Software Metadata Extraction Framework (SOMEF)

A command line interface for automatically extracting relevant metadata from code repositories (readme, configuration files, documentation, etc.).

Demo: See a demo running somef as a service, through the SOMEF-Vider tool.

Authors: Daniel Garijo, Allen Mao, Miguel Ángel García Delgado, Haripriya Dharmala, Vedant Diwanji, Jiaying Wang, Aidan Kelley, Jenifer Tabita Ciuciu-Kiss, Luca Angheluta, Juanje Mendoza, Anas El Hounsri and Thomas Vuillaume.

Features

Given a readme file (or a GitHub/Gitlab/Codeberg/Bitbucket repository) SOMEF will extract the following categories (if present), listed in alphabetical order:

• Acknowledgement: Text acknowledging funding sources or contributors
• Application domain: The application domain of the repository. Current supported domains include: Astrophysics, Audio, Computer vision, Graphs, Natural language processing, Reinforcement learning, Semantc web, Sequential. Domains are not mutually exclusive. These domains have been extracted from awesome lists and Papers with code. Find more information in our documentation
• Authors: Person(s) or organization(s) responsible for the project. We recognize the following properties:
  • Name: name of the author (including last name)
  • Given name: First name of an author
  • Family name: Last name of an author
  • Email: email of author
  • URL: website or ORCID associated with the author
• Application type: type of software (command line application, notebook, ontology, scientific workflow, etc.)
• Build file: Build file(s) of the project. For example, files used to create a Docker image for the target software, package files, etc.
• Citation: Preferred citation(s) as the authors have stated in their readme file. SOMEF recognizes Bibtex, Citation File Format files and other means by which authors cite their papers (e.g., by in-text citation). For CITATION.cff files, SOMEF now generates two separate entries: one for the software tool and another for the preferred citation (if available). This ensures metadata like DOI or version is correctly assigned to each entity. SOMEF now performs citation reconciliation: scholarly publications (articles) are assigned in codemeta to referencePublication, while the software itself is credited in creditText. (See https://somef.readthedocs.io/en/latest/output/#codemeta-format). When using -e, publication metadata is enriched via OpenAlex. We recognize the following properties:
  • Title: Title of the publication
  • Author: list of author names in the publication
  • URL: URL of the publication
  • DOI: Digital object identifier of the publication
  • Date published
  • Journal: Journal name where the paper was published
  • Year: Year of publication
  • Pages: Page range in the journal
  • openalex_id: OpenAlex ID of the publication
• Code of conduct: Link to the code of conduct of the project
• Code repository: Link to the GitHub/GitLab/Codeberg and Bitbucket repository used for the extraction
• Contact: Contact person responsible for maintaining a software component
• Continuous integration: Link to continuous integration service(s)
• Contribution guidelines: Text indicating how to contribute to this code repository
• Contributor: Contributors to a software component. Note: Contributor metadata is exported from metadata files (e.g., CodeMeta, CONTRIBUTORS, etc.) not from git logs.
• Creation date: Date when the repository was created
• Copyright holder: Entity or individual owning the rights to the software. The year is also extracted, if available.
• Date updated: Date of last release.
• Description: A description of what the software does
• Documentation: Where to find additional documentation about a software component
• Download URL: URL where to download the target software (typically the installer, package or a tarball to a stable version)
• Executable examples: Jupyter notebooks ready for execution (e.g., files, or through myBinder/colab links)
• FAQ: Frequently asked questions about a software component
• Features: List of features or capabilities of the software
• Forks count: Number of forks of the project
• Forks url: Links to forks made of the project
• Full name: Name + owner (owner/name)
• Full title: If the repository is a short name, we will attempt to extract the longer version of the repository name
• Funding: Funding information associated with the project. Note: This information is extracted from codemeta.json files within the repository and from the README (funding section headers and links to crowdfunding platforms). When using -e, the project data is enriched with OpenAIRE, adding:
• project_code: Project code
• project_title: Project title
• project_acronym: Project acronym
• grant_id: Call/grant identifier
• Identifier: Identifier associated with the software (if any), such as Digital Object Identifiers and Software Heritage identifiers (SWH). DOIs associated with publications will also be detected. When using -e, the following enrichment identifiers are also added:
• openaire_id: URL to the OpenAIRE explore page for the software
• swhid: Software Heritage identifier (for Zenodo DOIs)
• Images: Images used to illustrate the software component
• Installation instructions: A set of instructions that indicate how to install a target repository
• Invocation: Execution command(s) needed to run a scientific software component
• Issue tracker: Link where to open issues for the target repository
• Keywords: set of terms used to commonly identify a software component
• License: License and usage terms of a software component. Now we also extract license from citation.cff.
• Logo: Main logo used to represent the target software component
• Maintainer: Individuals or teams responsible for maintaining the software component, extracted from the CODEOWNERS file
• Name: Name identifying a software component
• Ontologies: Structured representation of the ontologies present in the repository
• Owner: Name and type of the user or organization in charge of the repository
• Package distribution: Links to package sites like pypi in case the repository has a package available.
• Package files: Links to package files used to wrap the project in a package.
• Programming languages: Languages used in the repository
• Related papers: URL to possible related papers within the repository stated within the readme file (from Arxiv)
• Releases: Pointer to the available versions of a software component. For each release, somef will track the following properties:
  • Description: Release notes
  • Author: Agent responsible of creating the release
  • Name: Name of the release
  • Tag: version number of the release
  • Date of publication
  • Date of creation
  • Link to the html page of the release
  • Id of the release
  • Link to the tarball zip and code of the release
• Repository status: Repository status as it is described in repostatus.org.
• Requirements: Pre-requisites and dependencies needed to execute a software component
• Run: Running instructions of a software component. It may be wider than the invocation category, as it may include several steps and explanations.
• Runtime platform: specifies the runtime environment or script interpreter dependencies (e.g., Python, Java).
• Script files: Bash script files contained in the repository
• Stargazers count: Total number of stargazers of the project
• Support: Guidelines and links of where to obtain support for a software component
• Support channels: Help channels one can use to get support about the target software component
• Usage examples: Assumptions and considerations recorded by the authors when executing a software component, or examples on how to use it
• Workflows: URL and path to the computational workflow files present in the repository

We use different supervised classifiers, header analysis, regular expressions, the GitHub/Gitlab/Codeberg and Bitbucket API to retrieve all these fields (more than one technique may be used for each field) and language specific metadata parsers (e.g., for package files). Each extraction records its provenance, with the confidence and technique used on each step. For more information check the output format description

Confidence values in header analysis

When extracting metadata through header analysis, SOMEF filters out headers whose confidence is below a certain threshold to avoid false positives. For instance, a header with 11+ words receives a confidence of 0.1, which is considered too low for a reliable classification — such headers are discarded from the results. The filtering ensures that only headers with a reasonable match quality are reported in the output.

Documentation

See full documentation at https://somef.readthedocs.io/en/latest/

Cite SOMEF:

Journal publication (preferred):

@article{10.1162/qss_a_00167,
    author = {Kelley, Aidan and Garijo, Daniel},
    title = "{A Framework for Creating Knowledge Graphs of Scientific Software Metadata}",
    journal = {Quantitative Science Studies},
    pages = {1-37},
    year = {2021},
    month = {11},
    issn = {2641-3337},
    doi = {10.1162/qss_a_00167},
    url = {https://doi.org/10.1162/qss_a_00167},
    eprint = {https://direct.mit.edu/qss/article-pdf/doi/10.1162/qss\_a\_00167/1971225/qss\_a\_00167.pdf},
}

Conference publication (first):

@INPROCEEDINGS{9006447,
author={A. {Mao} and D. {Garijo} and S. {Fakhraei}},
booktitle={2019 IEEE International Conference on Big Data (Big Data)},
title={SoMEF: A Framework for Capturing Scientific Software Metadata from its Documentation},
year={2019},
doi={10.1109/BigData47090.2019.9006447},
url={http://dgarijo.com/papers/SoMEF.pdf},
pages={3032-3037}
}

Requirements

• Python 3.11 + (default version support). Python 3.9 and 3.10 will work, but are not supported anymore.

SOMEF has been tested on Unix, MacOS and Windows Microsoft operating systems.

If you face any issues when installing SOMEF, please make sure you have installed the following packages: build-essential, libssl-dev, libffi-dev and python3-dev.

Install from Pypi

SOMEF is available in Pypi! To install it just type:

pip install somef

NLTK data

SOMEF-Core uses NLTK / WordNet to match README section headers (e.g. "Getting Started", "Citation") against metadata categories. Running somef configure (see below) takes care of downloading the required NLTK data automatically.

Install from GitHub

To run SOMEF, please follow the next steps:

Clone this GitHub repository

git clone https://github.com/KnowledgeCaptureAndDiscovery/somef.git

We use Poetry to ensure library compatibility. It can be installed as follows:

curl -sSL https://install.python-poetry.org | python3 -

This option is recommended over installing Poetry with pip install.

Now Poetry will handle the installation of SOMEF and all its dependencies configured in the toml file.

To test the correct installation of poetry run (poetry version > 2.0.0):

poetry --version

Install somef and all their dependencies.

cd /somef
poetry install

Now we need to access our virtual environment, to do so you can run the following command:

poetry env activate

If the environment is not active, paste the command shown when poetry env activate is run, typically something like the command below:

source /path_to_env/ENV_NAME/bin/activate

To learn more about poetry environment management, visit their official documentation here.

Test the SOMEF installation run:

somef --help

If everything goes fine, you should see:

Usage: somef [OPTIONS] COMMAND [ARGS]...

Options:
  -h, --help  Show this message and exit.

Commands:
  configure  Configure credentials
  describe   Running the Command Line Interface
  version    Show somef version.

Installing through Docker

We provide a Docker image with SOMEF already installed. To run through Docker, you may build the Dockerfile provided in the repository by running:

docker build -t somef .

Or just use the Docker image already built in DockerHub:

docker pull kcapd/somef

Then, to run your image just type:

docker run --rm -it kcapd/somef 

And you will be ready to use SOMEF (see section below). If you want to have access to the results we recommend mounting a volume. For example, the following command will mount the current directory as the out folder in the Docker image:

docker run -it --rm -v $PWD/:/out kcapd/somef 

If you move any files produced by somef into /out, then you will be able to see them in your current directory.

Configure

Before running SOMEF for the first time, you must configure it appropriately (you only need to do this once). Run:

somef configure

And you will be asked to provide the following:

• A GitHub authentication token [optional, leave blank if not used], which SOMEF uses to retrieve metadata from GitHub. If you don't include an authentication token, you can still use SOMEF. However, you may be limited to a series of requests per hour. For more information, see https://help.github.com/en/github/authenticating-to-github/creating-a-personal-access-token-for-the-command-line

• A GitLab authentication token [optional], used for GitLab.com and self-hosted GitLab instances (e.g., gitlab.in2p3.fr). Tokens are per-instance. Note: a token from GitLab.com does not work for self-hosted servers. Create one at https://gitlab.com/-/user_settings/personal_access_tokens (scope: read_api). Without a token, some self-hosted GitLab instances may not return rate limit information.

• A Codeberg authentication token [optional], used to retrieve metadata from Codeberg. Create one at https://codeberg.org/user/settings/applications (permissions: read:repository, read:user). Codeberg (Forgejo) does not expose rate limit headers even with a token.

• A Bitbucket authentication token [optional], used for Bitbucket Cloud. Create an API token with scopes at https://bitbucket.org/account/settings/api-tokens/ (permissions: read:repository:bitbucket, read:account). You will also need to provide your Atlassian account email, as Bitbucket API tokens use Basic authentication (email:token encoded in base64). Without a token you are limited to 60 requests/hour.

• The path to the trained classifiers (pickle files). If you have your own classifiers, you can provide them here. Otherwise, you can leave it blank.

• A download size limit in MB [optional, default 200]. SOMEF skips repository archives larger than this limit. Increase it if you need to process large repositories. You can also override it with the --download-limit parameter in the describe command.

If you want SOMEF to be automatically configured (without any tokens and using the default classifiers) just type:

somef configure -a

For showing help about the available options, run:

somef configure --help

Which displays:

Usage: somef configure [OPTIONS]

  Configure GitHub credentials and classifiers file path

Options:
  -a, --auto  Automatically configure SOMEF
  -b, --base_uri URL  Base URI for somef transformations
  -h, --help  Show this message and exit.

Commands:
  test  Test the configured API tokens

Testing your tokens

To verify that the authentication tokens stored in your configuration file are valid without having to run SOMEF, run:

somef configure test

This contacts each configured provider's API (api.github.com, gitlab.com, codeberg.org, bitbucket.org) with the token stored in ~/.somef/config.json and reports the status of each one, e.g.:

GitHub: token valid
GitLab: token valid
Codeberg: token valid
Bitbucket: token valid but with limited permissions (403)

401 means the token is invalid; 403 means it is valid but lacks the required scopes; any other non-200 response is reported as unexpected. Providers without a configured token are simply skipped. Bitbucket tokens must start with Basic (as set by somef configure); otherwise the test reports an incorrect format without contacting the API. The command exits with a non-zero status code if any configured token is invalid, which is useful for scripts and CI.

Updating SOMEF

If you update SOMEF to a newer version, you must configure again the library (by running somef configure). The rationale is that different versions may rely on classifiers which may be stored in a different path.

If you installed through poetry and you have upgraded the python environment (e.g., from 3.10 to 3.11), you may need to run poetry env use python3.11 and poetry install to update your environment.

Usage

$ somef describe --help
  SOMEF Command Line Interface
Usage: somef describe [OPTIONS]

  Running the Command Line Interface

Options:
  -t, --threshold FLOAT           Threshold to classify the text  [required]
  Input: [mutually_exclusive, required]
    -r, --repo_url URL            Github/Gitlab/Codeberg/Bitbucket Repository URL
    -d, --doc_src PATH            Path to the README file source
    -i, --in_file PATH            A file of newline separated links to GitHub/
                                  Gitlab/Codeberg/Bitbucket repositories
    -l, --local_repo PATH         Path to the local repository source. No APIs will be used

  Output: [required_any]
    -o, --output PATH             Path to the output file. If supplied, the
                                  output will be in JSON
    -c, --codemeta_out PATH       Path to an output codemeta file
    -g, --graph_out PATH          Path to the output Knowledge Graph export
                                  file. If supplied, the output will be a
                                  Knowledge Graph, in the format given in the
                                  --format option chosen (turtle, json-ld)
    -gc, --google_codemeta_out PATH Path to a Google-compliant Codemeta JSON-LD
                                    file. This output transforms the standard
                                    Codemeta to follow Google’s expected JSON-LD
                                    structure.
                                    
  -f, --graph_format [turtle|json-ld]
                                  If the --graph_out option is given, this is
                                  the format that the graph will be stored in

  -p, --pretty                    Pretty print the JSON output file so that it
                                  is easy to compare to another JSON output
                                  file.

  -m, --missing                   The JSON will include a field
                                  somef_missing_categories to report with the
                                  missing metadata fields that SOMEF was not
                                  able to find.

  -kt, --keep_tmp PATH            SOMEF will NOT delete the temporary folder
                                  where files are stored for analysis. Files
                                  will be stored at the
                                  desired path

  -all, --requirements_all        Export all detected requirements, including
                                  text and libraries (default).

  -v, --requirements_v            Export only requirements from structured
                                  sources (pom.xml, requirements.txt, etc.)


  -ra, --reconcile_authors         SOMEF will extract additional information 
                                  from certain files like CODEOWNERS. 
                                  This may require extra API
                                  requests and increase execution time
  --download-limit INTEGER        Download size limit in MB for repository
                                  archives. Overrides the value set in the
                                  configuration file.

  -h, --help                      Show this message and exit.

  -e, --enrichment                Enrich metadata with external APIs (OpenAlex, 
                                  OpenAIRE, Zenodo)

  --github-token TEXT             GitHub personal access token (if invalid,
                                  stored config is used instead)

  --gitlab-token TEXT             GitLab personal access token (if invalid,
                                  stored config is used instead)

  --codeberg-token TEXT           Codeberg personal access token (if invalid,
                                  stored config is used instead)

  --bitbucket-token TEXT          Bitbucket app password (if invalid, stored 
                                  config is used instead)

  --bitbucket-email TEXT          Bitbucket Atlassian account email (required
                                  with --bitbucket-token)
              

  
  Repoository versions [mutually_exclusive] (see section *Repository versions*t):
  -b, --branch name branch        Branch of the repository to analyze. Overrides the default branch.

  --tag text                      Tag of the repository to analyze. Cannot be used together with --branch and --commit

  --commit  TEXT                  Commit SHA to analyze. Cannot be used together 
                                  with --branch or --tag.

Alternatively, you can set tokens via environment variables or by running somef configure, which stores them permanently. The CLI flags take precedence over stored config when valid.

Enrichment with -e

The -e (or --enrichment) flag queries external APIs to complete the extracted metadata:

• OpenAlex: Adds openalex_id to DOIs of publications and reconciles missing author ORCIDs.
• OpenAIRE: Adds openaire_id to publications/identifiers and enriches project funding metadata.
• Zenodo: Adds swhid (Software Heritage ID) for records matching Zenodo DOIs.

For a detailed technical breakdown of the fields mapped by each external service, please refer to the specific documentation pages:

• See the OpenAlex Mapping Guide for citation and author properties.
• See the OpenAIRE and Zenodo Mapping Guide for funding and identifier properties.

Note: Enrichment makes additional network requests to external services, which may slow down the overall execution time. Use this flag only when you need the extra metadata.

Usage example:

The following command extracts all metadata available from https://github.com/dgarijo/Widoco/.

somef describe -r https://github.com/dgarijo/Widoco/ -o test.json -t 0.8

We recommend having a high value for the threshold parameter, 0.8 (default) or above. Additional configuration parameters (such as the similarity_threshold for header analysis) can be set in ~/.somef/config.json. See the usage documentation for details.

Try SOMEF in Binder with our sample notebook:

Contribute:

If you want to contribute with a pull request, please do so by submitting it to the dev branch.

Contributors:

Priyanka O.

Next features:

To see upcoming features, please have a look at our open issues and milestones

Extending SOMEF categories:

To run a classifier with an additional category or remove an existing one, a corresponding path entry in the config.json should be provided and the category type should be added/removed in the category variable in cli.py.

Metadata Support

SOMEF supports the extraction and analysis of metadata in package files of several programming languages. Current support includes: setup.py and pyproject.toml for Python, pom.xml for Java, .gemspec for Ruby, DESCRIPTION for R, bower.json for JavaScript, HTML or CSS, .cabal for Haskell, cargo.toml for RUST, composer for PHP, .juliaProject.toml for Julia , AUTHORS, codemeta.json, publiccode.yml, dockerfile and citation.cff This includes identifying dependencies, runtime requirements, and development tools specified in project configuration files.

Limitations

SOMEF is designed to work primarily with repositories written in English.
Repositories in other languages may not be processed as effectively, and results could be incomplete or less accurate.

Enrichment with -e

The -e (or --enrichment) flag queries external APIs to complete the extracted metadata:

• OpenAlex: adds openalex_id to DOIs of publications.
• OpenAIRE: adds openaire_id and enriches funding information (project code, title, acronym, grant id).
• Zenodo: adds swhid (Software Heritage ID) for Zenodo DOIs.

Note: Enrichment makes additional network requests to external services, which may slow down the overall execution time. Use this flag only when you need the extra metadata.

Repository versions: default behavior, branch tag and commit

SOMEF allows analyzing specific versions of a repository. If no version is specified, SOMEF will analyze the default branch of the repository (usually main or master). The following options let you control exactly which version of the codebase is inspected.

Default behavior

If neither --branch nor --tag is provided, SOMEF will:

• Clone the repository.
• Detect and analyze the default branch set up in github.

This is the recommended option when you want to describe the current version of a a project.

Using a branch

somef describe -r <repo_url> --branch <branch_name> ...
somef describe -r <repo_url> --b <branch_name> ...

Forces SOMEF to analyze a specific branch of the repository.

Useful when:

• The project maintains development, release or feature branches.
• You need to compare metadata across branches.
• You want to reproduce an analysis on a branch that is not the default one.

Using a tag

somef describe -r <repo_url> --tag <tag_name> ...

Analyzes a specific tagged version of the repository.

Recommended when:

• You need reproducible results (tags do not change over time).
• You want to document a released version of the software.
• You integrate SOMEF into pipelines that operate on versioned artifacts.

Using a commit

somef describe -r <repo_url> --commit <commit_sha> ...

Restrictions

• --branch, --tag and --commit are mutually exclusive.
• If either option is provided, it overrides the default branch behavior.