From a2fa58b09e746a5abb9ac8ea108b14475fea61f2 Mon Sep 17 00:00:00 2001
From: Svenja Mehringer <svenja.mehringer@fu-berlin.de>
Date: Wed, 23 Aug 2023 13:54:25 +0200
Subject: [PATCH] [FEATURE] Make hibf constructible from a layout file.

---
 .../hierarchical_interleaved_bloom_filter.hpp | 10 ++++
 src/hierarchical_interleaved_bloom_filter.cpp | 15 ++++++
 ...archical_interleaved_bloom_filter_test.cpp | 49 ++++++++++++++++++-
 3 files changed, 73 insertions(+), 1 deletion(-)

diff --git a/include/hibf/hierarchical_interleaved_bloom_filter.hpp b/include/hibf/hierarchical_interleaved_bloom_filter.hpp
index 5f026707..cc23275e 100644
--- a/include/hibf/hierarchical_interleaved_bloom_filter.hpp
+++ b/include/hibf/hierarchical_interleaved_bloom_filter.hpp
@@ -105,6 +105,16 @@ class hierarchical_interleaved_bloom_filter
     ~hierarchical_interleaved_bloom_filter() = default;            //!< Defaulted.
 
     hierarchical_interleaved_bloom_filter(config const & configuration);
+
+    /*!\brief [Advanced] Constructs the HIBF from a layout file (stream) and a given input function
+     * \details
+     * This constructor makes it possible to construct an hibf from a given layout file instead of calculating the
+     * layout based on the input function. A hibf::config object is not needed as it is assumed to be stored in the
+     * layout file. A layout file can be constructed manually or via chopper (https://github.com/seqan/chopper)
+     * or raptor-layout (https://github.com/seqan/raptor).
+     */
+    hierarchical_interleaved_bloom_filter(std::function<void(size_t const, insert_iterator &&)> input_fn,
+                                          std::istream & layout_stream);
     //!\}
 
     //!\brief The individual interleaved Bloom filters.
diff --git a/src/hierarchical_interleaved_bloom_filter.cpp b/src/hierarchical_interleaved_bloom_filter.cpp
index 8545010c..c5fa3772 100644
--- a/src/hierarchical_interleaved_bloom_filter.cpp
+++ b/src/hierarchical_interleaved_bloom_filter.cpp
@@ -199,4 +199,19 @@ hierarchical_interleaved_bloom_filter::hierarchical_interleaved_bloom_filter(con
     build_index(*this, configuration, layout);
 }
 
+hierarchical_interleaved_bloom_filter::hierarchical_interleaved_bloom_filter(
+    std::function<void(size_t const, insert_iterator &&)> input_fn,
+    std::istream & layout_stream)
+{
+    // read config and layout from file
+    config configuration;
+    layout::layout hibf_layout;
+    configuration.read_from(layout_stream);
+    hibf_layout.read_from(layout_stream);
+
+    configuration.input_fn = input_fn; // set input as it cannot be serialized.
+
+    build_index(*this, configuration, hibf_layout);
+}
+
 } // namespace hibf
diff --git a/test/unit/hibf/hierarchical_interleaved_bloom_filter_test.cpp b/test/unit/hibf/hierarchical_interleaved_bloom_filter_test.cpp
index 0638382c..2987de0f 100644
--- a/test/unit/hibf/hierarchical_interleaved_bloom_filter_test.cpp
+++ b/test/unit/hibf/hierarchical_interleaved_bloom_filter_test.cpp
@@ -9,7 +9,8 @@
 
 #include <cstddef>    // for size_t
 #include <functional> // for function
-#include <vector>     // for vector, allocator
+#include <sstream>
+#include <vector> // for vector, allocator
 
 #include <hibf/config.hpp>                                // for insert_iterator, config
 #include <hibf/hierarchical_interleaved_bloom_filter.hpp> // for hierarchical_interleaved_bloom_filter
@@ -41,6 +42,52 @@ TEST(hibf_test, test_specific_hash_values)
     }
 }
 
+TEST(hibf_test, build_from_layout)
+{
+    // range of range of sequences
+    std::vector<std::vector<size_t>> hashes{{1u, 2u, 3u, 4u, 5u, 6u, 7u, 8u, 9u, 10u}, {1u, 2u, 3u, 4u, 5u}};
+
+    auto input_fn = [&](size_t const num, hibf::insert_iterator it)
+    {
+        for (auto const hash : hashes[num])
+            it = hash;
+    };
+
+    std::stringstream stream{"@HIBF_CONFIG\n"
+                             "@{\n"
+                             "@    \"hibf_config\": {\n"
+                             "@        \"version\": 1,\n"
+                             "@        \"number_of_user_bins\": 2,\n"
+                             "@        \"number_of_hash_functions\": 2,\n"
+                             "@        \"maximum_false_positive_rate\": 0.05,\n"
+                             "@        \"threads\": 1,\n"
+                             "@        \"sketch_bits\": 12,\n"
+                             "@        \"tmax\": 64,\n"
+                             "@        \"alpha\": 1.2,\n"
+                             "@        \"max_rearrangement_ratio\": 0.5,\n"
+                             "@        \"disable_estimate_union\": false,\n"
+                             "@        \"disable_rearrangement\": true,\n"
+                             "@        \"disable_cutoffs\": false\n"
+                             "@    }\n"
+                             "@}\n"
+                             "@HIBF_CONFIG_END\n"
+                             "#TOP_LEVEL_IBF fullest_technical_bin_idx:0\n"
+                             "#USER_BIN_IDX\tTECHNICAL_BIN_INDICES\tNUMBER_OF_TECHNICAL_BINS\n"
+                             "1\t0\t34\n"
+                             "0\t34\t30\n"};
+
+    hibf::hierarchical_interleaved_bloom_filter hibf{input_fn, stream};
+
+    {
+        std::vector<size_t> query{1, 2, 3, 4, 5};
+
+        auto agent = hibf.membership_agent();
+        auto result = agent.bulk_contains(query, 2);
+
+        EXPECT_RANGE_EQ(result, (std::vector<size_t>{0u, 1u}));
+    }
+}
+
 // #ifdef HIBF_HAS_SEQAN3
 
 // #include <seqan3/alphabet/nucleotide/dna4.hpp>