From d51d99a421661071c57b5d7fbe9e98e69fc0acc8 Mon Sep 17 00:00:00 2001
From: Reece Hoffmann <reecehoffmann@outlook.com>
Date: Fri, 19 Sep 2025 13:04:57 -0700
Subject: [PATCH 1/2] feat: support cd-hit 4.8.1

---
 Dockerfile       | 13 +++++++++++--
 install_cdhit.sh | 14 ++++++++++++++
 test.sh          | 22 +++++++++++++++++++++-
 3 files changed, 46 insertions(+), 3 deletions(-)
 create mode 100644 install_cdhit.sh

diff --git a/Dockerfile b/Dockerfile
index 77b32e2..7d424d1 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -49,6 +49,14 @@ RUN apt-get update && apt-get install -y build-essential wget libbz2-dev zlib1g-
 COPY install_samtools.sh .
 RUN bash install_samtools.sh 1.22.1
 
+FROM debian:bookworm AS cdhit
+WORKDIR /build
+RUN apt-get update && apt-get install -y wget build-essential libz-dev && rm -rf /var/lib/apt/lists/*
+COPY install_cdhit.sh .
+RUN bash install_cdhit.sh 4.8.1 v4.8.1-2019-0228
+RUN ls
+RUN pwd
+
 # Combine tools into a single image.
 FROM debian:bookworm AS combine
 WORKDIR /tools
@@ -59,11 +67,12 @@ COPY --from=hmmer /hmmer ./hmmer
 COPY --from=pigz /pigz/ ./pigz
 COPY --from=samtools /samtools/ ./samtools
 COPY --from=skewer /skewer ./skewer
+COPY --from=cdhit /cd-hit ./cd-hit
 
 # Testing
 FROM debian:bookworm AS test
 WORKDIR /tools
-RUN apt-get update && apt-get install -y default-jre perl libcurl4 && rm -rf /var/lib/apt/lists/*
+RUN apt-get update && apt-get install -y default-jre perl libcurl4 libgomp1 && rm -rf /var/lib/apt/lists/*
 COPY --from=combine /tools /tools
 COPY test.sh .
 RUN bash test.sh
@@ -71,5 +80,5 @@ RUN bash test.sh
 
 FROM debian:bookworm
 WORKDIR /tools
-RUN apt-get update && apt-get install -y libcurl4 && apt-get clean && rm -rf /var/lib/apt/lists/*
+RUN apt-get update && apt-get install -y libcurl4 libgomp1 && apt-get clean && rm -rf /var/lib/apt/lists/*
 COPY --from=test /tools /tools
\ No newline at end of file
diff --git a/install_cdhit.sh b/install_cdhit.sh
new file mode 100644
index 0000000..0db8447
--- /dev/null
+++ b/install_cdhit.sh
@@ -0,0 +1,14 @@
+#!/bin/bash
+
+# Get the version from the command line argument
+version=$1
+slug=$2
+
+# Download, extract, configure, and install the version
+wget https://github.com/weizhongli/cdhit/releases/download/V${version}/cd-hit-${slug}.tar.gz
+tar -xf cd-hit-$slug.tar.gz
+cd cd-hit-$slug
+make && make install
+cd ..
+mkdir -p /cd-hit/${version}
+mv cd-hit-$slug/cd-hit* /cd-hit/${version}
\ No newline at end of file
diff --git a/test.sh b/test.sh
index 76364fe..a302091 100644
--- a/test.sh
+++ b/test.sh
@@ -137,4 +137,24 @@ for version in "${skewer_versions[@]}"; do
         echo "Skewer version ${version} is not installed or not executable."
         exit 1
     fi
-done
\ No newline at end of file
+done
+
+# Test cd-hit
+cd_hit_versions=("4.8.1")
+
+for version in "${cd_hit_versions[@]}"; do
+    cd_hit_path="/tools/cd-hit/${version}/cd-hit"
+
+    if [ -x "$cd_hit_path" ]; then
+        detected_version=$("$cd_hit_path" | head -n 1 | grep -oE '[0-9]+\.[0-9]+\.[0-9]+')
+        if [ "$detected_version" = "$version" ]; then
+            echo "cd-hit version ${version} is installed and correct."
+        else
+            echo "cd-hit version ${version} is installed but version mismatch (found $detected_version)."
+            exit 1
+        fi
+    else
+        echo "cd-hit version ${version} is not installed or not executable."
+        exit 1
+    fi
+done

From c95301582a04d418a3db68e44238e5691b604959 Mon Sep 17 00:00:00 2001
From: Reece Hoffmann <reecehoffmann@outlook.com>
Date: Wed, 24 Sep 2025 12:08:56 -0700
Subject: [PATCH 2/2] fix: statically link libgomp into cdhit binary

---
 Dockerfile       | 6 ++----
 install_cdhit.sh | 1 +
 2 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index 7d424d1..98cbc4e 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -54,8 +54,6 @@ WORKDIR /build
 RUN apt-get update && apt-get install -y wget build-essential libz-dev && rm -rf /var/lib/apt/lists/*
 COPY install_cdhit.sh .
 RUN bash install_cdhit.sh 4.8.1 v4.8.1-2019-0228
-RUN ls
-RUN pwd
 
 # Combine tools into a single image.
 FROM debian:bookworm AS combine
@@ -72,7 +70,7 @@ COPY --from=cdhit /cd-hit ./cd-hit
 # Testing
 FROM debian:bookworm AS test
 WORKDIR /tools
-RUN apt-get update && apt-get install -y default-jre perl libcurl4 libgomp1 && rm -rf /var/lib/apt/lists/*
+RUN apt-get update && apt-get install -y default-jre perl libcurl4 && rm -rf /var/lib/apt/lists/*
 COPY --from=combine /tools /tools
 COPY test.sh .
 RUN bash test.sh
@@ -80,5 +78,5 @@ RUN bash test.sh
 
 FROM debian:bookworm
 WORKDIR /tools
-RUN apt-get update && apt-get install -y libcurl4 libgomp1 && apt-get clean && rm -rf /var/lib/apt/lists/*
+RUN apt-get update && apt-get install -y libcurl4 && apt-get clean && rm -rf /var/lib/apt/lists/*
 COPY --from=test /tools /tools
\ No newline at end of file
diff --git a/install_cdhit.sh b/install_cdhit.sh
index 0db8447..2746079 100644
--- a/install_cdhit.sh
+++ b/install_cdhit.sh
@@ -8,6 +8,7 @@ slug=$2
 wget https://github.com/weizhongli/cdhit/releases/download/V${version}/cd-hit-${slug}.tar.gz
 tar -xf cd-hit-$slug.tar.gz
 cd cd-hit-$slug
+sed -i 's/LDFLAGS += -lz -o/LDFLAGS += -Wl,-Bstatic -lgomp -Wl,-Bdynamic -lz -o/' Makefile
 make && make install
 cd ..
 mkdir -p /cd-hit/${version}