From e75a00541d83c1df6ede20fec3a5611f960b6821 Mon Sep 17 00:00:00 2001
From: Luca Beltrame <einar@heavensinferno.net>
Date: Tue, 19 Oct 2010 10:45:54 +0200
Subject: [PATCH] Support for TSV-files. Unit-tested.

---
 tablib/core.py             | 14 ++++++++++
 tablib/formats/__init__.py |  3 ++-
 tablib/formats/_tsv.py     | 53 ++++++++++++++++++++++++++++++++++++++
 test_tablib.py             | 44 +++++++++++++++++++++++++++++++
 4 files changed, 113 insertions(+), 1 deletion(-)
 create mode 100644 tablib/formats/_tsv.py

diff --git a/tablib/core.py b/tablib/core.py
index 6e5e60b..c4071c8 100644
--- a/tablib/core.py
+++ b/tablib/core.py
@@ -306,6 +306,20 @@ class Dataset(object):
 		"""
 		pass
 
+	@property
+	def tsv():
+		"""A TSV representation of the :class:`Dataset` object. The top row will contain 
+	    headers, if they have been set. Otherwise, the top row will contain 
+	    the first row of the dataset.
+
+	    A dataset object can also be imported by setting the :class:`Dataset.csv` attribute. ::
+
+	        data = tablib.Dataset()
+	        data.tsv = 'age\tfirst_name\tlast_name\\n90\tJohn\tAdams'
+
+	    Import assumes (for now) that headers exist.
+		"""
+
 	@property
 	def yaml():
 		"""A YAML representation of the :class:`Dataset` object. If headers have been 
diff --git a/tablib/formats/__init__.py b/tablib/formats/__init__.py
index 0ce9b71..f5960b8 100644
--- a/tablib/formats/__init__.py
+++ b/tablib/formats/__init__.py
@@ -7,5 +7,6 @@ import _csv as csv
 import _json as json
 import _xls as xls
 import _yaml as yaml
+import _tsv as tsv
 
-available = (json, xls, yaml, csv)
+available = (json, xls, yaml, csv, tsv)
diff --git a/tablib/formats/_tsv.py b/tablib/formats/_tsv.py
new file mode 100644
index 0000000..8603c45
--- /dev/null
+++ b/tablib/formats/_tsv.py
@@ -0,0 +1,53 @@
+# -*- coding: utf-8 -*-
+
+""" Tablib - TSV (Tab Separated Values) Support.
+"""
+
+import cStringIO
+import csv
+import os
+
+import tablib
+
+
+title = 'tsv'
+extentions = ('tsv',)
+
+
+
+def export_set(dataset):
+	"""Returns a TSV representation of Dataset."""
+	stream = cStringIO.StringIO()
+	_tsv = csv.writer(stream, delimiter="\t")
+
+	for row in dataset._package(dicts=False):
+		_tsv.writerow(row)
+
+	return stream.getvalue()
+
+
+def import_set(dset, in_stream, headers=True):
+	"""Returns dataset from TSV stream."""
+
+	dset.wipe()
+
+	rows = csv.reader(in_stream.split("\r\n"), delimiter="\t")
+	for i, row in enumerate(rows):
+
+		# Skip empty rows
+		if not row:
+			continue
+
+		if (i == 0) and (headers):
+			dset.headers = row
+		else:
+			dset.append(row)
+
+
+def detect(stream):
+	"""Returns True if given stream is valid TSV."""
+	try:
+		rows = dialect = csv.Sniffer().sniff(stream, delimiters="\t")
+		return True
+	except csv.Error:
+		return False
diff --git a/test_tablib.py b/test_tablib.py
index f875cb7..a576ce8 100755
--- a/test_tablib.py
+++ b/test_tablib.py
@@ -178,6 +178,22 @@ class TablibTestCase(unittest.TestCase):
 
 		self.assertEqual(csv, self.founders.csv)
 
+	def test_tsv_export(self):
+		"""Verify exporting dataset object as CSV."""
+
+		# Build up the csv string with headers first, followed by each row
+		tsv = ''
+		for col in self.headers:
+			tsv += col + '\t'
+
+		tsv = tsv.strip('\t') + '\r\n'
+
+		for founder in self.founders:
+			for col in founder:
+				tsv += str(col) + '\t'
+			tsv = tsv.strip('\t') + '\r\n'
+
+		self.assertEqual(tsv, self.founders.tsv)
 
 	def test_unicode_append(self):
 		"""Passes in a single unicode charecter and exports."""
@@ -188,6 +204,7 @@ class TablibTestCase(unittest.TestCase):
 		data.json
 		data.yaml
 		data.csv
+		data.tsv
 		data.xls
 
  
@@ -268,6 +285,18 @@ class TablibTestCase(unittest.TestCase):
 
 		self.assertEqual(_csv, data.csv)
 
+	def test_tsv_import_set(self):
+		"""Generate and import TSV set serialization."""
+		data.append(self.john)
+		data.append(self.george)
+		data.headers = self.headers
+
+		_tsv = data.tsv
+
+		data.tsv = _tsv
+
+		self.assertEqual(_tsv, data.tsv)
+
 	def test_csv_format_detect(self):
 		"""Test CSV format detection."""
 		
@@ -283,6 +312,21 @@ class TablibTestCase(unittest.TestCase):
 		self.assertTrue(tablib.formats.csv.detect(_csv))
 		self.assertFalse(tablib.formats.csv.detect(_bunk))
 
+	def test_tsv_format_detect(self):
+		"""Test TSV format detection."""
+		
+		_tsv = (
+			'1\t2\t3\n'
+			'4\t5\t6\n'
+			'7\t8\t9\n'
+		)
+		_bunk = (
+			'¡¡¡¡¡¡¡¡£™∞¢£§∞§¶•¶ª∞¶•ªº••ª–º§•†•§º¶•†¥ª–º•§ƒø¥¨©πƒø†ˆ¥ç©¨√øˆ¥≈†ƒ¥ç©ø¨çˆ¥ƒçø¶'
+		)
+		
+		self.assertTrue(tablib.formats.tsv.detect(_tsv))
+		self.assertFalse(tablib.formats.tsv.detect(_bunk))
+
 	def test_json_format_detect(self):
 		"""Test JSON format detection."""