ripser.cpp

/*

 Ripser: a lean C++ code for computation of Vietoris-Rips persistence barcodes

 MIT License

 Copyright (c) 2015–2018 Ulrich Bauer

 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal
 in the Software without restriction, including without limitation the rights
 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 copies of the Software, and to permit persons to whom the Software is
 furnished to do so, subject to the following conditions:

 The above copyright notice and this permission notice shall be included in all
 copies or substantial portions of the Software.

 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.

 You are under no obligation whatsoever to provide any bug fixes, patches, or
 upgrades to the features, functionality or performance of the source code
 ("Enhancements") to anyone; however, if you choose to make your Enhancements
 available either publicly, or directly to the author of this software, without
 imposing a separate written license agreement for such Enhancements, then you
 hereby grant the following license: a non-exclusive, royalty-free perpetual
 license to install, use, modify, prepare derivative works, incorporate into
 other computer software, distribute, and sublicense such enhancements or
 derivative works thereof, in binary and source code form.

*/

//#define ASSEMBLE_REDUCTION_MATRIX
//#define USE_COEFFICIENTS

//#define INDICATE_PROGRESS
#define PRINT_PERSISTENCE_PAIRS

//#define USE_GOOGLE_HASHMAP

#include <algorithm>
#include <cassert>
#include <cmath>
#include <fstream>
#include <iostream>
#include <numeric>
#include <queue>
#include <sstream>
#include <unordered_map>

#ifdef USE_GOOGLE_HASHMAP
#include <sparsehash/sparse_hash_map>
template <class Key, class T> class hash_map : public google::sparse_hash_map<Key, T> {
public:
	inline void reserve(size_t hint) { this->resize(hint); }
};
#else
template <class Key, class T> class hash_map : public std::unordered_map<Key, T> {};
#endif

typedef float value_t;
typedef int64_t index_t;
typedef int16_t coefficient_t;

class binomial_coeff_table {
	std::vector<std::vector<index_t>> B;

public:
	binomial_coeff_table(index_t n, index_t k) : B(n + 1) {
		for (index_t i = 0; i <= n; i++) {
			B[i].resize(k + 1);
			for (index_t j = 0; j <= std::min(i, k); j++)
				if (j == 0 || j == i)
					B[i][j] = 1;
				else
					B[i][j] = B[i - 1][j - 1] + B[i - 1][j];
		}
	}

	index_t operator()(index_t n, index_t k) const {
		assert(n < B.size() && k < B[n].size());
		return B[n][k];
	}
};

bool is_prime(const coefficient_t n) {
	if (!(n & 1) || n < 2) return n == 2;
	for (coefficient_t p = 3, q = n / p, r = n % p; p <= q; p += 2, q = n / p, r = n % p)
		if (!r) return false;
	return true;
}

std::vector<coefficient_t> multiplicative_inverse_vector(const coefficient_t m) {
	std::vector<coefficient_t> inverse(m);
	inverse[1] = 1;
	// m = a * (m / a) + m % a
	// Multipying with inverse(a) * inverse(m % a):
	// 0 = inverse(m % a) * (m / a)  + inverse(a)  (mod m)
	for (coefficient_t a = 2; a < m; ++a) inverse[a] = m - (inverse[m % a] * (m / a)) % m;
	return inverse;
}

#ifdef USE_COEFFICIENTS
struct __attribute__((packed)) entry_t {
	index_t index : 8 * (sizeof(index_t) - sizeof(coefficient_t));
	coefficient_t coefficient;
	entry_t(index_t _index, coefficient_t _coefficient)
	    : index(_index), coefficient(_coefficient) {}
	entry_t(index_t _index) : index(_index), coefficient(1) {}
	entry_t() : index(0), coefficient(1) {}
};

static_assert(sizeof(entry_t) == sizeof(index_t), "size of entry_t is not the same as index_t");

entry_t make_entry(index_t _index, coefficient_t _coefficient) {
	return entry_t(_index, _coefficient);
}
index_t get_index(const entry_t& e) { return e.index; }
index_t get_coefficient(const entry_t& e) { return e.coefficient; }
void set_coefficient(entry_t& e, const coefficient_t c) { e.coefficient = c; }

bool operator==(const entry_t& e1, const entry_t& e2) {
	return get_index(e1) == get_index(e2) && get_coefficient(e1) == get_coefficient(e2);
}

std::ostream& operator<<(std::ostream& stream, const entry_t& e) {
	stream << get_index(e) << ":" << get_coefficient(e);
	return stream;
}

#else

typedef index_t entry_t;
const index_t get_index(const entry_t& i) { return i; }
index_t get_coefficient(const entry_t& i) { return 1; }
entry_t make_entry(index_t _index, coefficient_t _value) { return entry_t(_index); }
void set_coefficient(entry_t& e, const coefficient_t c) {}

#endif

const entry_t& get_entry(const entry_t& e) { return e; }

typedef std::pair<value_t, index_t> diameter_index_t;
value_t get_diameter(const diameter_index_t& i) { return i.first; }
index_t get_index(const diameter_index_t& i) { return i.second; }

typedef std::pair<index_t, value_t> index_diameter_t;
index_t get_index(const index_diameter_t& i) { return i.first; }
value_t get_diameter(const index_diameter_t& i) { return i.second; }

class diameter_entry_t : public std::pair<value_t, entry_t> {
public:
	diameter_entry_t() {}
	diameter_entry_t(const entry_t& e) : std::pair<value_t, entry_t>(0, e) {}
	diameter_entry_t(value_t _diameter, index_t _index, coefficient_t _coefficient)
	    : std::pair<value_t, entry_t>(_diameter, make_entry(_index, _coefficient)) {}
	diameter_entry_t(const diameter_index_t& _diameter_index, coefficient_t _coefficient)
	    : std::pair<value_t, entry_t>(get_diameter(_diameter_index),
	                                  make_entry(get_index(_diameter_index), _coefficient)) {}
	diameter_entry_t(const diameter_index_t& _diameter_index)
	    : diameter_entry_t(_diameter_index, 1) {}
};

const entry_t& get_entry(const diameter_entry_t& p) { return p.second; }
entry_t& get_entry(diameter_entry_t& p) { return p.second; }
const index_t get_index(const diameter_entry_t& p) { return get_index(get_entry(p)); }
const coefficient_t get_coefficient(const diameter_entry_t& p) {
	return get_coefficient(get_entry(p));
}
const value_t& get_diameter(const diameter_entry_t& p) { return p.first; }
void set_coefficient(diameter_entry_t& p, const coefficient_t c) {
	set_coefficient(get_entry(p), c);
}

template <typename Entry> struct greater_diameter_or_smaller_index {
	bool operator()(const Entry& a, const Entry& b) {
		return (get_diameter(a) > get_diameter(b)) ||
		       ((get_diameter(a) == get_diameter(b)) && (get_index(a) < get_index(b)));
	}
};

enum compressed_matrix_layout { LOWER_TRIANGULAR, UPPER_TRIANGULAR };

template <compressed_matrix_layout Layout> class compressed_distance_matrix {
public:
	std::vector<value_t> distances;
	std::vector<value_t*> rows;

	void init_rows();

	compressed_distance_matrix(std::vector<value_t>&& _distances)
	    : distances(std::move(_distances)), rows((1 + std::sqrt(1 + 8 * distances.size())) / 2) {
		assert(distances.size() == size() * (size() - 1) / 2);
		init_rows();
	}

	template <typename DistanceMatrix>
	compressed_distance_matrix(const DistanceMatrix& mat)
	    : distances(mat.size() * (mat.size() - 1) / 2), rows(mat.size()) {
		init_rows();

		for (index_t i = 1; i < size(); ++i)
			for (index_t j = 0; j < i; ++j) rows[i][j] = mat(i, j);
	}

	value_t operator()(const index_t i, const index_t j) const;

	size_t size() const { return rows.size(); }
};

class sparse_distance_matrix {
public:
	std::vector<std::vector<index_diameter_t>> neighbors;

	index_t num_edges;

	sparse_distance_matrix(std::vector<std::vector<index_diameter_t>>&& _neighbors,
	                       index_t _num_edges)
	    : neighbors(std::move(_neighbors)), num_edges(_num_edges) {}

	template <typename DistanceMatrix>
	sparse_distance_matrix(const DistanceMatrix& mat, value_t threshold)
	    : neighbors(mat.size()), num_edges(0) {

		for (index_t i = 0; i < size(); ++i)
			for (index_t j = 0; j < size(); ++j)
				if (i != j && mat(i, j) <= threshold) {
					++num_edges;
					neighbors[i].push_back(std::make_pair(j, mat(i, j)));
				}
	}

	size_t size() const { return neighbors.size(); }
};

template <> void compressed_distance_matrix<LOWER_TRIANGULAR>::init_rows() {
	value_t* pointer = &distances[0];
	for (index_t i = 1; i < size(); ++i) {
		rows[i] = pointer;
		pointer += i;
	}
}

template <> void compressed_distance_matrix<UPPER_TRIANGULAR>::init_rows() {
	value_t* pointer = &distances[0] - 1;
	for (index_t i = 0; i < size() - 1; ++i) {
		rows[i] = pointer;
		pointer += size() - i - 2;
	}
}

template <>
value_t compressed_distance_matrix<UPPER_TRIANGULAR>::operator()(const index_t i,
                                                                 const index_t j) const {
	return i == j ? 0 : i > j ? rows[j][i] : rows[i][j];
}

template <>
value_t compressed_distance_matrix<LOWER_TRIANGULAR>::operator()(const index_t i,
                                                                 const index_t j) const {
	return i == j ? 0 : i < j ? rows[j][i] : rows[i][j];
}

typedef compressed_distance_matrix<LOWER_TRIANGULAR> compressed_lower_distance_matrix;
typedef compressed_distance_matrix<UPPER_TRIANGULAR> compressed_upper_distance_matrix;

class euclidean_distance_matrix {
public:
	std::vector<std::vector<value_t>> points;

	euclidean_distance_matrix(std::vector<std::vector<value_t>>&& _points)
	    : points(std::move(_points)) {
		for (auto p : points) { assert(p.size() == points.front().size()); }
	}

	value_t operator()(const index_t i, const index_t j) const {
		assert(i < points.size());
		assert(j < points.size());
		return std::sqrt(std::inner_product(
		    points[i].begin(), points[i].end(), points[j].begin(), value_t(), std::plus<value_t>(),
		    [](value_t u, value_t v) { return (u - v) * (u - v); }));
	}

	size_t size() const { return points.size(); }
};

class union_find {
	std::vector<index_t> parent;
	std::vector<uint8_t> rank;

public:
	union_find(index_t n) : parent(n), rank(n, 0) {
		for (index_t i = 0; i < n; ++i) parent[i] = i;
	}

	index_t find(index_t x) {
		index_t y = x, z;
		while ((z = parent[y]) != y) y = z;
		while ((z = parent[x]) != y) {
			parent[x] = y;
			x = z;
		}
		return z;
	}
	void link(index_t x, index_t y) {
		if ((x = find(x)) == (y = find(y))) return;
		if (rank[x] > rank[y])
			parent[y] = x;
		else {
			parent[x] = y;
			if (rank[x] == rank[y]) ++rank[y];
		}
	}
};

template <typename Heap> diameter_entry_t pop_pivot(Heap& column, coefficient_t modulus) {
	if (column.empty())
		return diameter_entry_t(-1);
	else {
		auto pivot = column.top();

#ifdef USE_COEFFICIENTS
		coefficient_t coefficient = 0;
		do {
			coefficient = (coefficient + get_coefficient(column.top())) % modulus;
			column.pop();

			if (coefficient == 0) {
				if (column.empty())
					return diameter_entry_t(-1);
				else
					pivot = column.top();
			}
		} while (!column.empty() && get_index(column.top()) == get_index(pivot));
		if (get_index(pivot) != -1) { set_coefficient(pivot, coefficient); }
#else
		column.pop();
		while (!column.empty() && get_index(column.top()) == get_index(pivot)) {
			column.pop();
			if (column.empty())
				return diameter_entry_t(-1);
			else {
				pivot = column.top();
				column.pop();
			}
		}
#endif
		return pivot;
	}
}

template <typename Heap> diameter_entry_t get_pivot(Heap& column, coefficient_t modulus) {
	diameter_entry_t result = pop_pivot(column, modulus);
	if (get_index(result) != -1) column.push(result);
	return result;
}

template <typename ValueType> class compressed_sparse_matrix {
	std::vector<size_t> bounds;
	std::vector<ValueType> entries;

public:
	size_t size() const { return bounds.size(); }

	typename std::vector<ValueType>::const_iterator cbegin(size_t index) const {
		assert(index < size());
		return index == 0 ? entries.cbegin() : entries.cbegin() + bounds[index - 1];
	}

	typename std::vector<ValueType>::const_iterator cend(size_t index) const {
		assert(index < size());
		return entries.cbegin() + bounds[index];
	}

	template <typename Iterator> void append_column(Iterator begin, Iterator end) {
		for (Iterator it = begin; it != end; ++it) { entries.push_back(*it); }
		bounds.push_back(entries.size());
	}

	void append_column() { bounds.push_back(entries.size()); }

	void push_back(ValueType e) {
		assert(0 < size());
		entries.push_back(e);
		++bounds.back();
	}

	void pop_back() {
		assert(0 < size());
		entries.pop_back();
		--bounds.back();
	}

	template <typename Collection> void append_column(const Collection collection) {
		append_column(collection.cbegin(), collection.cend());
	}
};

template <typename DistanceMatrix> class ripser {
	DistanceMatrix dist;
	index_t n, dim_max;
	value_t threshold;
	float ratio;
	coefficient_t modulus;
	const binomial_coeff_table binomial_coeff;
	std::vector<coefficient_t> multiplicative_inverse;
	mutable std::vector<index_t> vertices;
	mutable std::vector<std::vector<index_diameter_t>::const_reverse_iterator> neighbor_it;
	mutable std::vector<std::vector<index_diameter_t>::const_reverse_iterator> neighbor_end;
	mutable std::vector<diameter_entry_t> coface_entries;

public:
	ripser(DistanceMatrix&& _dist, index_t _dim_max, value_t _threshold, float _ratio,
	       coefficient_t _modulus)
	    : dist(std::move(_dist)), n(dist.size()),
	      dim_max(std::min(_dim_max, index_t(dist.size() - 2))), threshold(_threshold),
	      ratio(_ratio), modulus(_modulus), binomial_coeff(n, dim_max + 2),
	      multiplicative_inverse(multiplicative_inverse_vector(_modulus)) {}

	index_t get_next_vertex(index_t& v, const index_t idx, const index_t k) const {
		if (binomial_coeff(v, k) > idx) {
			index_t count = v;
			while (count > 0) {
				index_t i = v;
				index_t step = count >> 1;
				i -= step;
				if (binomial_coeff(i, k) > idx) {
					v = --i;
					count -= step + 1;
				} else
					count = step;
			}
		}
		assert(binomial_coeff(v, k) <= idx && binomial_coeff(v + 1, k) > idx);
		return v;
	}

	index_t get_edge_index(const index_t i, const index_t j) const {
		return binomial_coeff(i, 2) + j;
	}

	template <typename OutputIterator>
	OutputIterator get_simplex_vertices(index_t idx, const index_t dim, index_t v,
	                                    OutputIterator out) const {
		--v;
		for (index_t k = dim + 1; k > 0; --k) {
			get_next_vertex(v, idx, k);
			*out++ = v;
			idx -= binomial_coeff(v, k);
		}
		return out;
	}

	value_t compute_diameter(const index_t index, index_t dim) const {
		value_t diam = -std::numeric_limits<value_t>::infinity();

		vertices.clear();
		get_simplex_vertices(index, dim, dist.size(), std::back_inserter(vertices));

		for (index_t i = 0; i <= dim; ++i)
			for (index_t j = 0; j < i; ++j) {
				diam = std::max(diam, dist(vertices[i], vertices[j]));
			}
		return diam;
	}

	class simplex_coboundary_enumerator;

	void assemble_columns_to_reduce(std::vector<diameter_index_t>& simplices,
	                                std::vector<diameter_index_t>& columns_to_reduce,
	                                hash_map<index_t, index_t>& pivot_column_index, index_t dim);

	void compute_dim_0_pairs(std::vector<diameter_index_t>& edges,
	                         std::vector<diameter_index_t>& columns_to_reduce) {
		union_find dset(n);

		edges = get_edges();

		std::sort(edges.rbegin(), edges.rend(),
		          greater_diameter_or_smaller_index<diameter_index_t>());

#ifdef PRINT_PERSISTENCE_PAIRS
		std::cout << "persistence intervals in dim 0:" << std::endl;
#endif

		std::vector<index_t> vertices_of_edge(2);
		for (auto e : edges) {
			vertices_of_edge.clear();
			get_simplex_vertices(get_index(e), 1, n, std::back_inserter(vertices_of_edge));
			index_t u = dset.find(vertices_of_edge[0]), v = dset.find(vertices_of_edge[1]);

			if (u != v) {
#ifdef PRINT_PERSISTENCE_PAIRS
				if (get_diameter(e) != 0)
					std::cout << " [0," << get_diameter(e) << ")" << std::endl;
#endif
				dset.link(u, v);
			} else
				columns_to_reduce.push_back(e);
		}
		std::reverse(columns_to_reduce.begin(), columns_to_reduce.end());

#ifdef PRINT_PERSISTENCE_PAIRS
		for (index_t i = 0; i < n; ++i)
			if (dset.find(i) == i) std::cout << " [0, )" << std::endl << std::flush;
#endif
	}

	template <typename Column, typename Iterator>
	diameter_entry_t add_coboundary_and_get_pivot(Iterator column_begin, Iterator column_end,
	                                              coefficient_t factor_column_to_add,
#ifdef ASSEMBLE_REDUCTION_MATRIX
	                                              Column& working_reduction_column,
#endif
	                                              Column& working_coboundary, const index_t& dim,
	                                              hash_map<index_t, index_t>& pivot_column_index,
	                                              bool& might_be_apparent_pair);

	void compute_pairs(std::vector<diameter_index_t>& columns_to_reduce,
	                   hash_map<index_t, index_t>& pivot_column_index, index_t dim) {

#ifdef PRINT_PERSISTENCE_PAIRS
		std::cout << "persistence intervals in dim " << dim << ":" << std::endl;
#endif

#ifdef ASSEMBLE_REDUCTION_MATRIX
		compressed_sparse_matrix<diameter_entry_t> reduction_matrix;
#else
#ifdef USE_COEFFICIENTS
		std::vector<diameter_entry_t> reduction_matrix;
#endif
#endif

		std::vector<diameter_entry_t> coface_entries;

		for (index_t index_column_to_reduce = 0; index_column_to_reduce < columns_to_reduce.size();
		     ++index_column_to_reduce) {
			auto column_to_reduce = columns_to_reduce[index_column_to_reduce];

#ifdef ASSEMBLE_REDUCTION_MATRIX
			std::priority_queue<diameter_entry_t, std::vector<diameter_entry_t>,
			                    greater_diameter_or_smaller_index<diameter_entry_t>>
			    working_reduction_column;
#endif

			std::priority_queue<diameter_entry_t, std::vector<diameter_entry_t>,
			                    greater_diameter_or_smaller_index<diameter_entry_t>>
			    working_coboundary;

			value_t diameter = get_diameter(column_to_reduce);

#ifdef INDICATE_PROGRESS
			if ((index_column_to_reduce + 1) % 1000000 == 0)
				std::cout << "\033[K"
				          << "reducing column " << index_column_to_reduce + 1 << "/"
				          << columns_to_reduce.size() << " (diameter " << diameter << ")"
				          << std::flush << "\r";
#endif

			index_t index_column_to_add = index_column_to_reduce;

			diameter_entry_t pivot;

			// start with factor 1 in order to initialize working_coboundary
			// with the coboundary of the simplex with index column_to_reduce
			coefficient_t factor_column_to_add = 1;

#ifdef ASSEMBLE_REDUCTION_MATRIX
			// initialize reduction_matrix as identity matrix
			reduction_matrix.append_column();
#endif
#ifdef USE_COEFFICIENTS
			reduction_matrix.push_back(diameter_entry_t(column_to_reduce, 1));
#endif

			bool might_be_apparent_pair = (index_column_to_reduce == index_column_to_add);

			while (true) {
#ifdef ASSEMBLE_REDUCTION_MATRIX
#ifdef USE_COEFFICIENTS
				auto reduction_column_begin = reduction_matrix.cbegin(index_column_to_add),
				     reduction_column_end = reduction_matrix.cend(index_column_to_add);
#else
				std::vector<diameter_entry_t> coeffs;
				coeffs.push_back(columns_to_reduce[index_column_to_add]);
				for (auto it = reduction_matrix.cbegin(index_column_to_add);
				     it != reduction_matrix.cend(index_column_to_add); ++it)
					coeffs.push_back(*it);
				auto reduction_column_begin = coeffs.begin(), reduction_column_end = coeffs.end();
#endif
#else
#ifdef USE_COEFFICIENTS
				auto reduction_column_begin = &reduction_matrix[index_column_to_add],
				     reduction_column_end = &reduction_matrix[index_column_to_add] + 1;
#else
				auto reduction_column_begin = &columns_to_reduce[index_column_to_add],
				     reduction_column_end = &columns_to_reduce[index_column_to_add] + 1;
#endif
#endif

				pivot = add_coboundary_and_get_pivot(
				    reduction_column_begin, reduction_column_end, factor_column_to_add,
#ifdef ASSEMBLE_REDUCTION_MATRIX
				    working_reduction_column,
#endif
				    working_coboundary, dim, pivot_column_index, might_be_apparent_pair);

				if (get_index(pivot) != -1) {
					auto pair = pivot_column_index.find(get_index(pivot));

					if (pair != pivot_column_index.end()) {
						index_column_to_add = pair->second;
						factor_column_to_add = modulus - get_coefficient(pivot);
					} else {
#ifdef PRINT_PERSISTENCE_PAIRS
						value_t death = get_diameter(pivot);
						if (death > diameter * ratio) {
#ifdef INDICATE_PROGRESS
							std::cout << "\033[K";
#endif
							std::cout << " [" << diameter << "," << death << ")" << std::endl
							          << std::flush;
						}
#endif
						pivot_column_index.insert(
						    std::make_pair(get_index(pivot), index_column_to_reduce));

#ifdef USE_COEFFICIENTS
						const coefficient_t inverse =
						    multiplicative_inverse[get_coefficient(pivot)];
#endif

#ifdef ASSEMBLE_REDUCTION_MATRIX
						// replace current column of reduction_matrix (with a single diagonal 1
						// entry) by reduction_column (possibly with a different entry on the
						// diagonal)
#ifdef USE_COEFFICIENTS
						reduction_matrix.pop_back();
#else
						pop_pivot(working_reduction_column, modulus);
#endif

						while (true) {
							diameter_entry_t e = pop_pivot(working_reduction_column, modulus);
							if (get_index(e) == -1) break;
#ifdef USE_COEFFICIENTS
							set_coefficient(e, inverse * get_coefficient(e) % modulus);
							assert(get_coefficient(e) > 0);
#endif
							reduction_matrix.push_back(e);
						}
#else
#ifdef USE_COEFFICIENTS
						reduction_matrix.pop_back();
						reduction_matrix.push_back(diameter_entry_t(column_to_reduce, inverse));
#endif
#endif
						break;
					}
				} else {
#ifdef PRINT_PERSISTENCE_PAIRS
					std::cout << " [" << diameter << ", )" << std::endl << std::flush;
#endif
					break;
				}
			}
		}

#ifdef INDICATE_PROGRESS
		std::cout << "\033[K";
#endif
	}

	std::vector<diameter_index_t> get_edges();

	void compute_barcodes() {

		std::vector<diameter_index_t> simplices, columns_to_reduce;

		compute_dim_0_pairs(simplices, columns_to_reduce);

		for (index_t dim = 1; dim <= dim_max; ++dim) {
			hash_map<index_t, index_t> pivot_column_index;
			pivot_column_index.reserve(columns_to_reduce.size());

			compute_pairs(columns_to_reduce, pivot_column_index, dim);

			if (dim < dim_max) {
				assemble_columns_to_reduce(simplices, columns_to_reduce, pivot_column_index,
				                           dim + 1);
			}
		}
	}
};

template <> class ripser<compressed_lower_distance_matrix>::simplex_coboundary_enumerator {
private:
	index_t idx_below, idx_above, v, k;
	std::vector<index_t> vertices;
	const diameter_entry_t simplex;
	const coefficient_t modulus;
	const compressed_lower_distance_matrix& dist;
	const binomial_coeff_table& binomial_coeff;

public:
	simplex_coboundary_enumerator(const diameter_entry_t _simplex, index_t _dim,
	                              const ripser& parent)
	    : idx_below(get_index(_simplex)), idx_above(0), v(parent.n - 1), k(_dim + 1),
	      vertices(_dim + 1), simplex(_simplex), modulus(parent.modulus), dist(parent.dist),
	      binomial_coeff(parent.binomial_coeff) {
		parent.get_simplex_vertices(get_index(_simplex), _dim, parent.n, vertices.begin());
	}

	bool has_next() {
		while ((v != -1) && (binomial_coeff(v, k) <= idx_below)) {
			idx_below -= binomial_coeff(v, k);
			idx_above += binomial_coeff(v, k + 1);
			--v;
			--k;
			assert(k != -1);
		}
		return v != -1;
	}

	diameter_entry_t next() {
		value_t coface_diameter = get_diameter(simplex);
		for (index_t w : vertices) coface_diameter = std::max(coface_diameter, dist(v, w));
		index_t coface_index = idx_above + binomial_coeff(v--, k + 1) + idx_below;
		coefficient_t coface_coefficient =
		    (k & 1 ? -1 + modulus : 1) * get_coefficient(simplex) % modulus;
		return diameter_entry_t(coface_diameter, coface_index, coface_coefficient);
	}
};

template <typename DistanceMatrix>
template <typename Column, typename Iterator>
diameter_entry_t ripser<DistanceMatrix>::add_coboundary_and_get_pivot(
    Iterator column_begin, Iterator column_end, coefficient_t factor_column_to_add,
#ifdef ASSEMBLE_REDUCTION_MATRIX
    Column& working_reduction_column,
#endif
    Column& working_coboundary, const index_t& dim, hash_map<index_t, index_t>& pivot_column_index,
    bool& might_be_apparent_pair) {
	for (auto it = column_begin; it != column_end; ++it) {
		diameter_entry_t simplex = *it;
		set_coefficient(simplex, get_coefficient(simplex) * factor_column_to_add % modulus);

#ifdef ASSEMBLE_REDUCTION_MATRIX
		working_reduction_column.push(simplex);
#endif

		coface_entries.clear();
		simplex_coboundary_enumerator cofaces(simplex, dim, *this);
		while (cofaces.has_next()) {
			diameter_entry_t coface = cofaces.next();
			if (get_diameter(coface) <= threshold) {
				coface_entries.push_back(coface);
				if (might_be_apparent_pair && (get_diameter(simplex) == get_diameter(coface))) {
					if (pivot_column_index.find(get_index(coface)) == pivot_column_index.end()) {
						return coface;
					}
					might_be_apparent_pair = false;
				}
			}
		}
		for (auto coface : coface_entries) working_coboundary.push(coface);
	}

	return get_pivot(working_coboundary, modulus);
}

template <> class ripser<sparse_distance_matrix>::simplex_coboundary_enumerator {
private:
	const ripser& parent;

	index_t idx_below, idx_above, v, k, max_vertex_below;
	const diameter_entry_t simplex;
	const coefficient_t modulus;
	const sparse_distance_matrix& dist;
	const binomial_coeff_table& binomial_coeff;

	std::vector<index_t>& vertices;
	std::vector<std::vector<index_diameter_t>::const_reverse_iterator>& neighbor_it;
	std::vector<std::vector<index_diameter_t>::const_reverse_iterator>& neighbor_end;
	index_diameter_t x;

public:
	simplex_coboundary_enumerator(const diameter_entry_t _simplex, index_t _dim,
	                              const ripser& _parent)
	    : parent(_parent), idx_below(get_index(_simplex)), idx_above(0), v(parent.n - 1),
	      k(_dim + 1), max_vertex_below(parent.n - 1), simplex(_simplex), modulus(parent.modulus),
	      dist(parent.dist), binomial_coeff(parent.binomial_coeff), vertices(parent.vertices),
	      neighbor_it(parent.neighbor_it), neighbor_end(parent.neighbor_end) {

		neighbor_it.clear();
		neighbor_end.clear();
		vertices.clear();

		parent.get_simplex_vertices(idx_below, _dim, parent.n, std::back_inserter(vertices));

		for (auto v : vertices) {
			neighbor_it.push_back(dist.neighbors[v].rbegin());
			neighbor_end.push_back(dist.neighbors[v].rend());
		}
	}

	bool has_next(bool all_cofaces = true) {
		for (auto &it0 = neighbor_it[0], &end0 = neighbor_end[0]; it0 != end0; ++it0) {
			x = *it0;
			for (size_t idx = 1; idx < neighbor_it.size(); ++idx) {
				auto &it = neighbor_it[idx], end = neighbor_end[idx];
				while (get_index(*it) > get_index(x))
					if (++it == end) return false;
				auto y = *it;
				if (get_index(y) != get_index(x))
					goto continue_outer;
				else
					x = std::max(x, y);
			}
			return all_cofaces || !(k > 0 && parent.get_next_vertex(max_vertex_below, idx_below,
			                                                        k) > get_index(x));
		continue_outer:;
		}
		return false;
	}

	diameter_entry_t next() {
		++neighbor_it[0];

		while (k > 0 && parent.get_next_vertex(max_vertex_below, idx_below, k) > get_index(x)) {
			idx_below -= binomial_coeff(max_vertex_below, k);
			idx_above += binomial_coeff(max_vertex_below, k + 1);
			--k;
		}

		value_t coface_diameter = std::max(get_diameter(simplex), get_diameter(x));

		coefficient_t coface_coefficient =
		    (k & 1 ? -1 + modulus : 1) * get_coefficient(simplex) % modulus;

		return diameter_entry_t(coface_diameter,
		                        idx_above + binomial_coeff(get_index(x), k + 1) + idx_below,
		                        coface_coefficient);
	}
};

template <> std::vector<diameter_index_t> ripser<compressed_lower_distance_matrix>::get_edges() {
	std::vector<diameter_index_t> edges;
	for (index_t index = binomial_coeff(n, 2); index-- > 0;) {
		value_t diameter = compute_diameter(index, 1);
		if (diameter <= threshold) edges.push_back(std::make_pair(diameter, index));
	}
	return edges;
}

template <> std::vector<diameter_index_t> ripser<sparse_distance_matrix>::get_edges() {
	std::vector<diameter_index_t> edges;
	for (index_t i = 0; i < n; ++i)
		for (auto n : dist.neighbors[i]) {
			index_t j = get_index(n);
			if (i > j) edges.push_back(std::make_pair(get_diameter(n), get_edge_index(i, j)));
		}
	return edges;
}

template <>
void ripser<compressed_lower_distance_matrix>::assemble_columns_to_reduce(
    std::vector<diameter_index_t>& simplices, std::vector<diameter_index_t>& columns_to_reduce,
    hash_map<index_t, index_t>& pivot_column_index, index_t dim) {
	index_t num_simplices = binomial_coeff(n, dim + 1);

	columns_to_reduce.clear();

#ifdef INDICATE_PROGRESS
	std::cout << "\033[K"
	          << "assembling " << num_simplices << " columns" << std::flush << "\r";
#endif

	for (index_t index = 0; index < num_simplices; ++index) {
		if (pivot_column_index.find(index) == pivot_column_index.end()) {
			value_t diameter = compute_diameter(index, dim);
			if (diameter <= threshold) columns_to_reduce.push_back(std::make_pair(diameter, index));
#ifdef INDICATE_PROGRESS
			if ((index + 1) % 1000000 == 0)
				std::cout << "\033[K"
				          << "assembled " << columns_to_reduce.size() << " out of " << (index + 1)
				          << "/" << num_simplices << " columns" << std::flush << "\r";
#endif
		}
	}

#ifdef INDICATE_PROGRESS
	std::cout << "\033[K"
	          << "sorting " << num_simplices << " columns" << std::flush << "\r";
#endif

	std::sort(columns_to_reduce.begin(), columns_to_reduce.end(),
	          greater_diameter_or_smaller_index<diameter_index_t>());
#ifdef INDICATE_PROGRESS
	std::cout << "\033[K";
#endif
}

template <>
void ripser<sparse_distance_matrix>::assemble_columns_to_reduce(
    std::vector<diameter_index_t>& simplices, std::vector<diameter_index_t>& columns_to_reduce,
    hash_map<index_t, index_t>& pivot_column_index, index_t dim) {

#ifdef INDICATE_PROGRESS
	std::cout << "\033[K"
	          << "assembling columns" << std::flush << "\r";
#endif

	--dim;
	columns_to_reduce.clear();

	std::vector<diameter_index_t> next_simplices;

	for (diameter_index_t simplex : simplices) {
		simplex_coboundary_enumerator cofaces(simplex, dim, *this);

		while (cofaces.has_next(false)) {
			auto coface = cofaces.next();

			next_simplices.push_back(std::make_pair(get_diameter(coface), get_index(coface)));

			if (pivot_column_index.find(get_index(coface)) == pivot_column_index.end())
				columns_to_reduce.push_back(
				    std::make_pair(get_diameter(coface), get_index(coface)));
		}
	}

	simplices.swap(next_simplices);

#ifdef INDICATE_PROGRESS
	std::cout << "\033[K"
	          << "sorting " << columns_to_reduce.size() << " columns" << std::flush << "\r";
#endif

	std::sort(columns_to_reduce.begin(), columns_to_reduce.end(),
	          greater_diameter_or_smaller_index<diameter_index_t>());
#ifdef INDICATE_PROGRESS
	std::cout << "\033[K";
#endif
}

enum file_format {
	LOWER_DISTANCE_MATRIX,
	UPPER_DISTANCE_MATRIX,
	DISTANCE_MATRIX,
	POINT_CLOUD,
	DIPHA,
	SPARSE,
	RIPSER
};

template <typename T> T read(std::istream& s) {
	T result;
	s.read(reinterpret_cast<char*>(&result), sizeof(T));
	return result; // on little endian: boost::endian::little_to_native(result);
}

compressed_lower_distance_matrix read_point_cloud(std::istream& input_stream) {
	std::vector<std::vector<value_t>> points;

	std::string line;
	value_t value;
	while (std::getline(input_stream, line)) {
		std::vector<value_t> point;
		std::istringstream s(line);
		while (s >> value) {
			point.push_back(value);
			s.ignore();
		}
		if (!point.empty()) points.push_back(point);
		assert(point.size() == points.front().size());
	}

	euclidean_distance_matrix eucl_dist(std::move(points));

	index_t n = eucl_dist.size();

	std::cout << "point cloud with " << n << " points in dimension "
	          << eucl_dist.points.front().size() << std::endl;

	std::vector<value_t> distances;

	for (int i = 0; i < n; ++i)
		for (int j = 0; j < i; ++j) distances.push_back(eucl_dist(i, j));

	return compressed_lower_distance_matrix(std::move(distances));
}

sparse_distance_matrix read_sparse_distance_matrix(std::istream& input_stream) {

	std::vector<std::vector<index_diameter_t>> neighbors;

	index_t num_edges = 0;

	std::string line;
	while (std::getline(input_stream, line)) {
		std::istringstream s(line);
		size_t i, j;
		value_t value;
		s >> i;
		s >> j;
		s >> value;
		if (i != j) {
			neighbors.resize(std::max({neighbors.size(), i + 1, j + 1}));
			neighbors[i].push_back(std::make_pair(j, value));
			neighbors[j].push_back(std::make_pair(i, value));
			++num_edges;
		}
	}

	for (index_t i = 0; i < neighbors.size(); ++i)
		std::sort(neighbors[i].begin(), neighbors[i].end());

	return sparse_distance_matrix(std::move(neighbors), num_edges);
}

compressed_lower_distance_matrix read_lower_distance_matrix(std::istream& input_stream) {
	std::vector<value_t> distances;
	value_t value;
	while (input_stream >> value) {
		distances.push_back(value);
		input_stream.ignore();
	}

	return compressed_lower_distance_matrix(std::move(distances));
}

compressed_lower_distance_matrix read_upper_distance_matrix(std::istream& input_stream) {
	std::vector<value_t> distances;
	value_t value;
	while (input_stream >> value) {
		distances.push_back(value);
		input_stream.ignore();
	}

	return compressed_lower_distance_matrix(compressed_upper_distance_matrix(std::move(distances)));
}

compressed_lower_distance_matrix read_distance_matrix(std::istream& input_stream) {
	std::vector<value_t> distances;

	std::string line;
	value_t value;
	for (int i = 0; std::getline(input_stream, line); ++i) {
		std::istringstream s(line);
		for (int j = 0; j < i && s >> value; ++j) {
			distances.push_back(value);
			s.ignore();
		}
	}

	return compressed_lower_distance_matrix(std::move(distances));
}

compressed_lower_distance_matrix read_dipha(std::istream& input_stream) {
	if (read<int64_t>(input_stream) != 8067171840) {
		std::cerr << "input is not a Dipha file (magic number: 8067171840)" << std::endl;
		exit(-1);
	}

	if (read<int64_t>(input_stream) != 7) {
		std::cerr << "input is not a Dipha distance matrix (file type: 7)" << std::endl;
		exit(-1);
	}

	index_t n = read<int64_t>(input_stream);

	std::vector<value_t> distances;

	for (int i = 0; i < n; ++i)
		for (int j = 0; j < n; ++j)
			if (i > j)
				distances.push_back(read<double>(input_stream));
			else
				read<double>(input_stream);

	return compressed_lower_distance_matrix(std::move(distances));
}

compressed_lower_distance_matrix read_ripser(std::istream& input_stream) {
	std::vector<value_t> distances;
	while (!input_stream.eof()) distances.push_back(read<value_t>(input_stream));
	return compressed_lower_distance_matrix(std::move(distances));
}

compressed_lower_distance_matrix read_file(std::istream& input_stream, file_format format) {
	switch (format) {
	case LOWER_DISTANCE_MATRIX:
		return read_lower_distance_matrix(input_stream);
	case UPPER_DISTANCE_MATRIX:
		return read_upper_distance_matrix(input_stream);
	case DISTANCE_MATRIX:
		return read_distance_matrix(input_stream);
	case POINT_CLOUD:
		return read_point_cloud(input_stream);
	case DIPHA:
		return read_dipha(input_stream);
	default:
		return read_ripser(input_stream);
	}
}

void print_usage_and_exit(int exit_code) {
	std::cerr
	    << "Usage: "
	    << "ripser "
	    << "[options] [filename]" << std::endl
	    << std::endl
	    << "Options:" << std::endl
	    << std::endl
	    << "  --help           print this screen" << std::endl
	    << "  --format         use the specified file format for the input. Options are:"
	    << std::endl
	    << "                     lower-distance (lower triangular distance matrix; default)"
	    << std::endl
	    << "                     upper-distance (upper triangular distance matrix)" << std::endl
	    << "                     distance       (full distance matrix)" << std::endl
	    << "                     point-cloud    (point cloud in Euclidean space)" << std::endl
	    << "                     dipha          (distance matrix in DIPHA file format)" << std::endl
	    << "                     sparse         (sparse distance matrix in Sparse Triplet format)"
	    << std::endl
	    << "                     ripser         (distance matrix in Ripser binary file format)"
	    << std::endl
	    << "  --dim <k>        compute persistent homology up to dimension <k>" << std::endl
	    << "  --threshold <t>  compute Rips complexes up to diameter <t>" << std::endl
#ifdef USE_COEFFICIENTS
	    << "  --modulus <p>    compute homology with coefficients in the prime field Z/<p>Z"
#endif
	    << std::endl;

	exit(exit_code);
}

int main(int argc, char** argv) {

	const char* filename = nullptr;

	file_format format = DISTANCE_MATRIX;

	index_t dim_max = 1;
	value_t threshold = std::numeric_limits<value_t>::max();

	float ratio = 1;

#ifdef USE_COEFFICIENTS
	coefficient_t modulus = 2;
#else
	const coefficient_t modulus = 2;
#endif

	for (index_t i = 1; i < argc; ++i) {
		const std::string arg(argv[i]);
		if (arg == "--help") {
			print_usage_and_exit(0);
		} else if (arg == "--dim") {
			std::string parameter = std::string(argv[++i]);
			size_t next_pos;
			dim_max = std::stol(parameter, &next_pos);
			if (next_pos != parameter.size()) print_usage_and_exit(-1);
		} else if (arg == "--threshold") {
			std::string parameter = std::string(argv[++i]);
			size_t next_pos;
			threshold = std::stof(parameter, &next_pos);
			if (next_pos != parameter.size()) print_usage_and_exit(-1);
		} else if (arg == "--ratio") {
			std::string parameter = std::string(argv[++i]);
			size_t next_pos;
			ratio = std::stof(parameter, &next_pos);
			if (next_pos != parameter.size()) print_usage_and_exit(-1);
		} else if (arg == "--format") {
			std::string parameter = std::string(argv[++i]);
			if (parameter == "lower-distance")
				format = LOWER_DISTANCE_MATRIX;
			else if (parameter == "upper-distance")
				format = UPPER_DISTANCE_MATRIX;
			else if (parameter == "distance")
				format = DISTANCE_MATRIX;
			else if (parameter == "point-cloud")
				format = POINT_CLOUD;
			else if (parameter == "dipha")
				format = DIPHA;
			else if (parameter == "sparse")
				format = SPARSE;
			else if (parameter == "ripser")
				format = RIPSER;
			else
				print_usage_and_exit(-1);
#ifdef USE_COEFFICIENTS
		} else if (arg == "--modulus") {
			std::string parameter = std::string(argv[++i]);
			size_t next_pos;
			modulus = std::stol(parameter, &next_pos);
			if (next_pos != parameter.size() || !is_prime(modulus)) print_usage_and_exit(-1);
#endif
		} else {
			if (filename) { print_usage_and_exit(-1); }
			filename = argv[i];
		}
	}

	std::ifstream file_stream(filename);
	if (filename && file_stream.fail()) {
		std::cerr << "couldn't open file " << filename << std::endl;
		exit(-1);
	}

	if (format == SPARSE) {
		sparse_distance_matrix dist =
		    read_sparse_distance_matrix(filename ? file_stream : std::cin);
		std::cout << "sparse distance matrix with " << dist.size() << " points and "
		          << dist.num_edges << "/" << (dist.size() * (dist.size() - 1)) / 2 << " entries"
		          << std::endl;

		ripser<sparse_distance_matrix>(std::move(dist), dim_max, threshold, ratio, modulus)
		    .compute_barcodes();
	} else {

		compressed_lower_distance_matrix dist =
		    read_file(filename ? file_stream : std::cin, format);

		value_t min = std::numeric_limits<value_t>::infinity(),
		        max = -std::numeric_limits<value_t>::infinity(), max_finite = max;
		int num_edges = 0;

		if (threshold == std::numeric_limits<value_t>::max()) {
			value_t enclosing_radius = std::numeric_limits<value_t>::infinity();
			for (index_t i = 0; i < dist.size(); ++i) {
				value_t r_i = -std::numeric_limits<value_t>::infinity();
				for (index_t j = 0; j < dist.size(); ++j) r_i = std::max(r_i, dist(i, j));
				enclosing_radius = std::min(enclosing_radius, r_i);
			}

			threshold = enclosing_radius;
		}

		for (auto d : dist.distances) {
			min = std::min(min, d);
			max = std::max(max, d);
			max_finite =
			    d != std::numeric_limits<value_t>::infinity() ? std::max(max, d) : max_finite;
			if (d <= threshold) ++num_edges;
		}

		std::cout << "value range: [" << min << "," << max_finite << "]" << std::endl;

		if (threshold >= max) {
			std::cout << "distance matrix with " << dist.size() << " points" << std::endl;
			ripser<compressed_lower_distance_matrix>(std::move(dist), dim_max, threshold, ratio,
			                                         modulus)
			    .compute_barcodes();
		} else {
			std::cout << "sparse distance matrix with " << dist.size() << " points and "
			          << num_edges << "/" << (dist.size() * dist.size() - 1) / 2 << " entries"
			          << std::endl;

			ripser<sparse_distance_matrix>(sparse_distance_matrix(std::move(dist), threshold),
			                               dim_max, threshold, ratio, modulus)
			    .compute_barcodes();
		}
	}
}