Develop merge for 2.2.2 release

.github/workflows/pypi_publish.yml

@@ -0,0 +1,49 @@
+name: Publish package python distribution to Pypi
+
+on:
+  release:
+    types: [published]
+  workflow_dispatch:
+
+jobs:
+    build:
+      name: Build distribution
+      runs-on: ubuntu-latest
+      steps:
+      - uses: actions/checkout@v4
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: 3.12.7
+      - name: Install pypi/build
+        run: >-
+          python3 -m
+          pip install
+          build
+          --user
+      - name: Build a binary wheel and a source tarball
+        run: python3 -m build
+      - name: Store the distribution packages
+        uses: actions/upload-artifact@v4
+        with:
+          name: python-package-distributions
+          path: dist/
+
+    publish-to-pypi:
+      name: Publish dist to PyPI
+      needs:
+      - build
+      runs-on: ubuntu-latest
+      environment:
+        name: pypi
+        url: https://pypi.org/p/buisciii-tools
+      permissions:
+        id-token: write
+      steps:
+      - name: Download all the dists
+        uses: actions/download-artifact@v4
+        with:
+          name: python-package-distributions
+          path: dist/
+      - name: Publish to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1

CHANGELOG.md

@@ -8,14 +8,63 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 
 ### Credits
 
+### Template fixes and updates
+
+### Modules
+
+#### Added enhancements
+
+#### Fixes
+
+#### Changed
+
+#### Removed
+
+### Requirements
+
+## [2.X.Xhot] - 2024-0X-0X : https://github.com/BU-ISCIII/buisciii-tools/releases/tag/2.X.2
+
+### Credits
+
+### Template fixes and updates
+
+### Modules
+
+#### Added enhancements
+
+#### Fixes
+
+#### Changed
+
+#### Removed
+
+### Requirements
+
+## [2.2.2] - 2024-10-28 : https://github.com/BU-ISCIII/buisciii-tools/releases/tag/2.2.2
+
+### Credits
+
 Code contributions to the new version:
 
+- [Pablo Mata](https://github.com/Shettland)
+- [Victor Lopez](https://github.com/victor5lm)
+
 ### Template fixes and updates
 
+- Updated the fix-permissions module in __main__.py [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+- Fixed the singularity cache directory in taxprofiler.config [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+- Updated sftp_user.json [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+- Fixed viralrecon's lablog and the remove_columns_mapping_table.sh auxiliary script [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+- Updated the singularity image in the mtbseq templates [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+- Fixed a bug in bioinfo_doc.py [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+- Updated new_service.py to check FASTQ integrity via md5sum [#356](https://github.com/BU-ISCIII/buisciii-tools/pull/356).
+
 ### Modules
 
 #### Added enhancements
 
+- Included a new github action to automatically publish releases to pypi [#351](https://github.com/BU-ISCIII/buisciii-tools/pull/351)
+
 #### Fixes
 
 #### Changed

README.md

@@ -31,7 +31,7 @@ BU-ISCIII provides a serie or services in its portfolio for supporting bioinform
 ```bash
 micromamba create -n buisciii -f environment.yml
 micromamba activate buisciii
-pip install --force-reinstall --upgrade git+https://github.com/bu-isciii/buisciii-tools.git@main
+pip install buisciii-tools
 ```
 
 or
@@ -40,7 +40,7 @@ or
 git checkout main
 conda create -n buisciii -f environment.yml
 conda activate 
-pip install .
+pip install buisciii-tools
 ```
 
 ### Dev version

bu_isciii/__main__.py

@@ -57,7 +57,7 @@ def run_bu_isciii():
     )
 
     # stderr.print("[green]                                          `._,._,'\n", highlight=False)
-    __version__ = "2.2.1"
+    __version__ = "2.2.2"
     stderr.print(
         "[grey39]    BU-ISCIII-tools version {}".format(__version__), highlight=False
     )
@@ -655,6 +655,7 @@ def autoclean_sftp(ctx, sftp_folder, days):
     "-d",
     "--input_directory",
     type=click.Path(),
+    multiple=True,
     default=None,
     required=True,
     help="Input directory to fix permissions (absolute path)",
@@ -664,15 +665,18 @@ def fix_permissions(ctx, input_directory):
     """
     Fix permissions
     """
-    if not os.path.isdir(input_directory):
-        exit("Invalid input directory")
     conf = bu_isciii.config_json.ConfigJson()
     permissions = conf.get_configuration("global").get("permissions")
-    bu_isciii.utils.remake_permissions(input_directory, permissions)
     stderr = rich.console.Console(
         stderr=True, force_terminal=bu_isciii.utils.rich_force_colors()
     )
-    stderr.print(f"[green]Correct permissions were applied to {input_directory}")
+
+    for directory in input_directory:
+        if not os.path.isdir(directory):
+            stderr.print(f"[red]Invalid input directory: {directory}")
+            continue
+        bu_isciii.utils.remake_permissions(directory, permissions)
+        stderr.print(f"[green]Correct permissions were applied to {directory}")
 
 
 if __name__ == "__main__":

bu_isciii/bioinfo_doc.py

@@ -715,7 +715,7 @@ def send_email(self, html_text, results_pdf_file):
             server.ehlo()
             server.login(user=email_host_user, password=email_host_password)
         except Exception as e:
-            stderr.print("[red] Unable to send e-mail" + e)
+            stderr.print("[red] Unable to send e-mail: " + str(e))
         default_cc = "[email protected]"
         msg = MIMEMultipart("alternative")
         msg["To"] = self.resolution_info["service_user_id"]["email"]

bu_isciii/new_service.py

@@ -8,6 +8,7 @@
 import json
 import shutil
 import rich
+import subprocess
 
 # Local imports
 import bu_isciii
@@ -88,7 +89,35 @@ def __init__(
             )
         self.full_path = os.path.join(self.path, self.service_folder)
 
+    def check_md5(self):
+        # Path to the .md5 file
+        project_name = self.service_samples[0]["project_name"]
+        md5_file_path = (
+            f'{self.conf["fastq_repo"]}/{project_name}/md5sum_{project_name}.md5'
+        )
+        if not os.path.exists(md5_file_path):
+            stderr.print(f"[red]ERROR: .md5 file not found at {md5_file_path}")
+            sys.exit(1)
+
+        original_dir = os.getcwd()
+        md5_dir = os.path.dirname(md5_file_path)
+        os.chdir(md5_dir)
+
+        # md5sum command
+        stderr.print(f"[blue]Checking MD5 integrity for {md5_file_path}")
+        try:
+            subprocess.run(
+                ["md5sum", "-c", os.path.basename(md5_file_path)], check=True
+            )
+            stderr.print("[green]MD5 check passed!")
+        except subprocess.CalledProcessError as e:
+            stderr.print(f"[red]ERROR: MD5 check failed: {e.stderr}")
+            sys.exit(1)
+        finally:
+            os.chdir(original_dir)
+
     def create_folder(self):
+        self.check_md5()
         if not self.no_create_folder:
             stderr.print(
                 "[blue]I will create the service folder for " + self.resolution_id + "!"

bu_isciii/templates/mag/DOC/taxprofiler.config

@@ -5,7 +5,7 @@
 singularity {
     enabled                 = true
     autoMounts              = true
-    singularity.cacheDir    = '/data/cnm/ratb/pipelines/singularity-images/'
+    singularity.cacheDir    = '/data/bi/pipelines/singularity-images/'
 }
 
 process {

bu_isciii/templates/mtbseq/ANALYSIS/ANALYSIS01_MTBSEQ/03-MTBSeq/all_samples/lablog

@@ -14,9 +14,9 @@ cd Mpileup; ln -s ../../*/Mpileup/* . ; cd -
 cd Position_Tables; ln -s ../../*/Position_Tables/* . ; cd -
 
 ##### Create join scripts
-echo "srun --job-name MTBSEQ_JOIN  --output logs/MTBSEQ_JOIN.%j.log --partition middle_obx --mem 48G --chdir ${scratch_dir} --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/mtbseq:1.0.4—hdfd78af_2 MTBseq --step TBjoin --threads 5 --samples ${scratch_dir}/samples.txt &" > _01_tb_join.sh
-echo "srun --job-name MTBSEQ_AMEND  --output logs/MTBSEQ_AMEND.%j.log --partition middle_obx --mem 48G --chdir ${scratch_dir} --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/mtbseq:1.0.4—hdfd78af_2 MTBseq --step TBamend --threads 5 --samples ${scratch_dir}/samples.txt &" > _02_tb_amend.sh
-echo "srun --job-name MTBSEQ_GROUPS --output logs/MTBSEQ_GROUPS.%j.log --partition middle_obx --mem 48G --chdir ${scratch_dir} --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/mtbseq:1.0.4—hdfd78af_2 MTBseq --step TBgroups --threads 5 --samples ${scratch_dir}/samples.txt &" > _03_tb_groups.sh
+echo "srun --job-name MTBSEQ_JOIN  --output logs/MTBSEQ_JOIN.%j.log --partition middle_obx --mem 48G --chdir ${scratch_dir} --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/mtbseq:1.1.0--hdfd78af_0 MTBseq --step TBjoin --threads 5 --samples ${scratch_dir}/samples.txt &" > _01_tb_join.sh
+echo "srun --job-name MTBSEQ_AMEND  --output logs/MTBSEQ_AMEND.%j.log --partition middle_obx --mem 48G --chdir ${scratch_dir} --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/mtbseq:1.1.0--hdfd78af_0 MTBseq --step TBamend --threads 5 --samples ${scratch_dir}/samples.txt &" > _02_tb_amend.sh
+echo "srun --job-name MTBSEQ_GROUPS --output logs/MTBSEQ_GROUPS.%j.log --partition middle_obx --mem 48G --chdir ${scratch_dir} --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/mtbseq:1.1.0--hdfd78af_0 MTBseq --step TBgroups --threads 5 --samples ${scratch_dir}/samples.txt &" > _03_tb_groups.sh
 
 #### Execute iqtree
 echo "srun --chdir ${scratch_dir} --output logs/IQTREEFULLALIGN.%j.log --job-name IQTREEFULLALIGN --cpus-per-task 20 --mem 15G --partition short_idx --time 08:00:00 singularity exec -B ${scratch_dir}/../../../../ /data/bi/pipelines/singularity-images/iqtree:2.1.4_beta--hdcc8f71_0 iqtree -s ${scratch_dir}/Amend/*amended_u95_phylo_w12.plainIDs.fasta -m K3Pu+F+I -T 20 -B 1000 -pre phylo.iqtree.bootstrap &" > _04_iqtreeall.sh

bu_isciii/templates/mtbseq/ANALYSIS/ANALYSIS01_MTBSEQ/03-MTBSeq/lablog

@@ -9,7 +9,7 @@ cat ../samples_id.txt | xargs -I % echo "mkdir %; ln -s ../../01-preprocessing/%
 cat ../samples_id.txt | xargs -I % echo "ln -s ../../01-preprocessing/%/%_R2_filtered.fastq.gz %/%_lib1_R2.fastq.gz" >> _00_prepareRaw.sh
 cat ../samples_id.txt | xargs -I % echo "cd %;ls *.fastq.gz | tr '_' '\t' | cut -f 1,2 | sort -u > samples.txt;cd -" > _01_preparesamples.sh
 
-cat ../samples_id.txt | xargs -I @@ echo -e "srun --job-name MTBSEQ.@@  --output logs/MTBSEQ.@@.%j.log --partition middle_idx --mem 48G --chdir ${scratch_dir}/@@ --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../ /data/bi/pipelines/singularity-images/mtbseq:1.0.4--hdfd78af_2 MTBseq --step TBfull --threads 10 --samples samples.txt &" > _02_mtbseq.sh
+cat ../samples_id.txt | xargs -I @@ echo -e "srun --job-name MTBSEQ.@@  --output logs/MTBSEQ.@@.%j.log --partition middle_idx --mem 48G --chdir ${scratch_dir}/@@ --cpus-per-task 10 singularity exec -B ${scratch_dir}/../../../ /data/bi/pipelines/singularity-images/mtbseq:1.1.0--hdfd78af_0 MTBseq --step TBfull --threads 10 --samples samples.txt &" > _02_mtbseq.sh
 
 # classification
 echo "mkdir classification_all" > _03_gather_results.sh 

bu_isciii/templates/sftp_user.json

@@ -54,5 +54,6 @@
     "bbaladron": ["SpainUDP"],
     "bioinfoadm": ["test"],
     "s.varona": ["misc"],
-    "nlabiod": ["Labarbovirus"]
+    "nlabiod": ["Labarbovirus"],
+    "jmgonzalez": ["Lablegionella"]
 }

bu_isciii/templates/viralrecon/ANALYSIS/lablog_viralrecon

@@ -479,6 +479,7 @@ do
     cp create_summary_report.sh ${FOLDER_NAME}/
     cp deduplicate_long_table.sh ${FOLDER_NAME}/
     cp percentajeNs.py ${FOLDER_NAME}/
+    cp remove_columns_mapping_table.sh ${FOLDER_NAME}/
     grep -i ${in} samples_ref.txt | cut -f1,2 > ${FOLDER_NAME}/samples_ref.txt
     echo "ln -s ../00-reads ." > ${FOLDER_NAME}/lablog
     printf "ln -s ../samples_id.txt .\n\n" >> ${FOLDER_NAME}/lablog

bu_isciii/templates/viralrecon/ANALYSIS/remove_columns_mapping_table.sh

@@ -3,7 +3,7 @@
 output_file=$(echo processed_mapping_illumina_$(date '+%Y%m%d').tab)
 
 # Removal of the first three columns of the mapping illumina tab file
-cut --complement -f1-3 mapping_*.tab > output_file
+cut --complement -f1-3 mapping_*.tab > $output_file
 mv $output_file mapping_illumina_$(date '+%Y%m%d').tab
 
 # Success message

bu_isciii/templates/viralrecon/RESULTS/lablog_viralrecon_results

@@ -25,7 +25,6 @@ ln -s ../../ANALYSIS/*/mapping_illumina*.tab ./mapping_illumina.csv
 ln -s ../../ANALYSIS/*/assembly_stats.csv ./assembly_stats.csv
 ln -s ../../ANALYSIS/*/01-PikaVirus-results/all_samples_virus_table_filtered.tsv ./pikavirus_table.tsv
 
-#conda activate viralrecon_report
 echo "python ./excel_generator.py -r ./references.tmp --merge_lineage_files" > _01_generate_excel_files.sh
 #Cleaning temp files and broken symbolic links
 echo "find . -xtype l -delete" > _02_clean_folders.sh

setup.py

@@ -2,7 +2,7 @@
 
 from setuptools import setup, find_packages
 
-version = "2.2.1"
+version = "2.2.2"
 
 with open("README.md") as f:
     readme = f.read()