Redesign the function recordpubli.to_intial.

modules/invenio_tools/recordpubli.py

@@ -30,33 +30,37 @@ MSG_INVALID_FMT = "Invalid format for author"
 # the keys containing paper reference
 PAPER_REFERENCE_KEYS = {"pagination", "title", "volume", "year"}
 
-# extract initial of a first name
-REG_INITIAL = initial = r"^(\w+)\.?(\-)* *(\w+)*\.?$"
 
+def to_initial(name):
+    """Help function to extract initial from a first name:
 
-def to_initial(x, y, z):
-    """Help function to extract initial from a first name split in x, y and z:
-
-        Albert                (x="Albert", y="", z="")
-        Antonio Augusto       (x="Antonio", y="", z="Augusto")
-        Jean-Pierre           (x="Jean", y="-", z="Pierre")
+            +------------------+----------+
+            | name             | initial  |
+            +------------------+----------+
+            | Albert           | A.       |
+            | Antonio Augusto  | A. A.    |
+            | Kristof Antoon M | K. A. M. |
+            | Jean-Pierre      | J.-P.    |
+            | Marie-Hélène     | M.-H.    |
+            +------------------+----------+
 
     Args:
-        x (str): first part
-        y (str): separator
-        z (str): second part
+        name (unicode):
 
     Returns:
-        str
+        unicode
 
     """
-    if z == "":
-        return "%s." % x[0:1]
+    if len(name) == 0:
+        return u""
 
-    if y == "":
-        return "%s. %s." % (x[0:1], z[0:1])
-    else:
-        return "%s.%s%s." % (x[0:1], y[0:1], z[0:1])
+    li = []
+    for el in re.finditer(r"(\w+|-)", name, re.UNICODE):
+        val = el.group(1)[0:1]
+        val = (val if val == u"-" else "%s." % val)
+        li.append(val)
+
+    return (u"".join(li) if "-" in li else u" ".join(li))
 
 
 def to_str(x):
@@ -710,11 +714,9 @@ class RecordPubli(Record):
         #
         if fmt in ("F. Last", "Last F."):
 
-            dfm = (df.first_name.str.extract(REG_INITIAL, expand=True)
-                   .fillna(""))
-
-            df["initial"] = dfm.apply(
-                lambda x: to_initial(x[0], x[1], x[2]), axis="columns")
+            df["initial"] = (df.first_name
+                             .fillna("")
+                             .apply(to_initial))
 
         # ....................................................................
         #

tests/basis/test_04_RecordPubli.py

@@ -26,6 +26,7 @@ import pandas as pd
 import pytest
 
 from invenio_tools import load_record
+from invenio_tools.recordpubli import to_initial
 
 
 @pytest.fixture(scope="module")
@@ -67,6 +68,15 @@ def test_constructor(reccds):
     assert paper.pagination == "P12005"
 
 
+def test_to_initial():
+
+    assert to_initial(u"Albert") == u"A."
+    assert to_initial(u"Antonio Augusto") == u"A. A."
+    assert to_initial(u"Kristof Antoon M") == u"K. A. M."
+    assert to_initial(u"Jean-Pierre") == u"J.-P."
+    assert to_initial(u"Marie-Hélène") == u"M.-H."
+    assert to_initial(u"Marie - Pierre") == u"M.-P."
+
 # ............................................................................
 #
 #    Section devoted to authors

tests/issues/test_acti.py

@@ -38,6 +38,21 @@ def test_acti_cds1411352(svc):
     assert reccds.submitted() == "2012-01-05"
 
 
+def test_acti_cds2270937(svc):
+    """
+    JSON:
+        * Author name not well formatted when using F. Last.
+        * More general design of the function recordpubli.to_initial
+
+    """
+    reccds = load_record('cds.cern.ch', 2270937)
+    assert reccds.first_author() == "De Bruyn, Kristof Antoon M"
+
+    svc.format_authors(reccds, fmt="F. Last")
+
+    assert reccds.first_author() == "K. A. M. De Bruyn"
+
+
 def test_acti_ins1276938(svc):
     """
     XML: