New upstream version 3.2.1+git20191130.5b131b9 (b1b6ea2d) · Commits · Debian Med / minia

CMakeLists.txt

+3 −1

Original line number	Diff line number	Diff line
		@@ -8,7 +8,7 @@ cmake_minimum_required (VERSION 2.6)
		# The default version number is the latest official build
		SET (gatb-tool_VERSION_MAJOR 3)
		SET (gatb-tool_VERSION_MINOR 2)
		SET (gatb-tool_VERSION_PATCH 0)
		SET (gatb-tool_VERSION_PATCH 1)

		# But, it is possible to define another release number during a local build
		IF (DEFINED MAJOR)
		@@ -84,6 +84,8 @@ link_directories (${gatb-core-extra-libraries-path})
		set (PROGRAM_SOURCE_DIR ${PROJECT_SOURCE_DIR}/src)
		set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)

		cmake_policy(SET CMP0009 NEW) # fixes cmake complaining about symlinks

		include_directories (${PROGRAM_SOURCE_DIR})
		file (GLOB_RECURSE ProjectFiles ${PROGRAM_SOURCE_DIR}/*)
		add_executable(${PROJECT_NAME} ${ProjectFiles})

README.md

+17 −7

Original line number	Diff line number	Diff line
		@@ -2,24 +2,27 @@

		[![License](http://img.shields.io/:license-affero-blue.svg)](http://www.gnu.org/licenses/agpl-3.0.en.html)

		<!---
		\| Linux \| Mac OSX \|
		\|-----------\|-------------\|
		[![Build Status](https://ci.inria.fr/gatb-core/view/Minia/job/tool-minia-build-debian7-64bits-gcc-4.7/badge/icon)](https://ci.inria.fr/gatb-core/view/Minia/job/tool-minia-build-debian7-64bits-gcc-4.7/) \| [![Build Status](https://ci.inria.fr/gatb-core/view/Minia/job/tool-minia-build-macos-10.9.5-gcc-4.2.1/badge/icon)](https://ci.inria.fr/gatb-core/view/Minia/job/tool-minia-build-macos-10.9.5-gcc-4.2.1/)
		--->

		# Before continuing..

		# What is Minia ?
		If you are looking to do high-quality genome or metagenome assemblies, please go here: https://github.com/GATB/gatb-minia-pipeline This is a pipeline built on top of Minia that does a similar algorithm to metaSpades and MEGAHIT (multi-k assembly).

		Minia is a short-read assembler based on a de Bruijn graph, capable of assembling a human genome on a desktop computer in a day. The output of Minia is a set of contigs. Minia produces results of similar contiguity and accuracy to other de Bruijn assemblers (e.g. Velvet).
		# Introduction

		# Getting the latest source code
		Minia is a short-read assembler based on a de Bruijn graph, capable of assembling a human genome on a desktop computer in a day. The output of Minia is a set of contigs. Back when it was released, Minia produced results of similar contiguity and accuracy to other de Bruijn assemblers (e.g. Velvet). Now (2015 onwards), genome assemblers have evolved and in order ot have high contiguity, see the previous section.

		## Requirements
		# Getting the latest source code

		CMake 2.6+; see http://www.cmake.org/cmake/resources/software.html
		## Instructions

		C++11 compiler; (g++ version>=4.7 (Linux), clang version>=4.3 (Mac OSX))
		It is recommended to use download the latest binary release (Linux or OSX) there: https://github.com/GATB/minia/releases

		## Instructions
		Otherwise, Minia may be compiled from sources as follows:

		# get a local copy of minia source code
		git clone --recursive https://github.com/GATB/minia.git
		@@ -28,6 +31,13 @@ C++11 compiler; (g++ version>=4.7 (Linux), clang version>=4.3 (Mac OSX))
		cd minia
		sh INSTALL

		## Requirements

		CMake 3.10+; see http://www.cmake.org/cmake/resources/software.html

		C++11 compiler; (g++ version>=4.7 (Linux), clang version>=4.3 (Mac OSX))


		# User manual

		Type `minia` without any arguments for usage instructions.

merci/merci.cpp

+78 −2

Original line number	Diff line number	Diff line
		@@ -374,6 +374,65 @@ static bool maybe_merge(uint64_t packed, connections_index_t &connections_index,
		}


		static void
		parse_unitig_header(string header, float& mean_abundance)
		{
		bool debug = false;
		if (debug) std::cout << "parsing unitig links for " << header << std::endl;
		std::stringstream stream(header);
		while(1) {
		string tok;
		stream >> tok;
		if(!stream)
		break;

		if (tok.size() < 3)
		// that's the id, skip it
		continue;

		string field = tok.substr(0,2);

		if (field == "km")
		{
		mean_abundance = atof(tok.substr(tok.find_last_of(':')+1).c_str());
		//std::cout << "unitig " << header << " mean abundance " << mean_abundance << std::endl;
		}
		}
		}


		void renumber_glue_file(string glue_filename, uint64_t nb_out_tigs)
		{
		{
		std::ifstream infile(glue_filename);
		std::ofstream outfile(glue_filename+".tmp");
		std::string line;
		uint64_t counter = 1;
		while (std::getline(infile, line))
		{
		if (line[0] == '>')
		{
		size_t space_pos = line.find(' ');
		/* // yolo
		if (space_pos >= line.size())
		{
		std::cout << "error: no space in this glue file header (" << line << ") contact a developer." << std::endl;
		exit(1);
		}
		*/
		auto end_header = line.substr(space_pos);
		string new_header = ">" + std::to_string(nb_out_tigs+counter) + end_header;
		outfile << new_header << std::endl;
		counter++;
		}
		else
		outfile << line << std::endl;
		}
		} // closes files
		file_copy(glue_filename+".tmp",glue_filename);
		System::file().remove (glue_filename+".tmp");
		}

		static void
		extend_assembly_with_connections(const string assembly, int k, int nb_threads, bool verbose, connections_index_t &connections_index, connections_t &connections, BankFasta &out, BankFasta &glue)
		{
		@@ -416,6 +475,13 @@ extend_assembly_with_connections(const string assembly, int k, int nb_threads, b
		s._comment = string(lmark?"1":"0")+string(rmark?"1":"0"); //We set the sequence comment.
		s._comment += " ";

		// add coverage information
		float mean_abundance;
		parse_unitig_header(comment,mean_abundance);
		uint nb_kmers = seq.size() - k + 1;
		for (uint i = 0; i < nb_kmers; i++)
		s._comment += std::to_string((uint)mean_abundance) + " ";

		if (lmark \|\| rmark)
		glue.insert(s);
		else
		@@ -438,7 +504,8 @@ void merci(int k, string reads, string assembly, int nb_threads, bool verbose)
		string linked_assembly = assembly + ".linked";
		file_copy(assembly, linked_assembly);
		uint64_t nb_tigs = 0;
		link_tigs<span>( linked_assembly, k, nb_threads, nb_tigs, verbose);
		bool renumber_unitigs = true; // let's allow the input to be anything. Here it doesn't amtter much. We anyway renumber at the end
		link_tigs<span>( linked_assembly, k, nb_threads, nb_tigs, verbose, renumber_unitigs);

		// real trick here
		// tigs of length exactly k are annoying, they need to be handled carefully with UNITIG_BOTH positions
		@@ -463,11 +530,20 @@ void merci(int k, string reads, string assembly, int nb_threads, bool verbose)
		glue.flush();

		// glue what needs to be glued. magic, we're re-using bcalm code
		bglue<span> (nullptr /no storage/, assembly+".glue", k, 0, nb_threads, verbose);
		bglue<span> (nullptr /no storage/, assembly+".glue", k, 0, nb_threads, false, verbose);

		// renumber the .glue file just to avoid ID collision with .merci file
		renumber_glue_file(assembly+".glue", nb_tigs );

		// append glued to merci
		out.flush();
		file_append(assembly+".merci", assembly+".glue");

		// bglue drop links so let's recreate them
		k += 1;
		file_copy(assembly+".merci", assembly+".merci.b4link");
		renumber_unitigs = true; // here it's absolutely mandatory to renumber if we want the output to be processed by minia
		link_tigs<span>( assembly+".merci", k, nb_threads, nb_tigs, verbose, false, renumber_unitigs);
		}

		class Merci : public gatb::core::tools::misc::impl::Tool

src/Minia.cpp

+3 −4

Original line number	Diff line number	Diff line
		@@ -154,8 +154,7 @@ struct MiniaFunctor { void operator () (Parameter parameter)
		// link contigs
		uint nb_threads = 1; // doesn't matter because for now link_tigs is single-threaded
		bool verbose = true;
		link_tigs<span>(output, minia.k, nb_threads, minia.nbContigs, verbose);

		link_tigs<span>(output, minia.k, nb_threads, minia.nbContigs, verbose, false);

		/** We gather some statistics. */
		minia.getInfo()->add (1, minia.getTimeInfo().getProperties("time"));
		@@ -274,8 +273,8 @@ string Minia::assemble (/*const, removed because Simplifications isn't const any
		graphSimplifications._bulgeLen_kAdd = getInput()->getDouble("-bulge-len-kadd");
		if (getParser()->saw("-bulge-altpath-kadd"))
		graphSimplifications._bulgeAltPath_kAdd = getInput()->getDouble("-bulge-altpath-kadd");
		if (getParser()->saw("-bulge-altpath-covMult"))
		graphSimplifications._bulgeAltPath_covMult = getInput()->getDouble("-bulge-altpath-covMult");
		if (getParser()->saw("-bulge-altpath-covmult"))
		graphSimplifications._bulgeAltPath_covMult = getInput()->getDouble("-bulge-altpath-covmult");

		if (getParser()->saw("-ec-len-kmult"))
		graphSimplifications._ecLen_kMult = getInput()->getDouble("-ec-len-kmult");

test/ERR039477.md5

+3 −3

Original line number	Diff line number	Diff line
		3732560f98d63897d2b7a122938d7a42 # osx CI
		037b126f9e37db1db55d23eadc40477d # gcc 7 blok-bok
		c6e5a2cf1b9c6246129ae4263da749cc # debian CI
		e92d66d1e5b7450e6f6d8f6cc1de24bf # osx CI
		3192031d3491f3488a210419c50b9d4d # gcc 7 blok-bok
		dc556ec0e91c9aad6c1a68e48a2d8456 # debian CI