Merge branch 'dev' of github.com:lskatz/Kalamari into dev

lskatz · Jun 5, 2024 · 50c8120 · 50c8120
2 parents 008e9da + 8cfa2e5
commit 50c8120
Show file tree

Hide file tree

Showing 6 changed files with 32 additions and 27 deletions.
diff --git a/.github/workflows/unit-testing.Listeria.Kraken1.yml b/.github/workflows/unit-testing.Listeria.Kraken1.yml
@@ -1,7 +1,7 @@
 # This is a subsampling unit test to get early results
 on: 
   push:
-    branches: [master, dev, validate-taxonomy]
+    branches: [master, dev]
 name: Listeria-with-Kraken1
 
 env:
@@ -41,18 +41,10 @@ jobs:
           tree $(realpath .)
       - name: install-edirect
         run:  |
-          sudo apt-get install ncbi-entrez-direct
-          echo "installed edirect the apt way"
-          exit
-          cd $HOME
-          perl -MNet::FTP -e '$ftp = new Net::FTP("ftp.ncbi.nlm.nih.gov", Passive => 1); $ftp->login; $ftp->binary; $ftp->get("/entrez/entrezdirect/edirect.tar.gz");'
-          gunzip -cv edirect.tar.gz | tar xf -
-          rm -v edirect.tar.gz
-          echo $GITHUB_WORKSPACE/edirect >> $GITHUB_PATH
+          sh -c "$(curl -fsSL https://ftp.ncbi.nlm.nih.gov/entrez/entrezdirect/install-edirect.sh)"
+          echo $HOME/edirect >> $GITHUB_PATH
           echo $GITHUB_WORKSPACE/Kalamari/bin >> $GITHUB_PATH
-          #export PATH=${PATH}:$HOME/edirect >& /dev/null || setenv PATH "${PATH}:$HOME/edirect"
-          yes Y | ./edirect/setup.sh
-          tree edirect
+          tree $HOME/edirect
       - name: check-env
         run:  echo "$PATH"
       - name: select for only Listeria

diff --git a/.github/workflows/unit-testing.Yersinia.Kraken2.yml b/.github/workflows/unit-testing.Yersinia.Kraken2.yml
@@ -1,7 +1,7 @@
 # This is a subsampling unit test to get early results
 on: 
   push:
-    branches: [master, dev, validate-taxonomy]
+    branches: [master, dev]
 name: Genera-with-Kraken2
 
 env:
@@ -34,6 +34,14 @@ jobs:
       - name: env check
         run:  |
           echo $PATH | tr ':' '\n' | sort
+
+      - name: install-edirect
+        run:  |
+          sh -c "$(curl -fsSL https://ftp.ncbi.nlm.nih.gov/entrez/entrezdirect/install-edirect.sh)"
+          echo $HOME/edirect >> $GITHUB_PATH
+          echo $GITHUB_WORKSPACE/Kalamari/bin >> $GITHUB_PATH
+          tree $HOME/edirect 
+                   
       - name: apt-get install
         run:  sudo apt-get install ca-certificates tree jellyfish ncbi-entrez-direct
       - name: select for only for this genus

diff --git a/.github/workflows/unit-testing.yml b/.github/workflows/unit-testing.yml
@@ -1,6 +1,6 @@
 on: 
   push:
-    branches: [master, dev, validate-taxonomy]
+    branches: [master, dev]
 name: Pull-down-all-accessions
 
 jobs:
@@ -25,18 +25,13 @@ jobs:
 
       - name: apt-get install
         run:  sudo apt-get install ca-certificates tree
+
       - name: install-edirect
         run:  |
-          sudo apt-get install ncbi-entrez-direct
-          echo "installed edirect the apt way"
-          exit
-          cd $HOME
-          perl -MNet::FTP -e '$ftp = new Net::FTP("ftp.ncbi.nlm.nih.gov", Passive => 1); $ftp->login; $ftp->binary; $ftp->get("/entrez/entrezdirect/edirect.tar.gz");'
-          gunzip -cv edirect.tar.gz | tar xf -
-          rm -v edirect.tar.gz
-          export PATH=${PATH}:$HOME/edirect >& /dev/null || setenv PATH "${PATH}:$HOME/edirect"
-          yes Y | ./edirect/setup.sh
-          tree edirect
+          sh -c "$(curl -fsSL https://ftp.ncbi.nlm.nih.gov/entrez/entrezdirect/install-edirect.sh)"
+          echo $HOME/edirect >> $GITHUB_PATH
+          echo $GITHUB_WORKSPACE/Kalamari/bin >> $GITHUB_PATH
+          tree $HOME/edirect
       - name: check-env
         run:  echo "$PATH"
       - name: download

diff --git a/.github/workflows/validateTaxonomy.yml b/.github/workflows/validateTaxonomy.yml
@@ -1,6 +1,6 @@
 on: 
   push:
-    branches: [master, dev, validate-taxonomy]
+    branches: [master, dev, esearch-input]
 name: Validate taxonomy
 
 jobs:
@@ -27,11 +27,18 @@ jobs:
           echo $PATH
           echo ""
           cat $GITHUB_PATH
+      - name: install taxonkit
+        run:  |
+          wget https://github.com/shenwei356/taxonkit/releases/download/v0.16.0/taxonkit_linux_amd64.tar.gz
+          tar -xvf taxonkit_linux_amd64.tar.gz
+          rm -v taxonkit_linux_amd64.tar.gz
+          chmod +x taxonkit
+          echo $(realpath .) >> $GITHUB_PATH
       - name: build taxonomy
         run:  |
           echo $PATH
-          bash Kalamari/bin/buildTaxonomy.sh
-          bash Kalamari/bin/filterTaxonomy.sh
+          bash -x Kalamari/bin/buildTaxonomy.sh
+          bash -x Kalamari/bin/filterTaxonomy.sh
           ls -lhR Kalamari/share/kalamari-*/taxonomy
       - name: validate taxonomy
         run:  |

diff --git a/bin/buildKraken1.sh b/bin/buildKraken1.sh
@@ -30,6 +30,7 @@ find $SRC -name '*.fasta.gz' | \
     done > $tmpfile
     echo -ne "ADDING to library:\n  "
     zgrep "^>" $tmpfile | sed "s/^>//" | tr "$nl" " "
+    echo
     echo "^^ contents of $tmpfile ^^"
     kraken-build --db $DB --add-to-library $tmpfile
   '

diff --git a/bin/filterTaxonomy.sh b/bin/filterTaxonomy.sh
@@ -3,7 +3,9 @@
 set -eu
 
 # Check for dependencies
+echo "Check for dependencies"
 which taxonkit
+echo
 
 thisdir=$(dirname $0)
 thisfile=$(basename $0)