From 7a002e83f25bf532fc19b5a09cdcd040ae81fd12 Mon Sep 17 00:00:00 2001
From: "C. Titus Brown" <titus@idyll.org>
Date: Mon, 28 Aug 2023 07:37:09 -0700
Subject: [PATCH] remove Jaccard calculation from manysearch (#72)

---
 src/lib.rs                      | 18 ++++--------------
 src/python/tests/test_search.py |  7 +------
 2 files changed, 5 insertions(+), 20 deletions(-)

diff --git a/src/lib.rs b/src/lib.rs
index 19d23e28..d5713a7f 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -143,10 +143,6 @@ fn prepare_query(search_sigs: &[Signature], template: &Sketch) -> Option<SmallSi
 ///
 /// Note: this function loads all _queries_ into memory, and iterates over
 /// database once.
-///
-/// TODO:
-///   - support jaccard as well as containment/overlap
-///   - support md5 output columns; other?
 
 fn manysearch<P: AsRef<Path>>(
     querylist: P,
@@ -208,10 +204,10 @@ fn manysearch<P: AsRef<Path>>(
     };
     let thrd = std::thread::spawn(move || {
         let mut writer = BufWriter::new(out);
-        writeln!(&mut writer, "query_name,query_md5,match_name,match_md5,containment,jaccard,intersect_hashes").unwrap();
-        for (query, query_md5, m, m_md5, cont, jaccard, overlap) in recv.into_iter() {
-            writeln!(&mut writer, "\"{}\",{},\"{}\",{},{},{},{}",
-                     query, query_md5, m, m_md5, cont, jaccard, overlap).ok();
+        writeln!(&mut writer, "query_name,query_md5,match_name,match_md5,containment,intersect_hashes").unwrap();
+        for (query, query_md5, m, m_md5, cont, overlap) in recv.into_iter() {
+            writeln!(&mut writer, "\"{}\",{},\"{}\",{},{},{}",
+                     query, query_md5, m, m_md5, cont, overlap).ok();
         }
     });
 
@@ -243,19 +239,13 @@ fn manysearch<P: AsRef<Path>>(
                         let overlap = q.minhash.count_common(&search_sm.minhash, false).unwrap() as f64;
                         let query_size = q.minhash.size() as f64;
 
-                        let mut merged = q.minhash.clone();
-                        merged.merge(&search_sm.minhash).ok();
-                        let total_size = merged.size() as f64;
-
                         let containment = overlap / query_size;
-                        let jaccard = overlap / total_size;
                         if containment > threshold {
                             results.push((q.name.clone(),
                                           q.md5sum.clone(),
                                           search_sm.name.clone(),
                                           search_sm.md5sum.clone(),
                                           containment,
-                                          jaccard,
                                           overlap))
                         }
                     }
diff --git a/src/python/tests/test_search.py b/src/python/tests/test_search.py
index 9970a366..6e207223 100644
--- a/src/python/tests/test_search.py
+++ b/src/python/tests/test_search.py
@@ -53,25 +53,20 @@ def test_simple(runtmp):
         if row['query_md5'] == row['match_md5']:
             assert row['match_name'] == row['query_name']
             assert float(row['containment'] == 1.0)
-            assert float(row['jaccard'] == 1.0)
         else:
             # confirm hand-checked numbers
             q = row['query_name'].split()[0]
             m = row['match_name'].split()[0]
-            jaccard = float(row['jaccard'])
             cont = float(row['containment'])
             intersect_hashes = int(row['intersect_hashes'])
 
-            jaccard = round(jaccard, 4)
             cont = round(cont, 4)
-            print(q, m, f"{jaccard:.04}", f"{cont:.04}")
+            print(q, m, f"{cont:.04}")
 
             if q == 'NC_011665.1' and m == 'NC_009661.1':
-                assert jaccard == 0.3207
                 assert cont == 0.4828
 
             if q == 'NC_009661.1' and m == 'NC_011665.1':
-                assert jaccard == 0.3207
                 assert cont == 0.4885