Merge pull request scikit-learn#4966 from lazywei/multilabel-dump-svm…

…light

ENH Add multilabel support to dump_svmlight_file

sklearn/datasets/svmlight_format.py

@@ -225,7 +225,7 @@ def load_svmlight_files(files, n_features=None, dtype=np.float64,
     dtype : numpy data type, default np.float64
         Data type of dataset to be loaded. This will be the data type of the
         output numpy arrays ``X`` and ``y``.
- 
+
     Returns
     -------
     [X1, y1, ..., Xn, yn]
@@ -275,18 +275,19 @@ def load_svmlight_files(files, n_features=None, dtype=np.float64,
     return result
 
 
-def _dump_svmlight(X, y, f, one_based, comment, query_id):
+def _dump_svmlight(X, y, f, multilabel, one_based, comment, query_id):
     is_sp = int(hasattr(X, "tocsr"))
     if X.dtype.kind == 'i':
         value_pattern = u("%d:%d")
     else:
         value_pattern = u("%d:%.16g")
 
     if y.dtype.kind == 'i':
-        line_pattern = u("%d")
+        label_pattern = u("%d")
     else:
-        line_pattern = u("%.16g")
+        label_pattern = u("%.16g")
 
+    line_pattern = u("%s")
     if query_id is not None:
         line_pattern += u(" qid:%d")
     line_pattern += u(" %s\n")
@@ -309,14 +310,22 @@ def _dump_svmlight(X, y, f, one_based, comment, query_id):
             row = zip(np.where(nz)[0], X[i, nz])
 
         s = " ".join(value_pattern % (j + one_based, x) for j, x in row)
+
+        if multilabel:
+            nz_labels = np.where(y[i] != 0)[0]
+            labels_str = ",".join(label_pattern % j for j in nz_labels)
+        else:
+            labels_str = label_pattern % y[i]
+
         if query_id is not None:
-            feat = (y[i], query_id[i], s)
+            feat = (labels_str, query_id[i], s)
         else:
-            feat = (y[i], s)
+            feat = (labels_str, s)
+
         f.write((line_pattern % feat).encode('ascii'))
 
 
-def dump_svmlight_file(X, y, f, zero_based=True, comment=None, query_id=None):
+def dump_svmlight_file(X, y, f, multilabel=False, zero_based=True, comment=None, query_id=None):
     """Dump the dataset in svmlight / libsvm file format.
 
     This format is a text-based format, with one sample per line. It does
@@ -339,6 +348,10 @@ def dump_svmlight_file(X, y, f, zero_based=True, comment=None, query_id=None):
         If file-like, data will be written to f. f should be opened in binary
         mode.
 
+    multilabel: boolean, optional
+        Samples may have several labels each (see
+        http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/multilabel.html)
+
     zero_based : boolean, optional
         Whether column indices should be written zero-based (True) or one-based
         (False).
@@ -368,7 +381,7 @@ def dump_svmlight_file(X, y, f, zero_based=True, comment=None, query_id=None):
             raise ValueError("comment string contains NUL byte")
 
     y = np.asarray(y)
-    if y.ndim != 1:
+    if y.ndim != 1 and not multilabel:
         raise ValueError("expected y of shape (n_samples,), got %r"
                          % (y.shape,))
 
@@ -396,7 +409,7 @@ def dump_svmlight_file(X, y, f, zero_based=True, comment=None, query_id=None):
     one_based = not zero_based
 
     if hasattr(f, "write"):
-        _dump_svmlight(X, y, f, one_based, comment, query_id)
+        _dump_svmlight(X, y, f, multilabel, one_based, comment, query_id)
     else:
         with open(f, "wb") as f:
-            _dump_svmlight(X, y, f, one_based, comment, query_id)
+            _dump_svmlight(X, y, f, multilabel, one_based, comment, query_id)

sklearn/datasets/tests/test_svmlight_format.py

@@ -249,6 +249,20 @@ def test_dump():
                 assert_array_equal(y, y2)
 
 
+def test_dump_multilabel():
+    X = [[1, 0, 3, 0, 5],
+         [0, 0, 0, 0, 0],
+         [0, 5, 0, 1, 0]]
+    y = [[0, 1, 0], [1, 0, 1], [1, 1, 0]]
+    f = BytesIO()
+    dump_svmlight_file(X, y, f, multilabel=True)
+    f.seek(0)
+    # make sure it dumps multilabel correctly
+    assert_equal(f.readline(), b("1 0:1 2:3 4:5\n"))
+    assert_equal(f.readline(), b("0,2 \n"))
+    assert_equal(f.readline(), b("0,1 1:5 3:1\n"))
+
+
 def test_dump_concise():
     one = 1
     two = 2.1