Update run-ngs-pipeline

* Update :command:`run-ngs-pipeline` command to check input files more vigorously.
NTNguyen13 · Nov 27, 2021 · ff6a5bf · ff6a5bf
1 parent 658d17f
commit ff6a5bf
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 8 deletions.
diff --git a/CHANGELOG.rst b/CHANGELOG.rst
@@ -13,6 +13,7 @@ Changelog
 * Rename :meth:`sdk.utils.Archive.check` method to :meth:`sdk.utils.Archive.check_type`.
 * Add new method :meth:`sdk.utils.Archive.check_metadata`.
 * Add new error ``sdk.utils.IncorrectMetadataError``.
+* Update :command:`run-ngs-pipeline` command to check input files more vigorously.
 
 0.8.0 (2021-11-20)
 ------------------

diff --git a/pypgx/api/pipeline.py b/pypgx/api/pipeline.py
@@ -156,6 +156,8 @@ def run_ngs_pipeline(
             depth_of_coverage = sdk.Archive.from_file(depth_of_coverage)
 
         depth_of_coverage.check_type('CovFrame[DepthOfCoverage]')
+        depth_of_coverage.check_metadata('Platform', platform)
+        depth_of_coverage.check_metadata('Assembly', assembly)
 
         if control_statistics is None:
             raise ValueError('SV detection requires SampleTable[Statistcs]')
@@ -164,9 +166,8 @@ def run_ngs_pipeline(
             control_statistics = sdk.Archive.from_file(control_statistics)
 
         control_statistics.check_type('SampleTable[Statistics]')
-
-        if depth_of_coverage.metadata['Platform'] != control_statistics.metadata['Platform']:
-            raise ValueError('Different platforms detected')
+        control_statistics.check_metadata('Platform', platform)
+        control_statistics.check_metadata('Assembly', assembly)
 
         read_depth = utils.import_read_depth(gene, depth_of_coverage)
         read_depth.to_file(f'{output}/read-depth.zip')

diff --git a/pypgx/sdk/utils.py b/pypgx/sdk/utils.py
@@ -8,14 +8,14 @@
 import pandas as pd
 from fuc import pyvcf, pycov, common, pybam
 
-class SemanticTypeNotFoundError(Exception):
-    """Raised when specified semantic type is not supported."""
+class IncorrectMetadataError(Exception):
+    """Raised when specified metadata is incorrect."""
 
 class IncorrectSemanticTypeError(Exception):
     """Raised when specified semantic type is incorrect."""
 
-class IncorrectMetadataError(Exception):
-    """Raised when specified metadata is incorrect."""
+class SemanticTypeNotFoundError(Exception):
+    """Raised when specified semantic type is not supported."""
 
 class Archive:
     """
@@ -122,10 +122,12 @@ def check_metadata(self, key, value):
         Raise IncorrectMetadataError if the archive does not have specified
         pair of key and value.
         """
+        semantic_type = self.metadata['SemanticType']
         actual_value = self.metadata[key]
         if actual_value != value:
             raise IncorrectMetadataError(
-                f"Expected '{key}={value}' but found '{key}={actual_value}'")
+                f"Expected '{key}={value}' but found '{key}={actual_value}' "
+                f"for semantic type '{semantic_type}'")
 
 def zipdir(dir, output):
     """