[gnue] r8581 - trunk/gnue-common/src/utils

[johannes]

Author: johannes
Date: 2006-08-21 08:58:28 -0500 (Mon, 21 Aug 2006)
New Revision: 8581

Modified:
   trunk/gnue-common/src/utils/ucsv.py
Log:
Added shorthand function for exporting to csv files; added error 
handling schema for encoding


Modified: trunk/gnue-common/src/utils/ucsv.py
===================================================================
--- trunk/gnue-common/src/utils/ucsv.py 2006-08-21 13:18:30 UTC (rev 8580)
+++ trunk/gnue-common/src/utils/ucsv.py 2006-08-21 13:58:28 UTC (rev 8581)
@@ -58,12 +58,13 @@
     # Constructor
     # -------------------------------------------------------------------------
 
-    def __init__(self, fileobject, fieldnames, encoding="utf-8", restval="",
-            extrasaction="raise", dialect="ExcelSemicolon", *args, **kwds):
+    def __init__(self, fileobject, fieldnames, dialect="ExcelSemicolon",
+            encoding="utf-8", encode_errors="replace"):
 
         self.encoding = encoding
-        csv.DictWriter.__init__(self, fileobject, fieldnames, restval,
-                extrasaction, dialect, *args, **kwds)
+        self.encode_errors = encode_errors
+        csv.DictWriter.__init__(self, fileobject, fieldnames, "", "ignore",
+                dialect)
 
 
     # -------------------------------------------------------------------------
@@ -79,17 +80,48 @@
         ndict = {}
         for (key, value) in rowdict.items():
             if isinstance(value, unicode):
-                value = value.encode(self.encoding)
+                value = value.encode(self.encoding, self.encode_errors)
             if isinstance(value, float):
-                decp = locale.localeconv().get('decimal_point', '.')
-                if (decp != '.'):
-                    value = ("%s" % value).replace('.', decp)
+                value = locale.str(value)
 
             ndict[key] = value
 
         return csv.DictWriter._dict_to_list(self, ndict)
 
+
 # =============================================================================
+# Shorthand for creating a CSV file from a dictionary
+# =============================================================================
+
+def write_file(fieldnames, data, filename, dialect="ExcelSemicolon",
+        encoding="utf-8", encode_errors="replace", caption=None):
+    """
+    Create a CSV file using the given fieldnames and data
+    @param fieldnames: list of fields to export
+    @param data: list of dictionaries with the data to export 
+    @param filename: name of the file to export the data to
+    @param dialect: name of a registered dialect (as used by the csv module)
+    @param encoding: an encoding used for unicode values
+    @param encode_errors: error handling schema for encoding
+    @param caption: dictionary giving a nice caption per column.  If no such
+        mapping is provided the fieldnames will be used as header line
+    """
+    fhd = open(filename, 'wb')
+    try:
+        writer = UDictWriter(fhd, fieldnames, dialect, encoding, encode_errors)
+
+        if not caption:
+            caption = {}
+            for field in fieldnames:
+                caption[field] = field
+        writer.writerow(caption)
+        writer.writerows(data)
+
+    finally:
+        fhd.close()
+
+
+# =============================================================================
 # Selftest
 # =============================================================================
 
@@ -97,14 +129,12 @@
 
     import datetime
 
-    data = [{'foo': u'Bl\xf6di', 'bar': None, 'sepp': 7.41, 'noo': 100},
-            {'foo': u'Bl\xf6di und so', 'bar': None, 'sepp':
+    xk = u'bar\xf6xy'
+    test = [{'foo': u'Bl\xf6di', xk: None, 'sepp': 7.41, 'noo': 100},
+            {'foo': u'Bl\xf6di und so', xk: None, 'sepp':
                 datetime.date.today(), 'noo': 100},
-            {'foo': u'Bl\xf6di, und;so', 'bar': None, 'sepp':
+            {'foo': u'Bl\xf6di, und;so', xk: None, 'sepp':
                 datetime.date.today(), 'noo': 100},
             ]
-    out = open('out.csv', 'w')
 
-    xcs = UDictWriter(out, ['foo', 'sepp', 'bar', 'noo'], 'cp1252')
-    xcs.writerows(data)
-    out.close()
+    write_file(test[0].keys(), test, 'out.csv', encoding="cp1250")