From 77843cd2c421693b84a694930918e1196854903c Mon Sep 17 00:00:00 2001
From: Dries Schaumont <5946712+DriesSchaumont@users.noreply.github.com>
Date: Thu, 2 May 2024 13:32:32 +0200
Subject: [PATCH] FEAT: untar input directory if needed. (#4)

* FEAT: untar input directory if needed.

* Fixes to untar component

* Fix workflow including untar

* Check for .tar and .tgz as well

* Remove debugging output from tar component

* Add target and testData folder to .gitignore

* Update to latest release candidate

* Add .nextflow and work to .gitignore

* Add runners to untar component

* Update workflow integration test

* Remove leftover view

---------

Co-authored-by: Toni Verbeiren <toni.verbeiren@gmail.com>
---
 .gitignore                           |  6 ++++++
 _viash.yaml                          |  2 +-
 src/demultiplex/config.vsh.yaml      |  2 ++
 src/demultiplex/integration_tests.sh | 10 ++++++++++
 src/demultiplex/main.nf              | 14 +++++++++++++-
 src/demultiplex/test.nf              |  9 ++++-----
 src/io/untar/config.vsh.yaml         |  6 +++++-
 7 files changed, 41 insertions(+), 8 deletions(-)
 create mode 100644 .gitignore

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..9e379c7
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,6 @@
+target
+testData
+
+# Nextflow related files
+.nextflow
+work
\ No newline at end of file
diff --git a/_viash.yaml b/_viash.yaml
index ee18185..e5e310d 100644
--- a/_viash.yaml
+++ b/_viash.yaml
@@ -7,4 +7,4 @@ links:
   issue_tracker: https://github.com/viash-hub/demultiplex.vsh/issues
   repository: https://github.com/viash-hub/demultiplex.vsh
 
-viash_version: 0.9.0-RC2
+viash_version: 0.9.0-RC3
diff --git a/src/demultiplex/config.vsh.yaml b/src/demultiplex/config.vsh.yaml
index 401e150..f840a76 100644
--- a/src/demultiplex/config.vsh.yaml
+++ b/src/demultiplex/config.vsh.yaml
@@ -29,6 +29,8 @@ test_resources:
     entrypoint: test_wf
 
 dependencies:
+  - name: io/untar
+    repository: local
   - name: bcl_convert
     repository: bb
 
diff --git a/src/demultiplex/integration_tests.sh b/src/demultiplex/integration_tests.sh
index 6299962..d327e36 100755
--- a/src/demultiplex/integration_tests.sh
+++ b/src/demultiplex/integration_tests.sh
@@ -1,3 +1,13 @@
+#!/usr/bin/env bash
+
+# get the root of the directory
+REPO_ROOT=$(git rev-parse --show-toplevel)
+
+# ensure that the command below is run from the root of the repository
+cd "$REPO_ROOT"
+
+viash ns build -q 'untar|demultiplex' --setup cb
+
 nextflow run . \
   -main-script src/demultiplex/test.nf \
   -profile docker,no_publish \
diff --git a/src/demultiplex/main.nf b/src/demultiplex/main.nf
index 56d3274..2cae23b 100644
--- a/src/demultiplex/main.nf
+++ b/src/demultiplex/main.nf
@@ -4,7 +4,19 @@ workflow run_wf {
 
   main:
     output_ch = input_ch
-
+      // untar input if needed
+      | untar.run(
+        runIf: {id, state -> 
+          def inputStr = state.input.toString()
+          inputStr.endsWith(".tar.gz") || inputStr.endsWith(".tar") || inputStr.endsWith(".tgz") ? true : false
+        },
+        fromState: [
+          "input": "input",
+        ],
+        toState: { id, result, state ->
+          state + [ input: result.output ]
+        },
+      )
       // run bcl_convert
       | bcl_convert.run(
           fromState: { id, state ->
diff --git a/src/demultiplex/test.nf b/src/demultiplex/test.nf
index 65e8378..51a5e03 100644
--- a/src/demultiplex/test.nf
+++ b/src/demultiplex/test.nf
@@ -3,13 +3,12 @@ nextflow.enable.dsl=2
 include { demultiplex } from params.rootDir + "/target/nextflow/demultiplex/main.nf"
 
 workflow test_wf {
-  // allow changing the resources_test dir
-  resources_test = file("${params.rootDir}/testData")
-
   output_ch = Channel.fromList([
         [
-          sample_sheet: resources_test.resolve("bcl_convert_samplesheet.csv"),
-          input: resources_test.resolve("iseq-DI/"),
+          // sample_sheet: resources_test.resolve("bcl_convert_samplesheet.csv"),
+          // input: resources_test.resolve("iseq-DI/"),
+          sample_sheet: "https://raw.githubusercontent.com/nf-core/test-datasets/demultiplex/testdata/NovaSeq6000/SampleSheet.csv",
+          input: "https://raw.githubusercontent.com/nf-core/test-datasets/demultiplex/testdata/NovaSeq6000/200624_A00834_0183_BHMTFYDRXX.tar.gz",
           publish_dir: "output_dir/",
         ]
       ])
diff --git a/src/io/untar/config.vsh.yaml b/src/io/untar/config.vsh.yaml
index 55eba97..a8ac4ac 100644
--- a/src/io/untar/config.vsh.yaml
+++ b/src/io/untar/config.vsh.yaml
@@ -33,4 +33,8 @@ test_resources:
     path: test.sh
 engines:
   - type: docker
-    image: debian:stable-slim
\ No newline at end of file
+    image: debian:stable-slim
+
+runners:
+  - type: executable
+  - type: nextflow
\ No newline at end of file