Parse MLModel file from MLFlow model

Author: smlindauer

src/sasctl/pzmm/mlflowModel.py

@@ -0,0 +1,39 @@
+# Copyright (c) 2020, SAS Institute Inc., Cary, NC, USA.  All Rights Reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+from pathlib import Path
+import json
+
+class mlflowModel:
+    def readMLmodelFile(self, mPath=Path.cwd()):
+        with open(mPath, 'r') as mFile:
+            mLines = mFile.readlines()
+        
+        # More verbose substring acceptance is needed for each possible model type
+        # For now, stick with those models which are based on pickle files and don't report model type or version
+        #ind = self.findSubstringIndex(mLines, 'loader_module')
+        #package = mLines[ind[0]].strip().split(' ')[1].split('.')[1]
+        
+        #ind = self.findSubstringIndex(mLines, package + '_version')
+        #packageVersion = mLines[ind[0]].strip().split(' ')[1]
+        
+        varList = ['python_version', 'serialization_format', 'run_id', 'model_path']
+        for i, varString in enumerate(varList):
+            index = [i for i, s in enumerate(mLines) if varString in s]
+            if not index:
+                raise ValueError('This MLFlow model type is not currently supported.')
+            varList[i] = {varList[i]: mLines[index[0]].strip().split(' ')[1]}
+            
+        varDict = {k: v for d in varList for k, v in d.items}
+        varDict['mlflowPath'] = mPath
+        
+        indIn = [i for i, s in enumerate(mLines) if 'inputs:' in s]
+        indOut = [i for i, s in enumerate(mLines) if 'outputs:' in s]
+        
+        inputs = mLines[indIn[0]:indOut[0]]
+        outputs = mLines[indOut[0]:-1]
+        
+        inputsDict = json.loads(''.join([s.strip() for s in inputs])[9:-1])
+        outputsDict = json.loads(''.join([s.strip() for s in outputs])[10:-1])
+        
+        return varDict, inputsDict, outputsDict #,package, packageVersion