v1.13 - Copper Crocodile

When we do releases, we sometimes like to go big. This is one of those times.

This is a truly huge release of nf-core/tools with a massive amount of new functionality. Please be sure to read the changelog and updated readme carefully!

Note to developers - please run and push the results of nf-core lint --fix before attempting to merge the automated synchronisation PR. This should reduce the size of the diff and make merging a lot easier.

The template and associated updates in this release is for Nextflow DSL1 pipelines. This is likely to be one of the last releases before we officially switch to DSL2. For those of you already with DSL2 pipelines you may need to treat the automated merge PR very carefully. Ask for help on Slack if in any doubt.

Template

• Major new feature - Validation of pipeline parameters [#426]
  • The addition runs as soon as the pipeline launches and checks the pipeline input parameters two main things:
    • No parameters are supplied that share a name with core Nextflow options (eg. --resume instead of -resume)
    • Supplied parameters validate against the pipeline JSON schema (eg. correct variable types, required values)
  • If either parameter validation fails or the pipeline has errors, a warning is given about any unexpected parameters found which are not described in the pipeline schema.
  • This behaviour can be disabled by using --validate_params false
• Added profiles to support the Charliecloud and Shifter container engines [#824]
  • Note that Charliecloud requires Nextflow version v21.03.0-edge or later.
• Profiles for container engines now explicitly disable all other engines [#867]
• Fixed typo in nf-core-lint CI that prevented the markdown summary from being automatically posted on PRs as a comment.
• Changed default for --input from data/*{1,2}.fastq.gz to null, as this is now validated by the schema as a required value.
• Removed support for --name parameter for custom run names.
  • The same functionality for MultiQC still exists with the core Nextflow -name option.
• Added to template docs about how to identify process name for resource customisation
• The parameters --max_memory and --max_time are now validated against a regular expression [#793]
  • Must be written in the format 123.GB / 456.h with any of the prefixes listed in the Nextflow docs
  • Bare numbers no longer allowed, avoiding people from trying to specify GB and actually specifying bytes.
• Switched from cookiecutter to Jinja2 [#880]
• Finally dropped the wonderful cookiecutter library that was behind the first pipeline template that led to nf-core [#880]
  • Now rendering templates directly using Jinja, which is what cookiecutter was doing anyway

Modules

Initial addition of a number of new helper commands for working with DSL2 modules:

• modules list - List available modules
• modules install - Install a module from nf-core/modules
• modules remove - Remove a module from a pipeline
• modules create - Create a module from the template
• modules create-test-yml - Create the test.yml file for a module with md5 sums, tags, commands and names added
• modules lint - Check a module against nf-core guidelines

You can read more about each of these commands in the main tools documentation (see README.md or https://nf-co.re/tools)

Tools helper code

• Fixed some bugs in the command line interface for nf-core launch and improved formatting [#829]
• New functionality for nf-core download to make it compatible with DSL2 pipelines [#832]
  • Singularity images in module files are now discovered and fetched
  • Direct downloads of Singularity images in python allowed (much faster than running singularity pull)
  • Downloads now work with $NXF_SINGULARITY_CACHEDIR so that pipelines sharing containers have efficient downloads
• Changed behaviour of nf-core sync command [#787]
  • Instead of opening or updating a PR from TEMPLATE directly to dev, a new branch is now created from TEMPLATE and a PR opened from this to dev.
  • This is to make it easier to fix merge conflicts without accidentally bringing the entire pipeline history back into the TEMPLATE branch (which makes subsequent sync merges much more difficult)

Linting

• Major refactor and rewrite of pipieline linting code
  • Much better code organisation and maintainability
  • New automatically generated documentation using Sphinx
  • Numerous new tests and functions, removal of some unnecessary tests
• Added lint check for merge markers [#321]
• Added new option --fix to automatically correct some problems detected by linting
• Added validation of default params to nf-core schema lint [#823]
• Added schema validation of GitHub action workflows to lint function [#795]
• Fixed bug in schema title and description validation
• Added second progress bar for conda dependencies lint check, as it can be slow [#299]
• Added new lint test to check files that should be unchanged from the pipeline.
• Added the possibility to ignore lint tests using a nf-core-lint.yml config file [#809]

1.12.1 - Silver Dolphin

Minor release, primarily to fix a problem introduced in v1.12 where not all instances of baseDir had been switched with projectDir. However, the elves have been busy and managed to squeeze in a few extra goodies along with the fix:

Template

• Finished switch from $baseDir to $projectDir in iGenomes.conf and main.nf
  • Main fix is for smail_fields which was a bug introduced in the previous release. Sorry about that!
• Ported a number of small content tweaks from nf-core/eager to the template [#786]
  • Better contributing documentation, more placeholders in documentation files, more relaxed markdownlint exceptions for certain HTML tags, more content for the PR and issue templates.

Tools helper code

• Pipeline schema: make parameters of type range to number. [#738]
• Respect $NXF_HOME when looking for pipelines with nf-core list [#798]
• Swapped PyInquirer with questionary for command line questions in launch.py [#726]
  • This should fix conda installation issues that some people had been hitting
  • The change also allows other improvements to the UI
• Fix linting crash when a file deleted but not yet staged in git [#796]

1.12 - Mercury Weasel

A bit of a fast cycle on this release, it is needed urgently because GitHub Actions CI tests will be failing on all pipelines due to the set-env and add-path commands being depreciated (see the GitHub blog post for details).

Tools helper code

• Updated nf_core documentation generator for building https://nf-co.re/tools-docs/

Template

• Make CI comments work with PRs from forks [#765]
  • Branch protection and linting results should now show on all PRs
• Updated GitHub issue templates, which had stopped working
• Refactored GitHub Actions so that the AWS full-scale tests are triggered after docker build is finished
  • DockerHub push workflow split into two - one for dev, one for releases
• Updated actions to no longer use set-env which is now depreciating [#739]
• Added config import for test_full in nextflow.config
• Switched depreciated $baseDir to $projectDir
• Updated minimum Nextflow version to 20.04.10
• Make Nextflow installation less verbose in GitHub Actions [#780]

Linting

• Updated code to display colours in GitHub Actions log output
• Allow tests to pass with dev version of nf-core/tools (previous failure due to base image version)
• Lint code no longer tries to post GitHub PR comments. This is now done in a GitHub Action only.

1.11 - Iron Tiger

Template

• Fix command error in awstest.yml GitHub Action workflow.
• Allow manual triggering of AWS test GitHub Action workflows.
• Remove TODO item, which was proposing the usage of additional files beside usage.md and output.md for documentation.
• Added a Podman profile, which enables Podman as container.
• Updated linting for GitHub actions AWS tests workflows.

Linting

• Made a base-level Dockerfile a warning instead of failure
• Added a lint failure if the old bin/markdown_to_html.r script is found
• Update rich package dependency and use new markup escaping to change [[!]] back to [!] again

Other

• Pipeline sync - fetch full repo when checking out before sync
• Sync - Add GitHub actions manual trigger option

1.10.2 - Copper Camel [brought back from the dead]

Second patch release to address some small errors discovered in the pipeline template.
Apologies for the inconvenience.

• Fix syntax error in /push_dockerhub.yml GitHub Action workflow
• Change params.readPaths -> params.input_paths in test_full.config
• Check results when posting the lint results as a GitHub comment
  • This feature is unfortunately not possible when making PRs from forks outside of the nf-core organisation for now.
• More major refactoring of the automated pipeline sync
  • New GitHub Actions matrix parallelisation of sync jobs across pipelines [#673]
  • Removed the --all behaviour from nf-core sync as we no longer need it
  • Sync now uses a new list of pipelines on the website which does not include archived pipelines [#712]
  • When making a PR it checks if a PR already exists - if so it updates it [#710]
  • More tests and code refactoring for more stable code. Hopefully fixes 404 error [#711]

1.10.1 - Copper Camel [patch]

Patch release to fix the automatic template synchronisation, which failed in the v1.10 release.

• Improved logging: nf-core --log-file log.txt now saves a verbose log to disk.
• nf-core/tools GitHub Actions pipeline sync now uploads verbose log as an artifact.
• Sync - fixed several minor bugs, made logging less verbose.
• Python Rich library updated to >=4.2.1
• Hopefully fix git config for pipeline sync so that commit comes from @nf-core-bot
• Fix sync auto-PR text indentation so that it doesn't all show as code
• Added explicit flag --show-passed for nf-core lint instead of taking logging verbosity

1.10 - Copper Camel

Pipeline schema

This release of nf-core/tools introduces a major change / new feature: pipeline schema.
These are JSON Schema files that describe all of the parameters for a given
pipeline with their ID, a description, a longer help text, an optional default value, a variable type
(eg. string or boolean) and more.

The files will be used in a number of places:

• Automatic validation of supplied parameters when running pipelines
  • Pipeline execution can be immediately stopped if a required param is missing,
    or does not conform to the patterns / allowed values in the schema.
• Generation of pipeline command-line help
  • Running nextflow run <pipeline> --help will use the schema to generate a help text automatically
• Building online documentation on the nf-core website
• Integration with 3rd party graphical user interfaces

To support these new schema files, nf-core/tools now comes with a new set of commands: nf-core schema.

• Pipeline schema can be generated or updated using nf-core schema build - this takes the parameters from
  the pipeline config file and prompts the developer for any mismatch between schema and pipeline.
  • Once a skeleton Schema file has been built, the command makes use of a new nf-core website tool to provide
    a user friendly graphical interface for developers to add content to their schema: https://nf-co.re/pipeline_schema_builder
• Pipelines will be automatically tested for valid schema that describe all pipeline parameters using the
  nf-core schema lint command (also included as part of the main nf-core lint command).
• Users can validate their set of pipeline inputs using the nf-core schema validate command.

In addition to the new schema commands, the nf-core launch command has been completely rewritten from
scratch to make use of the new pipeline schema. This command can use either an interactive command-line
prompt or a rich web interface to help users set parameters for a pipeline run.

The parameter descriptions and help text are fully used and embedded into the launch interfaces to make
this process as user-friendly as possible. We hope that it's particularly well suited to those new to nf-core.

Whilst we appreciate that this new feature will add a little work for pipeline developers, we're excited at
the possibilities that it brings. If you have any feedback or suggestions, please let us know either here on
GitHub or on the nf-core #json-schema Slack channel.

Python code formatting

We have adopted the use of the Black Python code formatter.
This ensures a harmonised code formatting style throughout the package, from all contributors.
If you are editing any Python code in nf-core/tools you must now pass the files through Black when
making a pull-request. See .github/CONTRIBUTING.md for details.

Template

• Add --publish_dir_mode parameter #585
• Isolate R library paths to those in container #541
• Added new style of pipeline parameters JSON schema to pipeline template
• Add ability to attach MultiQC reports to completion emails when using mail
• Update output.md and add in 'Pipeline information' section describing standard NF and pipeline reporting.
• Build Docker image using GitHub Actions, then push to Docker Hub (instead of building on Docker Hub)
• Add Slack channel badge in pipeline README
• Allow multiple container tags in ci.yml if performing multiple tests in parallel
• Add AWS CI tests and full tests GitHub Actions workflows
• Update AWS CI tests and full tests secrets names
• Added macs_gsize for danRer10, based on this post
• Add information about config files used for workflow execution (workflow.configFiles) to summary
• Fix markdown_to_html.py to work with Python 2 and 3.
• Change params.reads -> params.input
• Change params.readPaths -> params.input_paths
• Added a .github/.dockstore.yml config file for automatic workflow registration with dockstore.org

Linting

• Refactored PR branch tests to be a little clearer.
• Linting error docs explain how to add an additional branch protecton rule to the branch.yml GitHub Actions workflow.
• Adapted linting docs to the new PR branch tests.
• Failure for missing the readme bioconda badge is now a warn, in case this badge is not relevant
• Added test for template {{ cookiecutter.var }} placeholders
• Fix failure when providing version along with build id for Conda packages
• New --json and --markdown options to print lint results to JSON / markdown files
• Linting code now automatically posts warning / failing results to GitHub PRs as a comment if it can
• Added AWS GitHub Actions workflows linting
• Fail if params.input isn't defined.
• Beautiful new progress bar to look at whilst linting is running and awesome new formatted output on the command line 😍
  • All made using the excellent rich python library - check it out!
• Tests looking for TODO strings should now ignore editor backup files. #477

nf-core/tools Continuous Integration

• Added CI test to check for PRs against master in tools repo
• CI PR branch tests fixed & now automatically add a comment on the PR if failing, explaining what is wrong
• Move some of the issue and PR templates into HTML <!-- comments --> so that they don't show in issues / PRs

Other

• Describe alternative installation method via conda with conda env create
• nf-core/tools version number now printed underneath header artwork
• Bumped Conda version shipped with nfcore/base to 4.8.2
• Added log message when creating new pipelines that people should talk to the community about their plans
• Fixed 'on completion' emails sent using the mail command not containing body text.
• Improved command-line help text for nf-core/tools
• nf-core list now hides archived pipelines unless --show_archived flag is set
• Command line tools now checks if there is a new version of nf-core/tools available
  • Disable this by setting the environment variable NFCORE_NO_VERSION_CHECK, eg. export NFCORE_NO_VERSION_CHECK=1
• Better command-line output formatting of nearly all nf-core commands using rich

1.9 - Platinum Pigeon

Continuous integration

• Travis CI tests are now deprecated in favor of GitHub Actions within the pipeline template.
  • nf-core bump-version support has been removed for .travis.yml
  • nf-core lint now fails if a .travis.yml file is found
• Ported nf-core/tools Travis CI automation to GitHub Actions
• Fixed the build for the nf-core/tools API documentation on the website

Template

• Rewrote the documentation markdown > HTML conversion in Python instead of R
• Removed the requirement for R in the conda environment
• Make params.multiqc_config give an additional MultiQC config file instead of replacing the one that ships with the pipeline
• Ignore only tests/ and testing/ directories in .gitignore to avoid ignoring test.config configuration file
• Rephrase docs to promote usage of containers over Conda to ensure reproducibility
• Stage the workflow summary YAML file within MultiQC work directory

Linting

• Removed linting for CircleCI
• Allow any one of params.reads or params.input or params.design before warning
• Added whitespace padding to lint error URLs
• Improved documentation for lint errors
• Allow either >= or !>= in nextflow version checks (the latter exits with an error instead of just warning) #506
• Check that manifest.version ends in dev and throw a warning if not
  • If running with --release check the opposite and fail if not
• Tidied up error messages and syntax for linting GitHub actions branch tests
• Add YAML validator
• Don't print test results if we have a critical error

Other

• Fix automatic synchronisation of the template after releases of nf-core/tools
• Improve documentation for installing nf-core/tools
• Replace preprint by the new nf-core publication in Nature Biotechnology 🍾
• Use stderr instead of stdout for header artwork
• Tolerate unexpected output from nextflow config command
• Add social preview image
• Added a release checklist for the tools repo

1.8 - Black Sheep

Continuous integration

• GitHub Actions CI workflows are now included in the template pipeline
  • Please update these files to match the existing tests that you have in .travis.yml
• Travis CI tests will be deprecated from the next tools release
• Linting will generate a warning if GitHub Actions workflows do not exist and if applicable to remove Travis CI workflow file i.e. .travis.yml.

Tools helper code

• Refactored the template synchronisation code to be part of the main nf-core tool
• nf-core bump-version now also bumps the version string of the exported conda environment in the Dockerfile
• Updated Blacklist of synced pipelines
• Ignore pre-releases in nf-core list
• Updated documentation for nf-core download
• Fixed typo in nf-core launch final command
• Handle missing pipeline descriptions in nf-core list

Linting

• Adjusted linting to enable patch branches from being tested
• Warn if GitHub Actions workflows do not exist, warn if .travis.yml and circleCI are there
• Lint for Singularity file and raise error if found #458
• Added linting of GitHub Actions workflows linting.yml, ci.yml and branch.yml
• Warn if pipeline name contains upper case letters or non alphabetical characters #85
• Make CI tests of lint code pass for releases

Template pipeline

• Fixed incorrect paths in iGenomes config as described in issue #418
• Fixed incorrect usage of non-existent parameter in the template #446
• Add UCSC genomes to igenomes.config and add paths to all genome indices
• Change maxMultiqcEmailFileSize parameter to max_multiqc_email_size
• Export conda environment in Docker file #349
• Change remaining parameters from camelCase to snake_case #39
  • --singleEnd to --single_end
  • --igenomesIgnore to --igenomes_ignore
  • Having the old camelCase versions of these will now throw an error
• Add autoMounts=true to default singularity profile
• Add in markdownlint checks that were being ignored by default
• Disable ansi logging in the travis CI tests
• Move paramssection from base.config to nextflow.config
• Use env scope to export PYTHONNOUSERSITE in nextflow.config to prevent conflicts with host Python environment
• Bump minimum Nextflow version to 19.10.0 - required to properly use env scope in nextflow.config
• Added support for nf-tower in the travis tests, using public mailbox [email protected]
• Add link to Keep a Changelog and Semantic Versioning to CHANGELOG
• Adjusted .travis.yml checks to allow for patch branches to be tested
• Add Python 3.7 dependency to the environment.yml file
• Remove awsbatch profile cf nf-core/configs#71
• Make scrape_software_versions.py compatible with Python3 to enable miniconda3 in base image PR
• Add GitHub Actions workflows and respective linting
• Add NXF_ANSI_LOG as global environment variable to template GitHub Actions CI workflow
• Fixed global environment variable in GitHub Actions CI workflow
• Add --awscli parameter
• Add README.txt path for genomes in igenomes.config nf-core/atacseq#75
• Fix buggy ANSI codes in pipeline summary log messages
• Add a TODO line in the new GitHub Actions CI test files

Base Docker image

• Use miniconda3 instead of miniconda for a Python 3k base environment
  • If you still need Python 2 for your pipeline, add conda-forge::python=2.7.4 to the dependencies in your environment.yml
• Update conda version to 4.7.12

Other

• Updated Base Dockerfile to Conda 4.7.10
• Entirely switched from Travis-Ci.org to Travis-Ci.com for template and tools
• Improved core documentation (-profile)

1.7 - Titanium Kangaroo

PyPI package description

• The readme should now be rendered properly on PyPI.

Tools helper code

• The tools create command now sets up a TEMPLATE and a dev branch for syncing
• Fixed issue 379
• nf-core launch now uses stable parameter schema version 0.1.0
• Check that PR from patch or dev branch is acceptable by linting
• Made code compatible with Python 3.7
• The download command now also fetches institutional configs from nf-core/configs
• When listing pipelines, a nicer message is given for the rare case of a detached HEAD ref in a locally pulled pipeline. #297
• The download command can now compress files into a single archive.
• nf-core create now fetches a logo for the pipeline from the nf-core website

Syncing

• Can now sync a targeted pipeline via command-line
• Updated Blacklist of synced pipelines
• Removed chipseq from Blacklist of synced pipelines
• Fixed issue #314

Linting

• If the container slug does not contain the nf-core organisation (for example during development on a fork), linting will raise a warning, and an error with release mode on

Template

• Add new code for Travis CI to allow PRs from patch branches too
• Fix small typo in central readme of tools for future releases
• Small code polishing + typo fix in the template main.nf file
• Header ANSI codes no longer print [2m to console when using -with-ansi
• Switched to yaml.safe_load() to fix PyYAML warning that was thrown because of a possible exploit
• Add nf-core citation
• Add proper nf-core logo for tools
• Add Quick Start section to main README of template
• Fix Docker RunOptions to get UID and GID set in the template
• Dockerfile now specifically uses the proper release tag of the nfcore/base image
• Use file instead of new File
  to avoid weird behavior such as making an s3:/ directory locally when using
  an AWS S3 bucket as the --outdir.
• Fix workflow.onComplete() message when finishing pipeline
• Update URL for joining the nf-core slack to https://nf-co.re/join/slack
• Increased default time limit to 4h
• Add direct link to the pipeline slack channel in the contribution guidelines
• Add contributions and support heading with links to contribution guidelines and link to the pipeline slack channel in the main README
• Fix Parameters JSON due to new versionized structure
• Added conda-forge::r-markdown=1.1 and conda-forge::r-base=3.6.1 to environment
• Plain-text email template now has nf-core ASCII artwork
• Template configured to use logo fetched from website
• New option --email_on_fail which only sends emails if the workflow is not successful
• Add file existence check when checking software versions
• Fixed issue [https://github.com//issues/165] - Use checkIfExists
• Consistent spacing for if statements
• Add sensible resource labels to base.config

Other

• Bump conda to 4.6.14 in base nf-core Dockerfile
• Added a Code of Conduct to nf-core/tools, as only the template had this before
• TravisCI tests will now also start for PRs from patch branches, to allow fixing critical issues without making a new major release