From 886ecc5a48b455d271ad3bb66257f5de4a274550 Mon Sep 17 00:00:00 2001
From: "N. Tessa Pierce-Ward" <ntpierce@gmail.com>
Date: Wed, 30 Aug 2023 14:52:49 -0700
Subject: [PATCH] better help; avoid recalc threshold

---
 Cargo.toml                              | 2 +-
 src/lib.rs                              | 7 ++-----
 src/python/pyo3_branchwater/__init__.py | 3 ++-
 3 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 1d441084..49a2b9db 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -12,7 +12,7 @@ crate-type = ["cdylib"]
 pyo3 = { version = "0.18.1", features = ["extension-module", "anyhow"] }
 rayon = "1.5.1"
 serde = { version = "1.0.136", features = ["derive"] }
-sourmash = { git = "https://github.com/sourmash-bio/sourmash", branch = "ntp/mastiff_roaring_prot" }
+sourmash = { git = "https://github.com/sourmash-bio/sourmash", "rev" = "ff1092f8f366339caa59d7203f623813228f4356" }
 clap = { version = "3.1.0", features = ["derive"] }
 serde_json = "1.0.79"
 niffler = "2.4.0"
diff --git a/src/lib.rs b/src/lib.rs
index a74b5411..596e91d2 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -862,7 +862,6 @@ fn mastiff_manysearch<P: AsRef<Path>>(
     queries_file: P,
     index: P,
     template: Sketch,
-    threshold_bp: usize,
     minimum_containment: f64,
     output: Option<P>,
     ) -> Result<(), Box<dyn std::error::Error>> {
@@ -923,10 +922,9 @@ fn mastiff_manysearch<P: AsRef<Path>>(
             if let Ok(query_sig) = Signature::from_path(filename) {
                 if let Some(query) = prepare_query(&query_sig, &template) {
                     let query_size = query.minhash.size() as f64;
-                    let threshold = threshold_bp / query.minhash.scaled() as usize;
                     // search mastiff db
                     let counter = db.counter_for_query(&query.minhash);
-                    let matches = db.matches_from_counter(counter, threshold);
+                    let matches = db.matches_from_counter(counter, minimum_containment as usize);
 
                     // filter the matches for containment
 
@@ -1142,8 +1140,7 @@ fn do_manysearch(querylist_path: String,
     // if siglist_path is revindex, run mastiff_manysearch; otherwise run manysearch
     if is_revindex_database(siglist_path.as_ref()) {
         let template = build_template(ksize, scaled);
-        let threshold_bp = (threshold * scaled as f64) as usize;
-        match mastiff_manysearch(querylist_path, siglist_path, template, threshold_bp, threshold, output_path) {
+        match mastiff_manysearch(querylist_path, siglist_path, template, threshold, output_path) {
             Ok(_) => Ok(0),
             Err(e) => {
                 eprintln!("Error: {e}");
diff --git a/src/python/pyo3_branchwater/__init__.py b/src/python/pyo3_branchwater/__init__.py
index 307d040a..77bd4f16 100755
--- a/src/python/pyo3_branchwater/__init__.py
+++ b/src/python/pyo3_branchwater/__init__.py
@@ -77,7 +77,8 @@ class Branchwater_Fastgather(CommandLinePlugin):
     def __init__(self, p):
         super().__init__(p)
         p.add_argument('query_sig', help="metagenome sketch")
-        p.add_argument('against_paths', help="a text file containing paths to .sig/.sig.gz files")
+        p.add_argument('against_paths', help="a text file containing paths to .sig/.sig.gz files \
+                       OR a branchwater indexed database generated with 'sourmash scripts index'")
         p.add_argument('-o', '--output-gather', required=True,
                        help="save gather output (minimum metagenome cover) to this file")
         p.add_argument('--output-prefetch',