dnanexus / wdlTools

WDL tools for parsing, type-checking, and more
Apache License 2.0
25 stars 7 forks source link

Add corpora of public WDLs for testing #159

Closed jdidion closed 3 years ago

jdidion commented 3 years ago

Added src/test/resources/corpora_repos.json, which describes corpora of public WDLs that we use for testing. There is an entry for each corpus (GitHub repository), and within each corpus are any number of entrypoints, each of which may be an individual WDL or a directory with multiple WDLs. In the latter case, some WDLs may be specifically included or excluded. The WDLs in an entrypoint may need to be fixed using the wdlTools fix command ("fix": true). The WDLs in an entrypoint may be expected to fail parsing or type-checking ("fail": true); if so, a reason is given.

The corpora WDLs are not committed to GitHub. Instead, the clone-corpora.sh script needs to be run first. The corpora tests are skipped if the src/test/resources/corpora directory does not exist.

This PR also fixes a number of issues uncovered by the test cases: