Merge pull request #47 from nextstrain/test-data

feat(ci): Use cached ncbi dataset package to isolate tests

ingest/build-configs/ci/config.yaml

@@ -1,5 +1,5 @@
-# TODO: If the ingest workflow ever runs too long, we should figure out a way
-# to subset the ingest data. Currently, the CI just runs the default ingest workflow.
+custom_rules:
+  - "build-configs/ci/copy_example_data.smk"
 
 # Snakemake requires at least one top level key in a config file, so including
 # a bogus key here that should not be used anywhere in the Snakemake workflow

ingest/build-configs/ci/copy_example_data.smk

@@ -0,0 +1,14 @@
+rule copy_example_data:
+    input:
+        ncbi_dataset="example_data/ncbi_dataset.zip"
+    output:
+        ncbi_dataset=temp("data/ncbi_dataset.zip")
+    shell:
+        """
+        cp -f {input.ncbi_dataset} {output.ncbi_dataset}
+        """
+
+# Add a Snakemake ruleorder directive here if you need to resolve ambiguous rules
+# that have the same output as the copy_example_data rule.
+
+ruleorder: copy_example_data > fetch_ncbi_dataset_package

ingest/rules/fetch_from_ncbi.smk

@@ -107,4 +107,3 @@ rule format_ncbi_datasets_ndjson:
             --duplicate-reporting warn \
             2> {log} > {output.ndjson}
         """
-