Merge pull request #41 from jameskochubasas/master

python models with string modifications

Author: Jon Walker

src/sasctl/utils/pymas/core.py

@@ -58,11 +58,25 @@ def build_wrapper_function(func, variables, array_input, setup=None,
     args = input_names
     func = func.__name__ if callable(func) else func
 
+    # HELPER: SAS to python char issue where SAS char have spaces and python string does not.
+    #   NOTE: we assume SAS char always need white space to be trimmed.  This seems to match python model built so far
+    pythonStringInput = ('',)
+    for tmp1 in variables:
+        if not tmp1.out:
+            if tmp1.type == 'char':
+                pythonStringInput = pythonStringInput + ("        if " + tmp1.name + ": " + tmp1.name + " = " + tmp1.name + ".strip()",)
+
     # Statement to execute the function w/ provided parameters
     if array_input:
-        func_call = '{}(np.array([{}]).reshape((1, -1)))'.format(func, ','.join(args))
+        middle = pythonStringInput +\
+                 ('        inputarray = np.array([{}]).reshape((1, -1))'.format(','.join(args)),
+                  '        column=[{}]'.format(','.join('"{0}"'.format(w) for w in args)),
+                  '        import pandas as pd',
+                  '        inputrun=pd.DataFrame(data=inputarray, columns=column)',
+                  '        result = {}(inputrun)'.format(func))
     else:
         func_call = '{}({})'.format(func, ','.join(args))
+        middle = ('        result = {}'.format(func_call),)
 
     # TODO: Verify that # of values returned by wrapped func matches length of output_names
     # TODO: cast all return types before returning (DS2 errors out if not exact match)
@@ -80,6 +94,7 @@ def build_wrapper_function(func, variables, array_input, setup=None,
     else:
         header = ('', )
 
+
     definition = header +\
                  ('def wrapper({}):'.format(', '.join(args)),
                   '    "Output: {}"'.format(', '.join(output_names + ['msg']) if return_msg
@@ -90,9 +105,9 @@ def build_wrapper_function(func, variables, array_input, setup=None,
                   '        if _compile_error is not None:',
                   '            raise _compile_error',
                   '        msg = ""' if return_msg else '',
-                  '        import numpy as np',
-                  '        result = {}'.format(func_call),
-                  '        if result.size == 1:',
+                  '        import numpy as np') +\
+                 middle +\
+                 ('        if result.size == 1:',
                   '            result = np.asscalar(result)',
                   '    except Exception as e:',
                   '        msg = str(e)' if return_msg else '',
@@ -103,6 +118,7 @@ def build_wrapper_function(func, variables, array_input, setup=None,
                   '    else: ',
                   '        return result, msg')
 
+
     return '\n'.join(definition)
 
 

src/sasctl/utils/pymas/ds2.py

@@ -18,7 +18,7 @@ def __init__(self, variables, code=None, return_code=True,
         self._python_code = code or []
         code = code or []
 
-        self.methods = [DS2PyMASMethod(variables, code, return_code,
+        self.methods = [DS2PyMASMethod(self._id, variables, code, return_code,
                                        return_message, target)]
 
         self._body = ("dcl package pymas py;",
@@ -88,7 +88,7 @@ def code(self):
 
 
 class DS2PyMASMethod(DS2BaseMethod):
-    def __init__(self, variables, python_code, return_code=True,
+    def __init__(self, name, variables, python_code, return_code=True,
                  return_message=True, target='wrapper'):
 
         target = target or 'wrapper'
@@ -110,12 +110,12 @@ def __init__(self, variables, python_code, return_code=True,
                 self.private_variables] \
                + ["if null(py) then do;",
                   "    py = _new_ pymas();",
-                  "    rc = py.useModule('mypymodule', 1);",
+                  "    rc = py.useModule('%s', 1);" % name,
                   "    if rc then do;"] \
                + ["        rc = py.appendSrcLine('%s');" % l for l in
                   python_code] \
                + ["        pycode = py.getSource();",
-                  "        revision = py.publish(pycode, 'mypymodule');",
+                  "        revision = py.publish(pycode, '%s');" % name,
                   "        if revision lt 1 then do;",
                   "            logr.log('e', 'py.publish() failed.');",
                   "            rc = -1;",

tests/integration/test_pymas.py

@@ -165,7 +165,7 @@ def test_from_pickle(train_data, pickle_file):
     
         if null(py) then do;
             py = _new_ pymas();
-            rc = py.useModule('mypymodule', 1);
+            rc = py.useModule('DF74A4B18C9E41A2A34B0053E123AA67', 1);
             if rc then do;
                 rc = py.appendSrcLine('try:');
                 rc = py.appendSrcLine('    import pickle, base64');
@@ -184,7 +184,12 @@ def test_from_pickle(train_data, pickle_file):
                 rc = py.appendSrcLine('            raise _compile_error');
                 rc = py.appendSrcLine('        msg = ""');
                 rc = py.appendSrcLine('        import numpy as np');
-                rc = py.appendSrcLine('        result = obj.predict(np.array([SepalLength,SepalWidth,PetalLength,PetalWidth]).reshape((1, -1)))');
+                rc = py.appendSrcLine('');
+                rc = py.appendSrcLine('        inputarray = np.array([SepalLength,SepalWidth,PetalLength,PetalWidth]).reshape((1, -1))');
+                rc = py.appendSrcLine('        column=["SepalLength","SepalWidth","PetalLength","PetalWidth"]');
+                rc = py.appendSrcLine('        import pandas as pd');
+                rc = py.appendSrcLine('        inputrun=pd.DataFrame(data=inputarray, columns=column)');
+                rc = py.appendSrcLine('        result = obj.predict(inputrun)');
                 rc = py.appendSrcLine('        if result.size == 1:');
                 rc = py.appendSrcLine('            result = np.asscalar(result)');
                 rc = py.appendSrcLine('    except Exception as e:');
@@ -196,7 +201,7 @@ def test_from_pickle(train_data, pickle_file):
                 rc = py.appendSrcLine('    else: ');
                 rc = py.appendSrcLine('        return result, msg');
                 pycode = py.getSource();
-                revision = py.publish(pycode, 'mypymodule');
+                revision = py.publish(pycode, 'DF74A4B18C9E41A2A34B0053E123AA67');
                 if revision lt 1 then do;
                     logr.log('e', 'py.publish() failed.');
                     rc = -1;

tests/unit/test_pymas.py

@@ -357,7 +357,7 @@ def domath(a, b):
     d = py.getDouble('d');
 end;
 """
-    method = DS2PyMASMethod([
+    method = DS2PyMASMethod('mypymodule',[
         DS2Variable('a', 'double', False),
         DS2Variable('b', 'double', False),
         DS2Variable('c', 'double', True),
@@ -452,15 +452,15 @@ def test_ds2_package():
     
         if null(py) then do;
             py = _new_ pymas();
-            rc = py.useModule('mypymodule', 1);
+            rc = py.useModule('pyscore', 1);
             if rc then do;
                 rc = py.appendSrcLine('def domath(a, b):');
                 rc = py.appendSrcLine('    "Output: c, d"');
                 rc = py.appendSrcLine('    c = a * b');
                 rc = py.appendSrcLine('    d = a / b');
                 rc = py.appendSrcLine('    return c, d');
                 pycode = py.getSource();
-                revision = py.publish(pycode, 'mypymodule');
+                revision = py.publish(pycode, 'pyscore');
                 if revision lt 1 then do;
                     logr.log('e', 'py.publish() failed.');
                     rc = -1;