Changeset 7fb0101 in git

Timestamp:

Aug 5, 2019, 11:00:30 PM (4 years ago)

Author:

Murray Heymann <heymann.murray@…>

Branches:

(u'spielwiese', '828514cf6e480e4bafc26df99217bf2a1ed1ef45')

Children:

bb8cdbbda3604b0b59579d1fa969c8d5ca30f07b

Parents:

2f9e3f171a90f561e7a3c81734eb72998dc8fe82

git-author:

Murray Heymann <heymann.murray@gmail.com>2019-08-05 23:00:30+02:00

git-committer:

Murray Heymann <heymann.murray@gmail.com>2019-08-05 23:00:37+02:00

Message:

Begin writing c functions to interact with python3

Location:

machine_learning

Files:

• .coveragerc (modified) (1 diff)
• ml_python/common/constants.py (modified) (1 diff)
• ml_python/common/lookuptable.py (modified) (3 diffs)
• ml_python/tests/common/test_keyword_vectors.py (modified) (2 diffs)
• ml_python/tests/common/test_lookuptable.py (modified) (2 diffs)
• mlpredict.c (added)
• mlpredict.h (added)
• testml.c (added)

machine_learning/.coveragerc

@@ -1 +1 @@
 [run]
 branch = True
-omit = tests/*, *__init__.py, predictor_runner.py
+omit = *tests/*, *__init__.py, *predictor_runner.py
 
 

machine_learning/ml_python/common/constants.py

@@ -7 +7 @@
 EXTRACT_SCRIPT = "extract.lib"
 KEYWORDS_FILE = "keywords.txt"
-VECTORS_NPY = "vectors.npy"
-HELPFILE_NPY = "helpfilelist.npy"
+VECTORS_NPY = ".vectors.npy"
+HELPFILE_NPY = ".helpfilelist.npy"
 
 HELP_FILE_URL = "ftp://jim.mathematik.uni-kl.de/pub/Math/Singular/src/4-1-2/doc.tbz2"

machine_learning/ml_python/common/lookuptable.py

@@ -55 +55 @@
     # read from the file created by singular
     dictionary = read_dictionary()
-    print(dictionary)
 
     # sort alphabetically
     dictionary = np.sort(np.unique(dictionary))
-    print(dictionary)
 
     # write back to the same file
@@ -66 +64 @@
             file.write(word + "\n")
 
+    return dictionary
 
 
@@ -97 +96 @@
 
     return (vectors, file_list)
+
+def init_table_on_system():
+    """
+    check whether the various files exist, and create if necessary. 
+    """
+    # check for and download help files if necessary
+    tbz2_path = os.path.join(HELP_FILE_PATH, "helpfiles.tbz2")
+    if not os.path.isdir(HELP_FILE_PATH) or not os.path.isfile(tbz2_path):
+        fetch_tbz2_data()
+
+    # Use Singular to extract the keywords and save in a file.
+    if not os.path.isfile(KEYWORDS_FILE):
+        dictionary = extract_keywords()
+    else:
+        dictionary = None
+
+
+    if not os.path.isfile(VECTORS_NPY) or not os.path.isfile(HELPFILE_NPY):
+        vectors, file_list = create_table(dictionary=dictionary,
+                                          attempt_cached=False)

machine_learning/ml_python/tests/common/test_keyword_vectors.py

@@ -73 +73 @@
         print("Testing create_vector_dictionary function:")
 
-        read_dictionary()
+        dictionary = read_dictionary()
 
         print("Create Vector Dictionary with None as dictionary:")
@@ -82 +82 @@
                           create_vector_dictionary,
                           np.array([]))
-        print()
+        vec_dic = create_vector_dictionary(dictionary)
 
     def test_vector_distance(self):

machine_learning/ml_python/tests/common/test_lookuptable.py

@@ -6 +6 @@
 from common.lookuptable import *
 from common.keyword_vector import count_occurances
-from common.constants import KEYWORDS_FILE
+from common.constants import KEYWORDS_FILE, VECTORS_NPY, HELP_FILE_PATH
 
 class TestLookuptableMethods(unittest.TestCase):
@@ -33 +33 @@
         self.assertTrue((test_vec == vectors[1]).all())
 
+    def test_init_table_on_system(self):
+        tbz2_path = os.path.join(HELP_FILE_PATH, "helpfiles.tbz2")
+        os.remove(tbz2_path)
+        os.remove(KEYWORDS_FILE)
+        os.remove(VECTORS_NPY)
+
+        init_table_on_system()
+
+        self.assertTrue(os.path.isfile(tbz2_path))
+        self.assertTrue(os.path.isfile(KEYWORDS_FILE))
+        self.assertTrue(os.path.isfile(VECTORS_NPY))
+
+        init_table_on_system()
+
+
 if __name__ == '__main__':
     #cProfile.run("unittest.main()")