From 72ca008bc63aaadc4fb7b2e67b2e470f62be0253 Mon Sep 17 00:00:00 2001
From: lvreynoso <lreynoso@chanzuckerberg.com>
Date: Wed, 8 May 2024 14:47:18 -0700
Subject: [PATCH] Move diamond changes to another branch

---
 lib/idseq_utils/idseq_utils/diamond_scatter.py | 12 ++----------
 scripts/release.sh                             |  4 ----
 workflows/diamond/Dockerfile                   |  6 +++---
 workflows/diamond/diamond.wdl                  |  4 ++--
 4 files changed, 7 insertions(+), 19 deletions(-)

diff --git a/lib/idseq_utils/idseq_utils/diamond_scatter.py b/lib/idseq_utils/idseq_utils/diamond_scatter.py
index 10ad3c8af..385301c1d 100644
--- a/lib/idseq_utils/idseq_utils/diamond_scatter.py
+++ b/lib/idseq_utils/idseq_utils/diamond_scatter.py
@@ -3,7 +3,6 @@
 """
 
 import os
-import shlex
 import shutil
 import sys
 import errno
@@ -72,15 +71,8 @@ def diamond_blastx(
         database,
         "--out",
         out,
+        f"--{diamond_args}",
     ]
-
-    # backwards compatibility for function calls that expect this function
-    # to automatically append "--" to diamond_args
-    if diamond_args == "long-reads" or diamond_args == "mid-sensitive":
-        diamond_args = "--" + diamond_args
-
-    cmd.extend(shlex.split(diamond_args))
-
     for query in queries:
         cmd += ["--query", query]
     if chunk:
@@ -238,7 +230,7 @@ def blastx_join(chunk_dir: str, out: str, diamond_args: str, *query: str):
             diamond_blastx(
                 cwd=tmp_dir,
                 par_tmpdir="par-tmp",
-                block_size=1,
+                block_size=100 if "long-reads" in diamond_args else 10,
                 database=db.name,
                 out=out,
                 join_chunks=chunks,
diff --git a/scripts/release.sh b/scripts/release.sh
index ec0dc9884..da13b1e74 100755
--- a/scripts/release.sh
+++ b/scripts/release.sh
@@ -28,10 +28,6 @@ else
     exit 1
 fi
 
-if [[ $WORKFLOW_NAME == long-read-mngs ]]; then
-    TAG="long-read-mngs-v0.7.6-branch-diamondsensitive"
-fi
-
 if [[ $( git branch --show-current) != "main" ]]; then 
     COMMIT=$(git rev-parse --short HEAD)
     TAG=$TAG"-$COMMIT"
diff --git a/workflows/diamond/Dockerfile b/workflows/diamond/Dockerfile
index c24bb2e0f..806ef944a 100644
--- a/workflows/diamond/Dockerfile
+++ b/workflows/diamond/Dockerfile
@@ -19,7 +19,7 @@ WORKDIR /tmp
 RUN git clone https://github.com/chanzuckerberg/czid-workflows
 WORKDIR /tmp/czid-workflows
 RUN pip3 install -r requirements-dev.txt
+RUN git checkout rlim-add-diamond-modification
+RUN cp short-read-mngs/idseq_utils/idseq_utils/diamond_scatter.py /usr/local/bin/
 
-COPY --from=lib idseq_utils/idseq_utils/diamond_scatter.py /usr/local/bin/
-
-WORKDIR /workdir
+WORKDIR /workdir
\ No newline at end of file
diff --git a/workflows/diamond/diamond.wdl b/workflows/diamond/diamond.wdl
index 7a8f2f814..35af2d033 100644
--- a/workflows/diamond/diamond.wdl
+++ b/workflows/diamond/diamond.wdl
@@ -34,7 +34,7 @@ task RunDiamond {
     }
 
     command <<<
-        python3 /usr/local/bin/diamond_scatter.py blastx-chunk --db ~{db_chunk} --query ~{query_0} ~{if defined(query_1) then '--query ~{query_1}' else ''} --out-dir chunks --diamond-args="~{extra_args}"
+        python3 /usr/local/bin/diamond_scatter.py blastx-chunk --db ~{db_chunk} --query ~{query_0} ~{if defined(query_1) then '--query ~{query_1}' else ''} --out-dir chunks --diamond-args "~{extra_args}"
     >>>
 
     output {
@@ -45,4 +45,4 @@ task RunDiamond {
     runtime {
         docker: docker_image_id
     }
-}
+}
\ No newline at end of file