.bumpversion.cfg

 [bumpversion]
-current_version = 1.5.0
+current_version = 1.6.0
 parse = (?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\-(?P<stage>(dev|rc)))?(\.(?P<again>\d+))?
 serialize = 
 	{major}.{minor}.{patch}-{stage}.{again}
@@ -18,5 +18,5 @@ values =
 
 [bumpversion:file:configure.ac]
 
-[bumpversion:file:doc/sina.tex]
+[bumpversion:file:doc/source/install.rst]
 

.circleci/config.yml

@@ -79,6 +79,7 @@ jobs:
             make code-coverage-capture
             bash <(curl -s https://codecov.io/bash) -X gcov -F unittest \
               -R ~/project -n Unit-Linux-CircleCI
+          no_output_timeout: 20m # kmer_search_test is slow
       - run:
           <<: *configure
           environment:
@@ -95,6 +96,11 @@ jobs:
             make code-coverage-capture
             bash <(curl -s https://codecov.io/bash) -X gcov -F integration \
               -R ~/project -n Integration-Linux-CircleCI
+      - run:
+          name: Collecting binaries
+          when: on_fail
+          command: |
+            tar -C ~/build -czvf $ARTIFACT_PATH/build_src.tgz src
       - run:
           name: Copying logs
           when: always

.gitignore

@@ -29,27 +29,30 @@
 *.app
 
 # Automake/Autoconf stuff
-build-aux/*
-m4/libtool.m4
-m4/ltoptions.m4
-m4/ltsugar.m4
-m4/ltversion.m4
-m4/lt~obsolete.m4
-Makefile.in
-aclocal.m4
-config.h
-config.h.in
-config.log
-config.status
-configure
-autom4te.cache
-libtool
+/build-aux/*
+/m4/libtool.m4
+/m4/ltoptions.m4
+/m4/ltsugar.m4
+/m4/ltversion.m4
+/m4/lt~obsolete.m4
+/Makefile.in
+/aclocal.m4
+/config.h
+/config.h.in
+/config.log
+/config.status
+/configure
+/autom4te.cache
+/libtool
 .deps
 .dirstamp
-Makefile
+/Makefile
 stamp-h1
 
 # temp files
 *~
 \#*\#
-.\#*
\ No newline at end of file
+.\#*
+
+# build dirs
+/build*/
\ No newline at end of file

.travis.yml

@@ -14,7 +14,6 @@ env:
   global:
     MINICONDA: $HOME/miniconda
     BASH_ENV: $HOME/.bashrc
-    MAKEFLAGS: -j2
 
 install:
   - ./ci_scripts/install_conda.sh

Makefile.am

@@ -80,6 +80,8 @@ src_libsina_la_SOURCES = \
     src/aligned_base.h \
     src/alignment_stats.cpp \
     src/alignment_stats.h \
+    src/buffer.h \
+    src/cache.h \
     src/cseq.cpp \
     src/cseq.h \
     src/cseq_impl.h \
@@ -94,6 +96,7 @@ src_libsina_la_SOURCES = \
     src/mesh_debug.h \
     src/mseq.cpp \
     src/mseq.h \
+    src/progress.h \
     src/pseq.cpp \
     src/pseq.h \
     src/query_arb.cpp \
@@ -108,6 +111,7 @@ src_libsina_la_SOURCES = \
     src/log.h \
     src/scoring_schemes.h \
     src/search.h \
+    src/search.cpp \
     src/search_filter.cpp \
     src/search_filter.h \
     src/timer.h \
@@ -180,12 +184,12 @@ man1_MANS = doc/man/sina.1
 
 $(doc_DATA): doc/build.stamp
 doc/build.stamp: $(DOC_SOURCE)
-	sphinx-build -M text $$(dirname $<) doc
+	$(SPHINX_BUILD) -M text $$(dirname $<) doc
 	touch $@
 CLEANFILES += doc/build.stamp
 
 $(man1_MANS): $(DOC_SOURCE)
-	sphinx-build -M man $$(dirname $<) doc
+	$(SPHINX_BUILD) -M man $$(dirname $<) doc
 
 %: doc/text/%.txt
 	cp $< $@
@@ -225,7 +229,7 @@ bindistdir: all
           cp -fpR $(top_builddir)/$$f $(bindistdir)/$$f; done; \
 	find $(bindistdir) -depth -type d -empty -exec rmdir {} \;
 	binaries=$$(find $(bindistdir) -type f -exec file --mime {} \; | \
-                grep -E "application/x-.*(mach|executable|sharedlib)" | cut -d: -f1); \
+                $(GREP) -E "application/x-.*(mach|executable|sharedlib)" | cut -d: -f1); \
 	export LDPATHS="$(subst -L,,$(filter -L%, $(BOOST_LDFLAGS)))"; \
         for binary in $$binaries; do $(top_srcdir)/tools/fix_libpaths.sh "$$binary"; done; \
 	find $(bindistdir) \( -name \*.la -or -name \*.a \) -delete
@@ -252,7 +256,7 @@ test_libs = \
 LOG_COMPILER = $(top_srcdir)/tools/test_driver.sh
 
 TEST_LOG_DRIVER = env AM_TAP_AWK='$(AWK)' $(SHELL) \
-                  $(top_srcdir)/build-aux/tap-driver.sh
+                  $(top_srcdir)/build-aux/tap-driver.sh --comments
 
 check_PROGRAMS = \
     src/unit_tests/cseq_test \
@@ -260,7 +264,8 @@ check_PROGRAMS = \
     src/unit_tests/idset_test \
     src/unit_tests/kmer_test \
     src/unit_tests/kmer_search_test \
-    src/unit_tests/famfinder_test
+    src/unit_tests/famfinder_test \
+    src/unit_tests/progress_test
 
 src_unit_tests_cseq_test_LDADD = $(test_libs)
 src_unit_tests_cseq_test_LDFLAGS = $(ARB_LDFLAGS)
@@ -274,12 +279,14 @@ src_unit_tests_kmer_search_test_LDADD = $(test_libs)
 src_unit_tests_kmer_search_test_LDFLAGS = $(ARB_LDFLAGS)
 src_unit_tests_famfinder_test_LDADD = $(test_libs)
 src_unit_tests_famfinder_test_LDFLAGS = $(ARB_LDFLAGS)
+src_unit_tests_progress_test_LDADD = $(BOOST_UNIT_TEST_FRAMEWORK_LIB)
 
 test_scripts = \
     tests/readwrite.test \
     tests/align.test \
     tests/search.test \
     tests/infocommands.test \
+    tests/accuracy_kmer.test \
     tests/accuracy.test
 
 
@@ -305,7 +312,7 @@ check-filtered:
 ## pass ARGS to tests
 TEST_ARGS := $(check_DATA) $(TEST_EXTRA_ARGS)
 export TEST_ARGS
-$(test_scripts): $(bin_PROGRAMS) $(check_DATA)
+$(test_scripts): $(bin_PROGRAMS) $(check_DATA) tests/test_helper.sh
 
 ## unzip source test data into build folder
 test_data/%: test_data/%.xz

README.rst

 SINA - reference based multiple sequence alignment
 ==================================================
 
-|latest| |release| |Bioconda| |TravisCI| |CircleCI| |Read the Docs| |Codecov|
+|latest| |Bioconda| |downloads| |TravisCI| |CircleCI| |Read the Docs| |Codecov|
 
 .. |latest| image:: https://img.shields.io/github/release/epruesse/SINA/all.svg?label=latest
 .. |release| image:: https://img.shields.io/github/release/epruesse/SINA.svg
@@ -15,37 +15,65 @@ SINA - reference based multiple sequence alignment
    :target: https://codecov.io/gh/epruesse/SINA
 .. |Read the Docs| image:: https://img.shields.io/readthedocs/sina/latest.svg
    :target: https://readthedocs.org/projects/sina/builds
+.. |downloads| image:: https://img.shields.io/conda/dn/bioconda/sina.svg?style=flat
 
-SINA is a tool to add sequences to an existing multiple sequence
-alignment. It needs about 1 second on a single core to add one 16S
-full length sequence (about 100k/h on a 32-core workstation). It was
-developed to create the multi-million sequence alignment that is the
-core of the SILVA SSU and LSU rRNA databases.
 
-Installation
-------------
+SINA aligns nucleotide sequences to match a pre-existing MSA using
+a graph based alignment algorithm similar to PoA. The graph approach
+allows SINA to incorporate information from many reference sequences
+building without blurring highly variable regions. While
+pure NAST implementations depend highly on finding a good match in
+the reference database, SINA is able to align sequences relatively
+distant to references with good quality and will yield a robust result
+for query sequences with many close reference.
 
-- Use the `online <https://www.arb-silva.de/aligner>`_ version hosted
-  by the SILVA project to align small batches of LSU and SSU
-  sequences to their databses.
-- The preferred way to install SINA locally is via Bioconda
-  (`full instructions <https://sina.readthedocs.io/en/latest/install.html>`_)::
+Features
+--------
+
+- Speed. Aligning 100,000 full length rRNA against the SILVA NR takes 40 minutes on a mid-sized 2018 desktop computer. Aligning 1,000,000 V4 amplicons takes about 60 minutes.
+- Accuracy. SINA is used to build the SILVA_ SSU and LSU rRNA databases.
+- Classification. SINA includes an LCA based classification module.
+- ARB. SINA is able to directly read and write ARB_ format files such as distributed by the SILVA_ project.
+
+.. _SILVA: https://www.arb-silva.de
+.. _ARB: https://www.arb-home.de
+
+Online Version
+--------------
+
+An online version for submitting small batches of sequences is made
+available by the SILVA_ project as part of their
+`ACT: Alignment, Classification and Tree Service <https://www.arb-silva.de/aligner>`_.
+In addition to SINA's alignment and classification stages, ACT allows directly building
+phylogenetic trees with RAxML or FastTree from your sequences and (optionally)
+additional sequences chosen using SINA's add-neighbors feature.
+
+Installing SINA
+---------------
+
+The preferred way to install SINA locally is via `Bioconda <https://bioconda.github.io>`_.
+If you have a working Bioconda installation, just run::
 
     conda create -n sina sina
     conda activate sina
 
+Alternatively, self-contained images are available at
+https://github.com/epruesse/SINA/releases. Choose the most recent ``tar.gz``
+appropriate for your operating system and unpack::
+
+   tar xf sina-1.6.0-linux.tar.gz
+   cd sina-1.6.0
+   ./sina
+
+
 Documentation
 -------------
 
-- The `SINA manual <https://sina.readthedocs.io>`_ is hosted at readthedocs.
-
-- Please refer to the `publication in
-  Bioinformatics <https://doi.org/10.1093/bioinformatics/bts252>`_ for
-  a description of the algorithm.
+The full documentation is available at https://sina.readthedocs.io.
 
-  If you use SINA in your research, please don't forget to cite us:
+The algorithm is explained in the paper:
 
   Elmar Pruesse, Jörg Peplies, Frank Oliver Glöckner; *SINA: Accurate
   high-throughput multiple sequence alignment of ribosomal RNA
   genes.* Bioinformatics 2012; 28 (14): 1823-1829.
-  doi:10.1093/bioinformatics/bts252
+  `doi:10.1093/bioinformatics/bts252 <https://doi.org/10.1093/bioinformatics/bts252>`_

configure.ac

 # Init
 
-AC_INIT([SINA],[1.5.0],[elmar@pruesse.net],,[http://github.com/epruesse/SINA])
+AC_INIT([SINA],[1.6.0],[elmar@pruesse.net],,[http://github.com/epruesse/SINA])
 AC_COPYRIGHT([Copyright (c) 2005-2018 Elmar Pruesse])
 AC_CONFIG_AUX_DIR([build-aux])
 AC_CONFIG_MACRO_DIR([m4])
@@ -28,6 +28,12 @@ AC_PROG_SED
 AC_PROG_LN_S
 AC_PROG_INSTALL
 AC_PROG_MKDIR_P
+AC_CHECK_TOOL([BC], [bc], [:])
+AC_CHECK_TOOL([TEE], [tee], [:])
+AC_CHECK_TOOL([MKTEMP], [mktemp], [:])
+AC_CHECK_TOOL([SPHINX_BUILD], [sphinx-build], [:])
+AC_PROG_GREP
+
 LT_INIT([shared static])
 
 AC_REQUIRE_AUX_FILE([tap-driver.sh])
@@ -125,7 +131,7 @@ AX_ARG_ENABLE([debug], [disable optimizations and add debug symbols],
     AC_DEFINE([DEBUG])
   ])
   CXXFLAGS="$(echo $CXXFLAGS | sed 's/-O@<:@@<:@:alnum:@:>@@:>@*//g')"
-  CXXFLAGS="$CXXFLAGS -O0 -ggdb3"
+  CXXFLAGS="$CXXFLAGS -Og -ggdb3"
   AX_CHECK_COMPILE_FLAG([-fno-omit-frame-pointer],[
     CXXFLAGS="$CXXFLAGS -fno-omit-frame-pointer"
   ])
@@ -191,6 +197,7 @@ if test x"$ARB_HELIX_LIBS" = x""; then
 fi
 
 AX_ARB_CHECK_FUNC([GBT_FIND_SEQUENCE], [GBT_find_sequence(NULL, NULL)])
+AX_ARB_STATUS_RETURN_TYPE
 
 case "$host_os" in
   darwin*)
@@ -268,6 +275,7 @@ CXXFLAGS="$CXXFLAGS -W"
 AC_CONFIG_HEADERS([config.h])
 AC_CONFIG_FILES([
   Makefile
+  tests/test_helper.sh
 ])
 
 
@@ -304,7 +312,8 @@ AC_MSG_NOTICE([    CXXFLAGS:        $CODE_COVERAGE_CXXFLAGS])
 AC_MSG_NOTICE([    LIBS:            $CODE_COVERAGE_LIBS])
 AC_MSG_NOTICE([])
 ])
-AC_MSG_NOTICE([  ARB:               ${ARBHOME}])
+AC_MSG_NOTICE([  ARB:])
+AC_MSG_NOTICE([    ARBHOME:         ${ARBHOME}])
 AC_MSG_NOTICE([    CPPFLAGS:        ${ARB_CPPFLAGS}])
 AC_MSG_NOTICE([    LDFLAGS:         ${ARB_LDFLAGS}])
 AC_MSG_NOTICE([    LIBS:            ${ARB_LIBS}])

debian/changelog

-sina (1.5.0+dfsg-1) UNRELEASED; urgency=medium
+sina (1.6.0+dfsg-1) UNRELEASED; urgency=medium
 
   * Initial release (Closes: #<bug>)
-  TODO: Needs libarb which is non-free see
-     https://github.com/epruesse/SINA/issues/68
+  TODO: Needs libarb
+     https://salsa.debian.org/med-team/libarb
 
- -- Andreas Tille <tille@debian.org>  Fri, 29 Mar 2019 22:30:54 +0100
+ -- Andreas Tille <tille@debian.org>  Thu, 13 Jun 2019 08:25:26 +0200

debian/control

 Source: sina
 Maintainer: Debian Med Packaging Team <debian-med-packaging@lists.alioth.debian.org>
-Uploaders: Andreas Tille <tille@debian.org>
+Uploaders: Andreas Tille <tille@debian.org>,
+           Liubov Chuprikova <chuprikovalv@gmail.com>
 Section: science
 Priority: optional
 Build-Depends: debhelper (>= 12~),

debian/rules

@@ -4,15 +4,6 @@
 export LC_ALL=C.UTF-8
 
 include /usr/share/dpkg/default.mk
-# this provides:
-# DEB_SOURCE: the source package name
-# DEB_VERSION: the full version of the package (epoch + upstream vers. + revision)
-# DEB_VERSION_EPOCH_UPSTREAM: the package's version without the Debian revision
-# DEB_VERSION_UPSTREAM_REVISION: the package's version without the Debian epoch
-# DEB_VERSION_UPSTREAM: the package's upstream version
-# DEB_DISTRIBUTION: the distribution(s) listed in the current entry of debian/changelog
-# SOURCE_DATE_EPOCH: the source release date as seconds since the epoch, as
-#                    specified by <https://reproducible-builds.org/specs/source-date-epoch/>
 
 # for hardening you might like to uncomment this:
 # export DEB_BUILD_MAINT_OPTIONS=hardening=+all
@@ -20,12 +11,11 @@ include /usr/share/dpkg/default.mk
 %:
 	dh $@
 
+override_dh_auto_configure:
+	dh_auto_configure -- --with-arbhome=/usr/lib/x86_64-linux-gnu/arb
+
 ### When overriding auto_test make sure DEB_BUILD_OPTIONS will be respected
 #override_dh_auto_test:
 #ifeq (,$(filter nocheck,$(DEB_BUILD_OPTIONS)))
 #	do_stuff_for_testing
 #endif
-
-### If you **really** can not use uscan (even not with mode=git) use a debian/get-orig-script
-#get-orig-source:
-#	. debian/get-orig-source

doc/Makefile

+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line.
+SPHINXOPTS    =
+SPHINXBUILD   = sphinx-build
+SOURCEDIR     = source
+BUILDDIR      = build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	doxygen
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
\ No newline at end of file

doc/source/CHANGELOG.rst

@@ -3,6 +3,31 @@
 Changelog
 =========
 
+Version 1.6.0:
+--------------
+ - make internal kmer engine the default (:issue:`23`)
+ - add pretty progress monitor
+ - run search stage in parallel (:issue:`32`)
+ - :option:`--num-pts` defaults to number of cores available
+   (previous: 1)
+ - add :option:`--search-engine` setting search engine for search
+   module
+ - always run internal engine without thread limit
+ - split num pt servers evently between search and align
+ - use fixed point format for logging (instead of scientific format)
+ - rewrote family selection (use :option:`--fs-oldmatch` for old
+   implementation)
+ - replace boost::mutex with std::mutex (c++11)
+ - fix :option:`--show-dist` if alignment width don't match
+ - fix race starting pt servers (library code not threadsafe)
+ - fix engine type not shown in :option:`--show-conf`
+ - fix writing to ARB sequence cache not threadsafe
+ - use lock free map for ARB sequence cache (speedup)
+ - add pod buffer to replace std::vector (speedup)
+ - add FIFO cache for kmer search results (speedup for
+   :option:`--search` and :option:`--turn`)
+
+
 Version 1.5.0:
 --------------
  - update documentation (:issue:`20`)
@@ -26,7 +51,8 @@ Version 1.5.0:
  - fix out-of-bounds access on iterator in NAST implementation
  - remove dependency on boost serialization library
  - build release binaries with GCC 7 and C++11 ABI
- - add integration tests watching for accuracy regressions (:issue:`25`)
+ - add integration tests watching for accuracy regressions
+   (:issue:`25`)
 
 
 Version 1.4.0:

doc/source/commandline.rst

@@ -65,7 +65,7 @@ General Options
 
    ``sina -i reference.fasta --prealigned -o reference.arb``
 
-.. option:: -t [all], --turn[=all]
+.. option:: -t [all], --turn [=all]
 
    Enables turn check stage: Sequences not oriented in accordance with
    the reference database will be reverse complemented as needed.
@@ -267,6 +267,11 @@ at least the fraction :option:`--lca-quorum` of the search result.
    alignment reference, but wish to search a larger set of sequences
    for classification purposes.
 
+.. option:: --search-engine=[internal|pt-server]
+
+   Override the value of :option:`--fs-engine` for use within the
+   search module.
+
 .. option:: --search-min-sim=id (0.7)
 
    The minimum fractional identity each result sequence must have with
@@ -721,6 +726,11 @@ Advanced Reference Selection Options
    selection. See :option:`--lca-quorum` for information on how the
    value is interpreted.
 
+.. option:: --fs-oldmatch
+
+   Use the pre-1.6.0 implementation for composing the alignment
+   family. Requires :option:`--fs-engine` = ``pt-server``.
+
 Search & Classify Options
 -------------------------
 

doc/source/index.rst

-Reference based multiple sequence alignment using SINA
-======================================================
+Reference based multiple sequence alignment
+===========================================
 
 SINA allows incorporating additional sequences into an existing
 multiple sequence alignment (MSA) without modifying the original
@@ -24,8 +24,7 @@ format reference databases released by the project `here
 An `online version of SINA <https://www.arb-silva.de/aligner/>`_ is provided
 by the SILVA_ project.
 
-Publication
-~~~~~~~~~~~
+.. rubric:: Publication
 
 If you use SINA in your work, please cite:
 

doc/source/install.rst

@@ -5,9 +5,9 @@ Installing SINA
 
 You can install SINA
 
-- `using Bioconda`_ (simple, recommended)
-- from `pre-compiled tarballs`_ (alternative if conda is misbehaving for you)
-- or build SINA `from source`_ (for developers)
+1. `using Bioconda`_ (recommended)
+2. from `pre-compiled tarballs`_ (alternate)
+3. or build SINA `from source`_ (for developers)
 
 
 .. _`using Bioconda`:
@@ -25,11 +25,12 @@ To install, follow these steps:
    Download the Miniconda_ installer (links for MacOS_ and Linux_),
    execute it and follow the instructions it shows in the shell::
 
-      if test $(uname) == Linux; then
-        wget https://repo.continuum.io/miniconda/Miniconda3-latest-MacOSX-x86_64.sh
-      else
-        wget https://repo.continuum.io/miniconda/Miniconda3-latest-MacOSX-x86_64.sh
-      fi
+      # if you are on MacOS
+      wget https://repo.continuum.io/miniconda/Miniconda3-latest-MacOSX-x86_64.sh
+      # if you are on Linux
+      wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
+
+      # then
       sh Miniconda3-lastest-*-x86_64.sh
 
 2. Add the Conda-Forge_ and Bioconda_ channels::
@@ -70,16 +71,16 @@ are available on the `SINA releases`_ page at Github. Head on over
 there and download the Linux or Macos one. Inside the folder created
 by unpacking the archive, you should find a `sina` executable::
 
-  tar xf ~/Downloads/sina-1.4.0-linux.tar.gz
-  ~/Downloads/sina-1.4.0-linux/sina --help
+  tar xf ~/Downloads/sina-1.6.0-linux.tar.gz
+  ~/Downloads/sina-1.6.0-linux/sina --help
 
 To install SINA system wide, place the contents of the archive in
 `/opt` and create symlinks into `/usr/local/bin`::
 
-  wget https://github.com/epruesse/SINA/releases/download/v1.4.0/sina-1.4.0-linux.tar.gz
-  sudo tar xf sina-1.4.0-linux.tar.gz -C /opt
-  rm sina-1.4.0-linux.tar.gz
-  sudo ln -s /opt/sina-1.4.0-linux /opt/sina
+  wget https://github.com/epruesse/SINA/releases/download/v1.6.0/sina-1.5.0-linux.tar.gz
+  sudo tar xf sina-1.6.0-linux.tar.gz -C /opt
+  rm sina-1.6.0-linux.tar.gz
+  sudo ln -s /opt/sina-1.6.0-linux /opt/sina
   sudo ln -s /opt/sina/bin/sina /usr/local/bin/sina
 
 .. _`SINA releases`: https://github.com/epruesse/SINA/releases

m4/arb.m4

 AC_DEFUN([AX_LIB_ARBDB],
 [
-    AH_TEMPLATE([HAVE_ARB], [], [Defined to 1 if ARB libraries are present])
+    AH_TEMPLATE([HAVE_ARB], [Defined to 1 if ARB libraries are present])
     # Check for dynamic libraries
 
     AC_ARG_WITH([arbhome],
@@ -47,7 +47,7 @@ AC_DEFUN([AX_LIB_ARBDB],
                 success="no"
             fi
 	fi
-	if test x"ax_arb_path" != x""; then
+	if test x"$ax_arb_path" != x""; then
             saved_CPPFLAGS="$CPPFLAGS"
             CPPFLAGS="$CPPFLAGS -I$ax_arb_path/INCLUDE $GLIB_CFLAGS"
             AC_CHECK_HEADER([arbdb.h], [
@@ -247,7 +247,7 @@ AC_DEFUN([AX_LIB_ARB_HELIX],
 [
     AC_REQUIRE([AX_LIB_ARBDB])
 
-    AH_TEMPLATE([HAVE_ARB_HELIX], [], [Defined to 1 if ARB SL HELIX is present])
+    AH_TEMPLATE([HAVE_ARB_HELIX], [Defined to 1 if ARB SL HELIX is present])
     AC_MSG_CHECKING([for ARB HELIX static lib])
 
     saved_CPPFLAGS="$CPPFLAGS"
@@ -281,3 +281,56 @@ AC_DEFUN([AX_LIB_ARB_HELIX],
     LDFLAGS="$saved_LDFLAGS"
 ])
 
+AC_DEFUN([AX_ARB_STATUS_RETURN_TYPE],
+[
+    AH_TEMPLATE([ARB_STATUS_RETURN_TYPE], [void or bool depending on version...])
+    AH_TEMPLATE([ARB_STATUS_RETURN_VALUE], ['' or true depending on RETURN_TYPE])
+    AC_REQUIRE([AX_LIB_ARBDB])
+
+    saved_CPPFLAGS="$CPPFLAGS"
+    saved_LDFLAGS="$LDFLAGS"
+    saved_LIBS="$LIBS"
+    CPPFLAGS="$CPPFLAGS $ARB_CPPFLAGS"
+    LDFLAGS="$LDFLAGS $ARB_LDFLAGS"
+    LIBS="$LIBS $ARBHOME/SL/HELIX/HELIX.a $ARB_LIBS"
+    AC_LANG_PUSH(C++)
+
+    AC_CHECK_HEADER([arb_handlers.h])
+    AC_MSG_CHECKING([ARB status return type])
+
+    ARB_STATUS_RETURN_TYPE=error
+    for sig in bool void; do
+        AC_LINK_IFELSE([
+            AC_LANG_PROGRAM([[
+                #include <arb_handlers.h>
+            ]], [[
+	        arb_status_implementation status;
+		$sig (*x)(const char*) = status.set_title;
+            ]])
+        ],[
+            ARB_STATUS_RETURN_TYPE=$sig
+	    break
+        ])
+    done
+    AC_MSG_RESULT([$ARB_STATUS_RETURN_TYPE])
+    if test x"$ARB_STATUS_RETURN_TYPE" == x"error"; then
+      ARB_STATUS_RETURN_TYPE=
+    fi
+    AC_DEFINE_UNQUOTED(ARB_STATUS_RETURN_TYPE, [$ARB_STATUS_RETURN_TYPE])
+    AC_SUBST(ARB_STATUS_RETURN_TYPE)
+
+    if test x"$ARB_STATUS_RETURN_TYPE" == x"bool"; then
+      ARB_STATUS_RETURN_VALUE=true
+    else
+      ARB_STAUTS_RETURN_VALUE=
+    fi
+    AC_DEFINE_UNQUOTED(ARB_STATUS_RETURN_VALUE, [$ARB_STATUS_RETURN_VALUE])
+    AC_SUBST(ARB_STATUS_RETURN_VALUE)
+
+    AC_LANG_POP(C++)
+    LIBS="$saved_LIBS"
+    CPPFLAGS="$saved_CPPFLAGS"
+    LDFLAGS="$saved_LDFLAGS"
+])
+
+

src/align.cpp

@@ -94,10 +94,10 @@ auto logger = sina::Log::create_logger("align");
 namespace sina {
 
 template<typename SCORING_SCHEME, typename MASTER>
-void choose_transition(cseq& c, cseq& orig, MASTER& m, SCORING_SCHEME& s, ostream& log);
+void choose_transition(cseq& c, const cseq& orig, MASTER& m, SCORING_SCHEME& s, ostream& log);
 
 template<typename transition, typename MASTER>
-void do_align(cseq& c, cseq& orig, MASTER& m, transition& tr, ostream& log);
+void do_align(cseq& c, const cseq& orig, MASTER& m, transition& tr, ostream& log);
 
 struct aligner::options {
     bool realign;
@@ -297,23 +297,6 @@ make_datetime() {
         return string(buf);
 }
 
-struct not_icontains {
-    using result_type = bool;
-    const string bases;
-    explicit not_icontains(string  _bases) : bases(std::move(_bases)) {}
-    bool operator()(const cseq& c) {
-        return !boost::algorithm::icontains(c.getBases(), bases);
-    }
-};
-
-struct iequals_cmp {
-    using result_type = bool;
-    const string bases;
-    explicit iequals_cmp(string  _bases) : bases(std::move(_bases)) {}
-    bool operator()(const cseq& c) {
-        return iequals(bases, c.getBases());
-    }
-};
 
 aligner::aligner() = default;
 aligner::~aligner() = default;
@@ -324,65 +307,71 @@ aligner& aligner::operator=(const aligner&  /*a*/) = default;
 tray
 aligner::operator()(tray t) {
     // skip if requirements missing
-    // FIXME: add logging here
     if ((t.input_sequence == nullptr) ||
         (t.alignment_reference == nullptr) ||
         (t.astats == nullptr) ) {
+        logger->error("Internal error - incomplete data for alignment");
         return t;
     }
     // prepare variables
     cseq &c = *(new cseq(*t.input_sequence)); // working copy
-    vector<cseq> &vc = *t.alignment_reference;
-    string bases = c.getBases(); // unaligned sequence
+    search::result_vector &vc = *t.alignment_reference;
+    const string bases = c.getBases(); // unaligned sequence
 
     if (opts->lowercase != LOWERCASE_ORIGINAL) {
         c.upperCaseAll();
     }
 
     // sort reference sequences containing candidate to end of family
-    auto it = partition(vc.begin(), vc.end(), not_icontains(bases));
+    auto not_contains_query = [&](search::result_item& item) {
+        // FIXME: we can do this w/o converting to string
+        return !boost::algorithm::icontains(item.sequence->getBases(), bases);
+    };
+    auto begin_containing = partition(vc.begin(), vc.end(), not_contains_query);
 
     // if there are such sequences...
-    if (it != vc.end()) {
-        if (opts->realign) { // ...either realign (throw them out)
+    if (begin_containing != vc.end()) {
+        if (opts->realign) { // realign means ignore those sequences
+            // FIXME: this should be done in famfinder to re-fill family
             t.log << "sequences ";
-            for (auto it2 = it; it2 != vc.end(); ++it2) {
-                t.log << it->get_attr<string>(query_arb::fn_acc) << " ";
+            for (auto it = begin_containing; it != vc.end(); ++it) {
+                t.log << it->sequence->get_attr<string>(query_arb::fn_acc) << " ";
             }
             t.log << "containing exact candidate removed from family;";
-            vc.erase(it, vc.end());
-            if (it == vc.begin()) {
+            vc.erase(begin_containing, vc.end());
+            if (vc.empty()) {
                 t.log << "that's ALL of them. skipping sequence;";
                 return t;
             }
-        } else { // ...or steal their alignment
-            auto exact_match = find_if(it, vc.end(), iequals_cmp(bases));
+        } else { // otherwise, we steal the alignment
+            auto same_as_query = [&](search::result_item& item) {
+                return iequals(bases, item.sequence->getBases());
+            };
+            auto exact_match = find_if(begin_containing, vc.end(), same_as_query);
             if (exact_match != vc.end()) {
-                c.setAlignedBases(exact_match->getAlignedBases());
+                c.setAlignedBases(exact_match->sequence->getAlignedBases());
                 t.log << "copied alignment from identical template sequence "
-                      << exact_match->get_attr<string>(query_arb::fn_acc) << ":"
-                      << exact_match->get_attr<string>(query_arb::fn_start)
+                      << exact_match->sequence->get_attr<string>(query_arb::fn_acc) << ":"
+                      << exact_match->sequence->get_attr<string>(query_arb::fn_start, "0")
                       << "; ";
             } else {
-                vector<aligned_base> subalignment, refalignment;
-                string refsequence = it->getBases();
-                boost::iterator_range<string::iterator> substr;
-                refalignment = it->getAlignedBases();
-
-                substr = boost::ifind_first(refsequence,bases);
+                vector<aligned_base> subalignment;
+                const vector<aligned_base>& refalignment = begin_containing->sequence->getAlignedBases();
+                string refsequence = begin_containing->sequence->getBases();
+                boost::iterator_range<string::iterator> substr = boost::ifind_first(refsequence, bases);
                 subalignment.reserve(substr.size());
-                std::copy( refalignment.begin() + std::distance(refsequence.begin(), substr.begin()),
-                           refalignment.begin() + std::distance(refsequence.begin(), substr.end()),
-                           std::back_inserter(subalignment) );
-
+                std::copy(refalignment.begin() + std::distance(refsequence.begin(), substr.begin()),
+                          refalignment.begin() + std::distance(refsequence.begin(), substr.end()),
+                          std::back_inserter(subalignment));
                 c.setAlignedBases(subalignment);
                 t.log << "copied alignment from (longer) template sequence "
-                      << it->get_attr<string>(query_arb::fn_acc) << ":"
-                      << it->get_attr<string>(query_arb::fn_start)
+                      << begin_containing->sequence->get_attr<string>(query_arb::fn_acc) << ":"
+                      << begin_containing->sequence->get_attr<string>(query_arb::fn_start, "0")
                       << "; ";
                 BOOST_ASSERT(bases == c.getBases());
-           }
-            c.setWidth(it->getWidth());
+            }
+
+            c.setWidth(begin_containing->sequence->getWidth());
             c.set_attr(query_arb::fn_date, make_datetime());
             c.set_attr(query_arb::fn_qual, 100);
             if (opts->calc_idty) {
@@ -390,15 +379,21 @@ aligner::operator()(tray t) {
             }
             c.set_attr(query_arb::fn_head, 0);
             c.set_attr(query_arb::fn_tail, 0);
-            c.set_attr("align_filter_slv", "");
+            c.set_attr(query_arb::fn_filter, "");
             t.aligned_sequence = &c;
             return t;
         }
     }
 
+    std::vector<const cseq*> vcp;
+    vcp.reserve(vc.size());
+    for (auto& r : vc) {
+        vcp.push_back(r.sequence);
+    }
+
     if (!opts->fs_no_graph) {
         // prepare reference
-        mseq m(vc.begin(), vc.end(), opts->fs_weight);
+        mseq m(vcp.begin(), vcp.end(), opts->fs_weight);
         // (remove duplicate edges:)
         m.sort();
         m.reduce_edges();
@@ -416,11 +411,11 @@ aligner::operator()(tray t) {
                 choose_transition(c, *t.input_sequence, m, s, t.log);
             }
         } else {
-            vector<float> weights(vc.begin()->getWidth(), 1.f);
+            vector<float> weights(vc.begin()->sequence->getWidth(), 1.f);
             if (t.astats->getWidth() == 0) { // FIXME: this looks broken
                 weights = t.astats->getWeights();
             }
-            float dist = vc.begin()->getScore();
+            float dist = vc.begin()->score;
             t.log << "using dist: " << dist << endl;
             scoring_scheme_matrix<aligned_base::matrix_type>
                 s(opts->gap_penalty, opts->gap_ext_penalty, weights,
@@ -428,7 +423,7 @@ aligner::operator()(tray t) {
             choose_transition(c, *t.input_sequence, m, s, t.log);
         }
     } else {
-        pseq p(vc.begin(), vc.end());
+        pseq p(vcp.begin(), vcp.end());
         scoring_scheme_profile s(-opts->match_score, -opts->mismatch_score,
                                  opts->gap_penalty, opts->gap_ext_penalty);
         choose_transition(c, *t.input_sequence, p, s, t.log);
@@ -436,8 +431,8 @@ aligner::operator()(tray t) {
 
     if (opts->write_used_rels) {
         stringstream tmp;
-        for (const cseq &s: vc) {
-            tmp << s.getName() << " ";
+        for (auto &s: vc) {
+            tmp << s.sequence->getName() << " ";
         }
         c.set_attr(query_arb::fn_used_rels, tmp.str());
     }
@@ -448,14 +443,14 @@ aligner::operator()(tray t) {
                                 CMP_COVER_OVERLAP,
                                 false);
         float idty = 0;
-        for (const cseq &s: vc) {
-            idty = std::max(idty, calc_id(c, s));
+        for (auto &s: vc) {
+            idty = std::max(idty, calc_id(c, *s.sequence));
         }
         c.set_attr(query_arb::fn_idty, 100.f*idty);
     }
 
     c.set_attr(query_arb::fn_date, make_datetime());
-    c.set_attr("align_filter_slv", t.astats->getName());
+    c.set_attr(query_arb::fn_filter, t.astats->getName());
     t.aligned_sequence = &c;
 
     return t;
@@ -463,7 +458,7 @@ aligner::operator()(tray t) {
 
 template<typename SCORING_SCHEME, typename MASTER>
 void
-sina::choose_transition(cseq& c, cseq& orig, MASTER& m,
+sina::choose_transition(cseq& c, const cseq& orig, MASTER& m,
                   SCORING_SCHEME& s, ostream& log) {
     if (aligner::opts->insertion == INSERTION_FORBID) {
         transition_aspace_aware<SCORING_SCHEME, MASTER, cseq> tr(s);
@@ -476,9 +471,8 @@ sina::choose_transition(cseq& c, cseq& orig, MASTER& m,
 
 template<typename transition, typename MASTER>
 void
-sina::do_align(cseq& c, cseq& orig, MASTER &m,
+sina::do_align(cseq& c, const cseq& orig, MASTER &m,
                transition &tr, ostream& log) {
-
     using cnsts_type = compute_node_simple<transition>;
     using data_type = typename cnsts_type::data_type;
     cnsts_type cns(tr);
@@ -489,9 +483,6 @@ sina::do_align(cseq& c, cseq& orig, MASTER &m,
     mesh_t A(m, c);
 
     int oh_head, oh_tail;
-#ifdef DEBUG
-    log << "refsize: " << m.size() << "; ";
-#endif
 
     // compute values of mesh nodes
     compute(A, cns);
@@ -500,15 +491,15 @@ sina::do_align(cseq& c, cseq& orig, MASTER &m,
     c.clearSequence();
 
     // run backtracking on the mesh
-    backtrack(A, c, tr,
-              aligner::opts->overhang,
-              aligner::opts->lowercase,
-              aligner::opts->insertion,
-              oh_head, oh_tail, log);
+    float score = backtrack(A, c, tr,
+                            aligner::opts->overhang,
+                            aligner::opts->lowercase,
+                            aligner::opts->insertion,
+                            oh_head, oh_tail, log);
     // alignment done :-)
     c.set_attr(query_arb::fn_head, oh_head);
     c.set_attr(query_arb::fn_tail, oh_tail);
-    c.set_attr(query_arb::fn_qual, (int)std::min(100.f, std::max(0.f, 100.f * c.getScore())));
+    c.set_attr(query_arb::fn_qual, (int)std::min(100.f, std::max(0.f, 100.f * score)));
 
     if (aligner::opts->debug_graph) {
         ofstream out(fmt::format("mseq_{}.dot", c.getName()));

src/aligned_base.h

@@ -62,7 +62,7 @@ public:
 
     class bad_character_exception : public std::exception {
     public:
-        bad_character_exception(value_type c) noexcept
+        explicit bad_character_exception(value_type c) noexcept
             : character(c)
         {
         }
@@ -102,9 +102,7 @@ public:
     }
 
     /* construct from base_type */
-    base_iupac(base_types b) {
-        _data = 1 << b;
-    }
+    base_iupac(base_types b) : _data(1<<b) {}
 
     /* explicit cast to base_type */
     base_types getBaseType() const {
@@ -217,7 +215,7 @@ public:
     }
 
 protected:
-    int count_bits(unsigned char c) const {
+    static int count_bits(unsigned char c) {
 #define HAVE_BUILTIN_POPCOUNT
 #ifdef HAVE_BUILTIN_POPCOUNT
         return __builtin_popcount(c);