#include <legacy_caller.hpp>

Inheritance diagram for vg::LegacyCaller:

Public Member Functions
	LegacyCaller (const PathPositionHandleGraph &graph, SupportBasedSnarlCaller &snarl_caller, SnarlManager &snarl_manager, const string &sample_name, const vector< string > &ref_paths={}, const vector< size_t > &ref_path_offsets={}, const vector< int > &ref_path_ploidies={})
virtual	~LegacyCaller ()
virtual bool	call_snarl (const Snarl &snarl)
	Call a given snarl, and print the output to out_stream.
virtual string	vcf_header (const PathHandleGraph &graph, const vector< string > &contigs, const vector< size_t > &contig_length_overrides={}) const
	Write the vcf header (version and contigs and basic info).
Public Member Functions inherited from vg::GraphCaller
	GraphCaller (SnarlCaller &snarl_caller, SnarlManager &snarl_manager)
virtual	~GraphCaller ()
virtual void	call_top_level_snarls (const HandleGraph &graph, RecurseType recurse_type=RecurseOnFail)
virtual void	call_top_level_chains (const HandleGraph &graph, size_t max_edges, size_t max_trivial, RecurseType recurise_type=RecurseOnFail)
void	set_show_progress (bool show_progress)
	toggle progress messages
Public Member Functions inherited from vg::VCFOutputCaller
	VCFOutputCaller (const string &sample_name)
virtual	~VCFOutputCaller ()
bool	add_variant (vcflib::Variant &var) const
void	write_variants (ostream &out_stream, const SnarlManager *snarl_manager=nullptr)
void	vcf_fixup (vcflib::Variant &var) const
	Run vcffixup from vcflib.
void	set_translation (const unordered_map< nid_t, pair< string, size_t > > *translation)
	Add a translation map.
void	set_nested (bool nested)
	Assume writing nested snarls is enabled.
void	set_allele_merge (double threshold, int64_t min_len)
unordered_set< string >	get_output_contigs () const
string	prune_header_contigs (const string &header, const unordered_set< string > &keep) const

Protected Member Functions
pair< vector< SnarlTraversal >, vector< int > >	top_down_genotype (const Snarl &snarl, TraversalFinder &trav_finder, int ploidy, const string &ref_path_name, pair< size_t, size_t > ref_interval) const
SnarlTraversal	get_reference_traversal (const Snarl &snarl, TraversalFinder &trav_finder) const
	we need the reference traversal for VCF, but if the ref is not called, the above method won't find it.
tuple< vector< SnarlTraversal >, vector< int >, unique_ptr< SnarlCaller::CallInfo > >	re_genotype (const Snarl &snarl, TraversalFinder &trav_finder, const vector< SnarlTraversal > &in_traversals, const vector< int > &in_genotype, int ploidy, const string &ref_path_name, pair< size_t, size_t > ref_interval) const
bool	is_traversable (const Snarl &snarl)
	check if a site can be handled by the RepresentativeTraversalFinder
pair< string, PathIndex * >	find_index (const Snarl &snarl, const vector< PathIndex * > path_indexes) const
	look up a path index for a site and return its name too
Protected Member Functions inherited from vg::GraphCaller
vector< Chain >	break_chain (const HandleGraph &graph, const Chain &chain, size_t max_edges, size_t max_trivial)
	Break up a chain into bits that we want to call using size heuristics.
Protected Member Functions inherited from vg::VCFOutputCaller
void	add_allele_path_to_info (const HandleGraph *graph, vcflib::Variant &v, int allele, const Traversal &trav, bool reversed, bool one_based) const
	add a traversal to the VCF info field in the format of a GFA W-line or GAF path
void	add_allele_path_to_info (vcflib::Variant &v, int allele, const SnarlTraversal &trav, bool reversed, bool one_based) const
	legacy version of above
string	trav_string (const HandleGraph &graph, const SnarlTraversal &trav) const
	convert a traversal into an allele string
bool	merge_similar_alleles (const PathPositionHandleGraph &graph, const vector< SnarlTraversal > &site_traversals, vector< int > &site_genotype, const string &sample_name, vcflib::Variant &out_variant) const
bool	emit_variant (const PathPositionHandleGraph &graph, SnarlCaller &snarl_caller, const Snarl &snarl, const vector< SnarlTraversal > &called_traversals, const vector< int > &genotype, int ref_trav_idx, const unique_ptr< SnarlCaller::CallInfo > &call_info, const string &ref_path_name, int ref_offset, bool genotype_snarls, int ploidy, function< string(const vector< SnarlTraversal > &, const vector< int > &, int, int, int)> trav_to_string=nullptr)
tuple< int64_t, int64_t, bool, step_handle_t, step_handle_t >	get_ref_interval (const PathPositionHandleGraph &graph, const Snarl &snarl, const string &ref_path_name) const
pair< string, int64_t >	get_ref_position (const PathPositionHandleGraph &graph, const Snarl &snarl, const string &ref_path_name, int64_t ref_path_offset) const
	used for making gaf traversal names
void	flatten_common_allele_ends (vcflib::Variant &variant, bool backward, size_t len_override) const
string	print_snarl (const HandleGraph *grpah, const handle_t &snarl_start, const handle_t &snarl_end, bool in_brackets=false) const
string	print_snarl (const Snarl &snarl, bool in_brackets=false) const
	legacy version of above
string	print_flipped_snarl (const Snarl &snarl, bool in_brackets=false) const
	The same as above, but print the snarl as if its orientation has been flipped.
void	scan_snarl (const string &allele_string, function< void(const string &, Snarl &)> callback) const
void	update_nesting_info_tags (const SnarlManager *snarl_manager)

Protected Attributes
const PathPositionHandleGraph &	graph
	the graph
bool	is_vg
RepresentativeTraversalFinder *	traversal_finder
vector< PathIndex * >	path_indexes
	Needed by above (only used when working on vg inputs – generated on the fly otherwise).
vector< string >	ref_paths
	keep track of the reference paths
map< string, size_t >	ref_offsets
	keep track of offsets in the reference paths
map< string, int >	ref_ploidies
	keep track of ploidies in the reference paths
int	max_search_depth = 1000
	Tuning.
int	max_search_width = 1000
size_t	max_bubble_paths = 100
Protected Attributes inherited from vg::GraphCaller
SnarlCaller &	snarl_caller
	Our Genotyper.
SnarlManager &	snarl_manager
	Our snarls.
bool	show_progress
	Toggle progress messages.
Protected Attributes inherited from vg::VCFOutputCaller
vcflib::VariantCallFile	output_vcf
	output vcf
string	sample_name
	Sample name.
vector< vector< pair< pair< string, size_t >, string > > >	output_variants
vector< unordered_map< string, SuppressedRef > >	suppressed_ref_info
size_t	max_uncalled_alleles = 5
	print up to this many uncalled alleles when doing ref-genotpes in -a mode
const unordered_map< nid_t, pair< string, size_t > > *	translation
bool	include_nested
double	allele_merge_threshold = 1.0
int64_t	allele_merge_min_len = 0

Additional Inherited Members
Public Types inherited from vg::GraphCaller
enum	RecurseType { RecurseOnFail , RecurseAlways , RecurseNever }
Static Protected Member Functions inherited from vg::VCFOutputCaller
static bool	snarl_traversal_to_handles (const HandleGraph &graph, const SnarlTraversal &trav, Traversal &out_trav)
static int64_t	allele_core_length (const vector< string > &alleles)
static string	nesting_info_headers ()
Static Protected Attributes inherited from vg::VCFOutputCaller
static const int64_t	max_vcf_line_length = 2000000000

Detailed Description

LegacyCaller : Preserves (most of) the old vg call logic by using the RepresentativeTraversalFinder to recursively find traversals through arbitrary sites.

Constructor & Destructor Documentation

◆ LegacyCaller()

vg::LegacyCaller::LegacyCaller	(	const PathPositionHandleGraph &	graph,
		SupportBasedSnarlCaller &	snarl_caller,
		SnarlManager &	snarl_manager,
		const string &	sample_name,
		const vector< string > &	ref_paths = {},
		const vector< size_t > &	ref_path_offsets = {},
		const vector< int > &	ref_path_ploidies = {} )

◆ ~LegacyCaller()

vg::LegacyCaller::~LegacyCaller ( )

virtual

Member Function Documentation

◆ call_snarl()

bool vg::LegacyCaller::call_snarl ( const Snarl & snarl )

virtual

Call a given snarl, and print the output to out_stream.

Implements vg::GraphCaller.

◆ find_index()

pair< string, PathIndex * > vg::LegacyCaller::find_index	(	const Snarl &	snarl,
		const vector< PathIndex * >	path_indexes ) const

protected

look up a path index for a site and return its name too

◆ get_reference_traversal()

SnarlTraversal vg::LegacyCaller::get_reference_traversal	(	const Snarl &	snarl,
		TraversalFinder &	trav_finder ) const

protected

we need the reference traversal for VCF, but if the ref is not called, the above method won't find it.

◆ is_traversable()

bool vg::LegacyCaller::is_traversable ( const Snarl & snarl )

protected

check if a site can be handled by the RepresentativeTraversalFinder

◆ re_genotype()

tuple< vector< SnarlTraversal >, vector< int >, unique_ptr< SnarlCaller::CallInfo > > vg::LegacyCaller::re_genotype	(	const Snarl &	snarl,
		TraversalFinder &	trav_finder,
		const vector< SnarlTraversal > &	in_traversals,
		const vector< int > &	in_genotype,
		int	ploidy,
		const string &	ref_path_name,
		pair< size_t, size_t >	ref_interval ) const

protected

re-genotype output of top_down_genotype. it may give slightly different results as it's working with fully-defined traversals and can exactly determine lengths and supports it will also make sure the reference traversal is in the beginning of the output

◆ top_down_genotype()

pair< vector< SnarlTraversal >, vector< int > > vg::LegacyCaller::top_down_genotype	(	const Snarl &	snarl,
		TraversalFinder &	trav_finder,
		int	ploidy,
		const string &	ref_path_name,
		pair< size_t, size_t >	ref_interval ) const

protected

recursively genotype a snarl todo: can this be pushed to a more generic class?

◆ vcf_header()

string vg::LegacyCaller::vcf_header	(	const PathHandleGraph &	graph,
		const vector< string > &	contigs,
		const vector< size_t > &	contig_length_overrides = {} ) const

virtual

Write the vcf header (version and contigs and basic info).

Reimplemented from vg::VCFOutputCaller.

Member Data Documentation

◆ graph

const PathPositionHandleGraph& vg::LegacyCaller::graph

protected

the graph

◆ is_vg

bool vg::LegacyCaller::is_vg

protected

non-vg inputs are converted into vg as-needed, at least until we get the traversal finding ported

◆ max_bubble_paths

size_t vg::LegacyCaller::max_bubble_paths = 100

protected

What's the maximum number of bubble path combinations we can explore while finding one with maximum support?

◆ max_search_depth

int vg::LegacyCaller::max_search_depth = 1000

protected

Tuning.

How many nodes should we be willing to look at on our path back to the primary path? Keep in mind we need to look at all valid paths (and all combinations thereof) until we find a valid pair.

◆ max_search_width

int vg::LegacyCaller::max_search_width = 1000

protected

How many search states should we allow on the DFS stack when searching for traversals?

◆ path_indexes

vector<PathIndex*> vg::LegacyCaller::path_indexes

protected

Needed by above (only used when working on vg inputs – generated on the fly otherwise).

◆ ref_offsets

map<string, size_t> vg::LegacyCaller::ref_offsets

protected

keep track of offsets in the reference paths

◆ ref_paths

vector<string> vg::LegacyCaller::ref_paths

protected

keep track of the reference paths

◆ ref_ploidies

map<string, int> vg::LegacyCaller::ref_ploidies

protected

keep track of ploidies in the reference paths

◆ traversal_finder

RepresentativeTraversalFinder* vg::LegacyCaller::traversal_finder

protected

The old vg call traversal finder. It is fairly efficient but daunting to maintain. We keep it around until a better replacement is implemented. It is not compatible with the Handle Graph API because it relise on PathIndex. We convert to VG as needed in order to use it.

The documentation for this class was generated from the following files:

src/legacy_caller.hpp
src/legacy_caller.cpp

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ LegacyCaller()

◆ ~LegacyCaller()

Member Function Documentation

◆ call_snarl()

◆ find_index()

◆ get_reference_traversal()

◆ is_traversable()

◆ re_genotype()

◆ top_down_genotype()

◆ vcf_header()

Member Data Documentation

◆ graph

◆ is_vg

◆ max_bubble_paths

◆ max_search_depth

◆ max_search_width

◆ path_indexes

◆ ref_offsets

◆ ref_paths

◆ ref_ploidies

◆ traversal_finder