0.3.1 (2022-08-23)#

  • Fixed bug related to displaying figure caption icons for cached datasets.

  • Added recursive dependency checks; if any upstream dependency of a figure script is a dataset or cached Zenodo deposit, an icon to that dataset/deposit will be displayed next to the figure.

  • Added functionality to obtain the Snakemake DAG prior to running any of the rules (used to infer upstream dependencies above). Metadata from the DAG is stored in the global config.

  • Reintroduced option to render the workflow DAG when building the article; users can now provide DAG-related settings under the dag entry in the config file (set render: true to enable the DAG generation). Note that showyourwork now depends on the graphviz package, so users may have to clear their ~/.showyourwork cache to force regeneration of the default showyourwork conda environment.

  • Workaround for issue with incomplete files not getting deleted when passing the --rerun-incomplete flag. Running showyourwork clean now deletes the .snakemake/incomplete directory.

  • Server errors on Zenodo / Zenodo Sandbox no longer lead to fatal errors when accessing the remote file cache.

  • Fixed issue with broken GitHub links in the PDF when using SSH to authenticate with GitHub.

  • Implemented better testing for showyourwork features; test repositories now persist on GitHub (showyourwork/test-**).

  • Implemented major fix for issue #124, which allows us to optimize the DAG by eliminating unnecessary jobs upstream of jobs with cache hits.

  • Users can now clone an empty remote repository prior to running showyourwork setup without triggering an error.

  • Added option to pass custom arguments to the tectonic build.

  • Removed dependency on fontspec, which caused issues with TT fonts. The shoyourwork logo is no longer displayed as text in a custom font, but as a PDF image.

  • Fix for newer versions of git that use main as the default branch name, which led to errors when integrating with Overleaf (which uses master).

  • Revamped the way tests for the showyourwork/showyourwork repo are run. Split tests into unit tests, local integration tests, and remote integration tests. The first two can be executed on either the base repo or any fork; the last one can only be executed on the base repo or upon the pull_request_target trigger from a PR, which occurs when a PR is labeled safe to test by a maintainer. This change will make contributing to showyourwork much easier going forward!

  • Changed the syntax for specifying the showyourwork version number in the showyourwork.yml config file. Users should now specify whether the version spec is a pip-installable version, a path to a local installation, or a ref on the GitHub repository (optionally on a given fork). The previous syntax (in which the version is provided as a string and the code attempts to interpret it as one of the above options) is deprecated, but will still be supported going forward. However, this change is still backwards-incompatible, as versions of showyourwork prior to 0.3.1 will not be able to build repositories that require versions greater than or equal to 0.3.1 (since previous versions of the code will not be able to parse the new version mapping syntax). Users can fix this by simply upgrading their local installation of showyourwork.

  • Removed built-in LaTeX class files for AAS journals, MNRAS, and A&A. Users who upgrade to 0.3.1 (or later) must now include all necessary class files and auxiliary TeX files in src/tex (and track them with git). This allows users to customize these files as needed and means we no longer have to maintain them as they get updated by the journals.

  • Removed the dependence on the marginnote package, which was causing issues with some other packages that also defined a \marginnote command. The margin link in the abstract now uses the built-in \marginpar.

0.3.0 (2022-06-22)#

  • Brand new release, featuring a complete re-write of the workflow. Below is a list of the major changes:

  • showyourwork! is now a pip-installable Python package. It is no longer a git submodule. The showyourwork.yml file must now specify which version of showyourwork to use.

  • There is no longer a Makefile. Articles should be built using the showyourwork command, which creates and activates a clean conda environment containing Snakemake and all dependencies needed to run the pipeline for each article. Users therefore no longer need to install Snakemake or mamba in the base environment.

  • New articles can now be created using the showyourwork setup command rather than via a GitHub repository template.

  • The syntax for many of the settings in the showyourwork.yml config file has changed, particularly for specifying Zenodo datasets.

  • The directory structure for article repositories has changed slightly. Figure scripts should now be placed in the src/scripts directory (used to be src/figures). The TeX files should now be placed in the src/tex directory (used to be src). Figure output files should now be generated in src/tex/figures (used to be src/figures).

  • To help with the transition to the new directory structure, new repositories include a file src/scripts/ that specifies absolute Pathlib paths to the main directories in the repository.

  • Figure scripts are no longer inferred from the \label command in LaTeX. Instead, users should specify the script associated with a given figure using the new \script command.

  • Added support for programmatically-generated files that can be included in the TeX manuscript via the new \variable command.

  • Users must now manually add \usepackage{showyourwork} to their LaTeX manuscript.

  • Overhauled the way showyourwork integrates with Zenodo. Static datasets should now be specified using their full DOI. Dynamic datasets are deprecated in favor of “cached” datasets. These are intermediate results that get cached on Zenodo Sandbox alongside a hash of the rules and all upstream dependencies used to generate them, making it possible to automatically restore results from the cloud in a way that preserves the full reproducibility of the workflow.

  • Implemented (experimental) integration with Overleaf projects, allowing users to pull changes to the manuscript and push changes to the figures.

  • Drastically improved the command-line interface, suppressing most of the noise generated by Snakemake in favor of succinct informational messages describing the build process. All messages now get logged to files in .showyourwork/logs for easier debugging. Similarly, improved error catching and added informational error messages for most of the failure modes of the workflow.

  • Several other tweaks, bugfixes, and improvements. Lots of changes to the back end to make showyourwork easier to develop, maintain, and extend!

0.2.3 (2022-02-21)#

  • Bump jinja2 version to fix issue with markupsafe

  • Changelog: v0.2.2…v0.2.3

0.2.2 (2022-01-05)#

  • Tweaks to the logo (now an actual font!)

  • Added option to exclude files from article cache on CI

  • Fixed behavior of figures labeled with an asterisk (e.g., \label{fig*:...})

  • Changelog: v0.2.1…v0.2.2

0.2.1 (2021-12-18)#

  • Implemented custom DAG generation (cleaner, prettier)

  • Added several entries to the Projects page on the docs

  • Streamlined arXiv tarball generation step

  • Added a basic make lint command to check for repo issues

  • Users can now disable caching on CI by setting the cache number to null

  • Switch to installing graphviz with conda

  • Pinned all conda dependencies to specific versions

  • Changelog: v0.2.0…v0.2.1

0.2.0 (2021-12-07)#

  • Major changes to the Zenodo interface! Please read the Zenodo section of the documentation on the showyourwork.yml file for details on what changed. The tl;dr is that all deposits now require either a concept or a version id (read more about that here:; this id now uniquely identifies the deposit (previously, we relied on the uniqueness of the deposit title & creators).

  • Added support for non-Python scripts to generate datasets and other dependencies

  • Fixed issue with large datasets lingering in the arxiv tarball

  • Added a make fast option to reproduce the results locally w/out running expensive steps

  • Signficiant improvements to the documentation, now with detailed info on LaTeX features

  • Added a marginicon command for custom margin icons next to figures

  • Added a progress bar to Zenodo uploads

  • Now re-downloading datasets on GitHub Actions if a newer version is available

  • Better path resolution when extracting tarballs

  • Changelog: v0.1.35…v0.2.0

0.1.35 (2021-11-22)#

  • Fixed issue with unnecessary downloads of Zenodo datasets on CI.

  • Fixed issue that prevented downloading the Zenodo datasets if the Zenodo API key belonged to someone other than the creator of the deposit.

  • Fixed issue that caused the conda env creation to fail. We are now installing snakemake-minimal and pinning the mamba version; this is only a temporary solution.

  • Changelog: v0.1.34…v0.1.35

0.1.34 (2021-11-18)#

  • Now uploads a build artifact whenever the workflow fails on CI for easier debugging.

  • Changelog: v0.1.33…v0.1.34

0.1.33 (2021-11-17)#

0.1.32 (2021-11-17)#

0.1.31 (2021-11-15)#

  • Migrated to the new Zenodo API; previously the limit for uploading files was 100 MB (now 50 GB).

  • Changelog: v0.1.30…v0.1.31

0.1.30 (2021-11-12)#

  • Added an example on how to use jinja templating to simplify the showyourwork.yml config file.

  • Developers: Undo the reset build cache operation from the previous patch, since this causes race conditions when accessing the cache during the unit tests (since we are concurrently running dozens of actions on a single repo!)

  • Changelog: v0.1.29…v0.1.30

0.1.29 (2021-11-10)#

  • Support for creation/download of Zenodo tarballs.

  • Implements the idea in #48 for specifying custom manuscript dependencies.

  • Bugfix for rules that subclass the main showyourwork figure rule.

  • Implements the idea in #47 for custom Zenodo dataset generation.

  • Developers: Now resetting the build cache before each unit test on showyourwork-example and then re-running the cached build.

  • Changelog: v0.1.28…v0.1.29

0.1.28 (2021-11-09)#

  • Added support for non-Python scripts; users can now define instructions in the YAML config file to execute other kinds of scripts.

  • Implemented better error messages when figure scripts fail.

  • Allow users to specify a graphicspath for all figures in the document.

  • Allow users to customize the name of the manuscript (it no longer needs to be called ms.tex).

  • Changelog: v0.1.27…v0.1.28

0.1.27 (2021-11-03)#

  • Added support for installing a minimal TeX distribution so that TeX can be rendered in matplotlib; see Custom workflows.

  • Changelog: v0.1.26…v0.1.27

0.1.26 (2021-11-02)#

0.1.25 (2021-11-02)#

  • Fixed issue that prevented ORCID badges from showing up when building the PDF on GitHub Actions

  • Changelog: v0.1.24…v0.1.25

0.1.24 (2021-11-02)#

  • Fixed issue with os.get_terminal_size breaking CI builds when displaying error messages

  • Changelog: v0.1.23…v0.1.24

0.1.23 (2021-11-02)#

  • Added explicit support for MNRAS and A&A LaTeX document classes

  • Improved support for new Apple M1 chips

  • Fixed options clash for package hyperref

  • Changelog: v0.1.22…v0.1.23

0.1.22 (2021-11-02)#

  • Updated LaTeX package fontawesome to fontawesome5

  • Developers: Can now run tests on PR branches to generate showyourwork-example-dev branches

  • Changelog: v0.1.21…v0.1.22

0.1.21 (2021-11-01)#

  • Fixed minor issue with error messages for custom figures

  • Improved documentation page on projects that use showyourwork

  • Changelog: v0.1.20…v0.1.21

0.1.20 (2021-10-28)#

  • Fixed issue with figure link formatting when enabling linenumbers in AASTeX

  • Made arxiv_tarball_exclude paths relative to the repository root

  • Added a make update option to update showyourwork to the latest release.

  • Changelog: v0.1.19…v0.1.20

0.1.19 (2021-10-25)#

0.1.18 (2021-10-25)#

  • Added more informative error messages that are displayed at the very end of the build logs. Still more work to be done on this front, but error logs should now be much easier to parse.

  • Implemented the new Zenodo config structure in the showyourwork.yml file, as per #31.

  • Changelog: v0.1.17…v0.1.18

0.1.17 (2021-10-22)#

  • Changed the way Zenodo dependencies are provided in the showyourwork.yml file. Dependencies like datasets should still be listed as entries under the corresponding figure scripts in figure_dependencies, but all information on how to generate or download them should now go in a separate top-level zenodo: key. This makes it much easier to, e.g., specify datasets used by multiple figures. Please see the Custom workflows section of the docs for more information.

  • Improved the API documentation.

  • Changelog: v0.1.16…v0.1.17

0.1.16 (2021-10-22)#

  • Template repo update: Pared down the Makefile in the template repository. This now calls a Makefile in the showyourwork submodule (this repo), which contains all the directives. This makes it easier to improve/update the workflow, since we can just update showyourwork.

  • Changelog: v0.1.15…v0.1.16

0.1.15 (2021-10-21)#

  • Template repo update: Added options to the Makefile to generate a report and a DAG. Added a submodule setup check; if the user didn’t init the showyourwork submodule, does it automatically before building.

  • Changelog: v0.1.14…v0.1.15

0.1.14 (2021-10-21)#

0.1.13 (2021-10-21)#

  • Fixed API documentation

  • Fixed error with arxiv_tarball_exclude and arxiv tarball issue (#21)

  • Changelog: v0.1.12…v0.1.13

0.1.12 (2021-10-20)#

  • Revert code that prevents the Snakefile from being loaded more than once. Turns out that is expected behavior, and is required in order for the module import syntax to work!

  • Switched to adding checks within the script to prevent dependencies from getting ingested multiple times.

  • Changelog: v0.1.11…v0.1.12

0.1.11 (2021-10-20)#

  • Fix bug preventing figures from being cached properly when one script generates multiple figures

  • Fixed issues due to Snakefile being loaded multiple times

  • Auto-populate the projects page on the docs via a GitHub API search on every release

  • Changelog: v0.1.10…v0.1.11

0.1.10 (2021-10-20)#

  • Cleaned up the workflow, separating rules into their own files with better documentation.

  • Added a fix for nested figures (figures under subdirectories in the src/figures folder).

  • Fixed issue with multiple Zenodo datasets causing the build to fail.

  • Added support for figures in figure* environments.

  • Fixed issue with occasional missing </HTML> closing tags in the showyourwork XML tree.

  • Added some API documentation; more coming soon.

  • Changelog: v0.1.9…v0.1.10

0.1.9 (2021-10-18)#

  • Template repo update: Added a Makefile for quick article generation; added docs on how to use it.

  • Changelog: v0.1.8…v0.1.9

0.1.8 (2021-10-18)#

  • Added “One script, multiple figures” example

  • Improved the documentation for script dependencies and datasets

  • Fixed a bug when downloading deposits from Zenodo

  • Added release testing

  • Changelog: v0.1.7…v0.1.8

0.1.7 (2021-10-18)#

  • Added explicit support for Zenodo-hosted datasets.

  • Template repo update: Added the environment variable ZENODO_TOKEN to .github/workflows/showyourwork.yml.

  • Changelog: v0.1.6…v0.1.7

0.1.6 (2021-10-14)#

  • Added documentation for the expensive-figure example.

  • Changelog: v0.1.5…v0.1.6

0.1.5 (2021-10-14)#

  • Added the expensive-figure example for computationally expensive figure generation.

  • Changelog: v0.1.4…v0.1.5

0.1.4 (2021-10-13)#

  • Initial release of the workflow.