Add workflows that are not directly using pathogen-repo-build

[joverlee521]

Context

Prompted by nextstrain/zika#55 (comment)

We want to monitor GH Action workflows using the pathogen-repo-build reusable workflow and we are currently filtering for them with

status/pathogen-workflows.sql

Lines 87 to 88 in d744ff7

	where
	workflow_file_content_json @@ '$.jobs.*.uses starts with "nextstrain/.github/.github/workflows/pathogen-repo-build.yaml@"'

However, this does not include workflows that are nesting reusable workflows, such as the current zika ingest-to-phylo.yaml workflow.

Possible solutions

1. We could add another proxy for filtering if we use a stable pattern across pathogen repos. e.g filter for path ingest-to-phylogenetic.yaml
2. We could traverse the nested workflows to check if any of them are using the pathogen-repo-build workflow. (GH Action currently allows four levels of nested reusable workflows)

[joverlee521]

Ah right, just remembered this issue since I changed avian flu GH Action workflows in nextstrain/avian-flu#71 but the name is ingest-to-phylogenetic-ncbi.yaml instead of ingest-to-phylogenetic.yaml.

I'm going to add a separate check for the avian-flu workflow, but other pathogens that only use NCBI data should continue to follow the generic ingest-to-phylogenetic.yaml naming.

[joverlee521]

Or maybe just update exact match to pattern match

diff --git a/pathogen-workflows.sql b/pathogen-workflows.sql
index d68d368..cf23d33 100644
--- a/pathogen-workflows.sql
+++ b/pathogen-workflows.sql
@@ -93,7 +93,7 @@ workflow as materialized (
                  *
                  * <https://github.com/nextstrain/status/issues/12>
                  */
-                or path = '.github/workflows/ingest-to-phylogenetic.yaml')
+                or path like '.github/workflows/ingest-to-phylogenetic%.yaml')
         and name != 'CI'
 ),

[genehack]

@joverlee521 it looks like, reviewing the above, that the work here is done and this issue can be closed?

[joverlee521]

Ah yeah, I'll consider #24 good enough. We can re-open this if there are other cases not supported.