From 40df65dec296b81650987c8ee4f832b703ab8f74 Mon Sep 17 00:00:00 2001
From: lltommy
Date: Tue, 7 Apr 2020 19:51:49 +0200
Subject: adding dummy metadata qc to the project

---
 bh20sequploader/qc_metadata.py | 13 +++++++++++++
 example/dummyschema.yaml       | 16 ++++++++++++++++
 example/metadata.json          |  0
 example/metadata.yaml          | 17 +++++++++++++++++
 4 files changed, 46 insertions(+)
 create mode 100644 bh20sequploader/qc_metadata.py
 create mode 100644 example/dummyschema.yaml
 delete mode 100644 example/metadata.json
 create mode 100644 example/metadata.yaml

diff --git a/bh20sequploader/qc_metadata.py b/bh20sequploader/qc_metadata.py
new file mode 100644
index 0000000..0632777
--- /dev/null
+++ b/bh20sequploader/qc_metadata.py
@@ -0,0 +1,13 @@
+import yamale
+
+## NOTE: this is just a DUMMY. Everything about this can and will change
+def qc_metadata(metadatafile):
+    print("Start metadata validation...")
+    schema = yamale.make_schema('../example/dummyschema.yaml')
+    data = yamale.make_data(metadatafile)
+    # Validate data against the schema. Throws a ValueError if data is invalid.
+    yamale.validate(schema, data)
+    print("...complete!")
+
+#qc_metadata("../example/metadata.yaml")
+
diff --git a/example/dummyschema.yaml b/example/dummyschema.yaml
new file mode 100644
index 0000000..e428324
--- /dev/null
+++ b/example/dummyschema.yaml
@@ -0,0 +1,16 @@
+#sampleInformation: include('sampleInformation')
+#InstituteInformation: include('InstituteInformation')
+---
+sampleInformation:
+  location : str()
+  host : str()
+  sequenceTechnology: str()
+  assemblyMethod: str()
+
+InstituteInformation:
+  OriginatingLab: str()
+  SubmittingLab: str()
+
+VirusDetail:
+  VirusName: str()
+  AccessionId: str()
diff --git a/example/metadata.json b/example/metadata.json
deleted file mode 100644
index e69de29..0000000
diff --git a/example/metadata.yaml b/example/metadata.yaml
new file mode 100644
index 0000000..587d0be
--- /dev/null
+++ b/example/metadata.yaml
@@ -0,0 +1,17 @@
+sampleInformation:
+  location: "USA"
+  host : "Homo Sapiens"
+  sequenceTechnology: "Sanger"
+  assemblyMethod: "CLC Genomics"
+
+InstituteInformation:
+  OriginatingLab: "Erik's kitchen"
+  SubmittingLab: "National Institute for Viral Disease Control and Prevention, China CDC"
+
+SubmitterInformation:
+  Submitter: "National Institute for Viral Disease Control and Prevention, China CDC"
+  submissionDate: "04-04-2020"
+
+VirusDetail:
+  VirusName: "hCoV-19/USA/identifer/2020"
+  AccessionId: "EPI_ISL_Random"
-- 
cgit v1.2.3