codemeta-harvester

• added tagignore parameter and set default value to ignore alpha, beta, rc suffixes
• detect Cargo.toml (Rust)
• assign the top contributor as fallback author/maintainer in case none of the metadata specifies any
• implemented 'stub' detection, repository that only have metadata and no code, will negatively impact validation
• expanded README detection and README URL inferrence, now also works for bitbucket and for subdirs
• remove newlines from scraped titles
• metadata update

• Container image uses new codemetapy 2.5.1

• made harvester more robust against failures in the pipeline
• add README URL also for codemeta.json if missing

• Keep prefix when checking out tag (thanks to @ddeboer)

Release for use with codemetapy 2.4.1, no functional changes.

• own codemeta metadata update
• adding some support for inferring developmentStatus from stability badges (https://masterminds.github.io/stability/)
• no need for the technologyReadinessLevel property anymore, use developmentStatus directly
• Makefile: build containers with no-cache
• fixed TRL URI (CLARIAH/tool-discovery#34)
• Dockerfile: no longer need to grab pandoc from alpine edge, it's in stable now
• Implemented "tagprefix" option to detect git version tags with a custom prefix.

• support for setting group in configuration, adding a means to aggregate tools into groups #10

This releases builds on codemetapy 2.3.0.

• Logs are more verbose
• Implemented DOI harvesting via Zenodo API #3
• Implemented support for Technology Readiness Level detection
• Added support for alternative spellings of README (thanks to @broeder-j) #4
• added fallbacks to self-correct in case git checkout fails, and some other improvements in error catching and overall robustness
• fix for 'Inferring repostatus information from git activity', wasn't always run
• documentation update

• Attempt to infer repostatus from git activty
• Fixed running without configuration mode, infer project in current working dir (e.g. no positional arguments)

• Dockerfile: allow building against any development version of codemetapy
• Allow validation against a SHACL schema (https://github.com/proycon/codemetapy/issues/21)
• Implemented support for Gitlab API harvesting (#1, thanks to @xmichele)
  • don't query api for unsupported git forges
• Limit 999kb for downloaded files from git clone (thanks to @xmichele)
• Improved overall error feedback, logging, and verbosity
• enabled new codemetapy --enrich option for some automatic enrichment
• derived some properties we from the very latest master branch instead of a specific version and backport to older versions (such as maintainer and repostatus) because these are more repo-bound than version bound. (we do this even if an explicit codemeta.json was provided!)
• allow single letter suffixes in version numbers (as used by @dirkroorda in text-fabric)
• improved getting latest version (git tag): let git itself handle the sorting of tags based on semantic version
• implemented CI detection from README
• implemented support for harvesting services without source (limited)

First release, should be fairly functional all-around but does still require some further testing.