivsufsort-2.0.1/lib/CMakeFiles/divsufsort.dir/utils.o
Linking C static library libdivsufsort.a
[ 22%] Built target divsufsort
Scanning dependencies of target SyntenyFinder
[ 27%] Building CXX object CMakeFiles/SyntenyFinder.dir/main.cpp.o
In file included from /Users/vyahhi/projects/SyntenyBuilder/src/main.cpp:2:
In file included from /Users/vyahhi/projects/SyntenyBuilder/src/outputgenerator.h:4:
In file included from /Users/vyahhi/projects/SyntenyBuilder/src/blockfinder.h:4:
In file included from /Users/vyahhi/projects/SyntenyBuilder/src/hashing.h:4:
In file included from /Users/vyahhi/projects/SyntenyBuilder/src/dnasequence.h:4:
In file included from /Users/vyahhi/projects/SyntenyBuilder/src/fasta.h:5:
/Users/vyahhi/projects/SyntenyBuilder/src/common.h:9:10: fatal error: 'cstdint' file not found
# include <cstdint>

 ^

1 error generated.
make[2]: **\* [CMakeFiles/SyntenyFinder.dir/main.cpp.o] Error 1
make[1]: **\* [CMakeFiles/SyntenyFinder.dir/all] Error 2
make: **\* [all] Error 2```

Add return value for insert method -- iterator to the first inserted element

Make it produce non-overlapping blocks for different multiplicities

Sibelia with --d3 flag generates d3.json. It should be automatically wrapped into d3.html
https://github.com/bioinf/Sibelia/blob/master/src/d3.html#L6753

So Sibelia output should not be d3.json, but customized d3.html.

One button (script) to generate distributions from the master branch:
Sibelia-X.Y.Z-Linux-x86-x64.tar.gz
Sibelia-X.Y.Z-Linux-x86-x64.zip
Sibelia-X.Y.Z-OSX-x86-x64.tar.gz
Sibelia-X.Y.Z-OSX-x86-x64.zip
Sibelia-X.Y.Z-Win-x86--x64.zip

Make a version of unrolled list that has vectors instead of static arrays:

1. two vectors, one for T, other for A (meta)
2. erase physically erases elements (and invalidates them, like insert)
3. size of vectors is limited by NODE_SIZE

Implement https://github.com/bioinf/Sibelia/blob/master/src/unrolledlist.h
At first, read http://en.wikipedia.org/wiki/Unrolled_linked_list and http://en.literateprograms.org/Unrolled_linked_list_(C_Plus_Plus).

Implementation details:

1. Each node contains NODE_SIZE elements (static array)
2. Erased elements are indicated by value erased_indicator (passed in constructor)
3. Iterator's ++ and -- iterators move current iterator to the next (previous) valid (not erased value)
4. Method erase sets all valid values in range [begin, end) to erased_indicator
5. Method insert inserts range [source_begin, source_end) after the place target. If there is any unvalidated iterator x, then x should put in vector invalidated iff notify_predicate(x).
6. Note that implementation must support allocators (as any STL data structure)
7. Write tests for unrolled_list

Current version can produce incorrect results for high cycle size threshold

Get rid of hashtables, use plain vectors to store bifurcation index

[ 75%] Building CXX object CMakeFiles/SyntenyFinder.dir/hashing.cpp.o [ 83%] Building CXX object CMakeFiles/SyntenyFinder.dir/serialization.cpp.o /Users/vyahhi/projects/SyntenyBuilder/src/serialization.cpp:16:48: warning: format specifies type 'unsigned long' but the argument has type 'ull' (aka 'unsigned long long') [-Wformat] sprintf(&buf[0], "[color=\"%s\", label=\"(%lu, %lu)\"];", color.c_str(), static_cast<ull>(chr), static_cast<ull>(pos)); ~~^ ~~~~~~~~~~~~~~~~~~~~~ %llu /Users/vyahhi/projects/SyntenyBuilder/src/serialization.cpp:16:53: warning: format specifies type 'unsigned long' but the argument has type 'ull' (aka 'unsigned long long') [-Wformat] sprintf(&buf[0], "[color=\"%s\", label=\"(%lu, %lu)\"];", color.c_str(), static_cast<ull>(chr), static_cast<ull>(pos)); ~~^ ~~~~~~~~~~~~~~~~~~~~~ %llu /Users/vyahhi/projects/SyntenyBuilder/src/serialization.cpp:94:51: warning: format specifies type 'unsigned long' but the argument has type 'ull' (aka 'unsigned long long') [-Wformat] sprintf(&buf[0], "[color=\"%s\", label=\"chr=%lu pos=%lu len=%lu ch=%c\"];", color.c_str(), uchr, upos, ulength, edge[i].firstChar); ~~^ ~~~~ %llu /Users/vyahhi/projects/SyntenyBuilder/src/serialization.cpp:94:59: warning: format specifies type 'unsigned long' but the argument has type 'ull' (aka 'unsigned long long') [-Wformat] sprintf(&buf[0], "[color=\"%s\", label=\"chr=%lu pos=%lu len=%lu ch=%c\"];", color.c_str(), uchr, upos, ulength, edge[i].firstChar); ~~^ ~~~~ %llu /Users/vyahhi/projects/SyntenyBuilder/src/serialization.cpp:94:67: warning: format specifies type 'unsigned long' but the argument has type 'ull' (aka 'unsigned long long') [-Wformat] sprintf(&buf[0], "[color=\"%s\", label=\"chr=%lu pos=%lu len=%lu ch=%c\"];", color.c_str(), uchr, upos, ulength, edge[i].firstChar); ~~^ ~~~~~~~ %llu 5 warnings generated. [ 91%] Building CXX object CMakeFiles/SyntenyFinder.dir/synteny.cpp.o

Want to get FASTA header of chromosome, not only some arbitrary int GetChr().

Log with progressbars should be printed in stdout and more detailed log should be printed in file.

Implement insert(iterator, const T&) and push_back(const T&) methods
Implement insert methods with new notifications

So that EraseConsistencyTest works

$ make install
[ 20%] Built target divsufsort
Scanning dependencies of target Sibelia
[ 25%] Building CXX object CMakeFiles/Sibelia.dir/main.cpp.o
[ 30%] Building CXX object CMakeFiles/Sibelia.dir/outputgenerator.cpp.o
[ 35%] Building CXX object CMakeFiles/Sibelia.dir/blockfinder.cpp.o
[ 40%] Building CXX object CMakeFiles/Sibelia.dir/bulgeremoval.cpp.o
[ 45%] Building CXX object CMakeFiles/Sibelia.dir/edge.cpp.o
[ 50%] Building CXX object CMakeFiles/Sibelia.dir/graphalgorithm.cpp.o
[ 55%] Building CXX object CMakeFiles/Sibelia.dir/serialization.cpp.o
[ 60%] Building CXX object CMakeFiles/Sibelia.dir/synteny.cpp.o
[ 65%] Building CXX object CMakeFiles/Sibelia.dir/platform.cpp.o
[ 70%] Building CXX object CMakeFiles/Sibelia.dir/vertexenumeration.cpp.o
Linking CXX executable Sibelia
[100%] Built target Sibelia
Install the project...
-- Install configuration: "Release"
-- Installing: /usr/local/bin/Sibelia
CMake Error at cmake_install.cmake:43 (FILE):
file cannot create directory: /usr/share/sibelia. Maybe need
administrative privileges.

make: *** [install] Error 1

Show chromosome ids (FASTARecord::description) not their numbers on the diagram.

Get familiar with Circos (http://circos.ca/). Understand how to use it to generate cool pictures.

Add d3 diagram generation (similar to circos): http://mbostock.github.com/d3/ex/chord.html

If this switch is ON, then all computations are performed in RAM without creating temporary files.

It's too slow now

Add constructor by default
Add method set_erased_value(const T& value)

If someone tries to use unrolled list without setting proper erased value, then program must crash in debug mode (use assert)

Subj

Add return value (next valid iterator) for erase method

Run:

cd build
Sibelia -s loose Helicobacter_pylori.fasta

Got:

Segmentation fault: 11

P.S. I have no file Helicobacter_pylori.fasta in build folder, but Sibelia should explain me this by error message, not segfault.

There should be following parameters to set:

1. --infile= $input_file #input file name, mandatory
2. --parset= fine | loose #parameter set, optional, default = loose
3. --minlength = #minimum block length, optional, default=5000
4. --stagefile = $stage_file #stage file name, optional, incompatible with parset
5. --graphfile = $graph_file #dot file for condensed graph, optional
6. --repfile = $rep_file #path to report file, optional, default = $input_file + "_report"
7. --chrfile = $chr_file #path to chr file, optional, default = $input_file + "_chr"
8. --coordfile = $coord_file #path to coordfile, optional, default = $input_file + "_coord"
9. --blockfile = $block_file #path to blocks file, optional
10. --maxiterations = #maximum number of simplification iterations, default = 4

Extend class OutputGenerator so that it will contain method:

void GenerateCircosOutput(std::ostream & out, ...) const

It must take one (or more, if necessary) std::ostream's and write synteny blocks so that circos can visualize them.

Add a cmd option that allows user to generate circos files. By default, it should be not set.

Update website with new text and interactive picture.

Incorporate suffix array to do this.