add reader, csv reader, jsonlines reader

rltk/io/__init__.py

@@ -0,0 +1,3 @@
+from rltk.io.reader import Reader
+from rltk.io.csv_reader import CSVReader
+from rltk.io.jsonlines_reader import JsonLinesReader

rltk/io/csv_reader.py

@@ -0,0 +1,32 @@
+import csv
+import codecs
+
+from rltk.io.reader import Reader
+
+
+class CSVReader(Reader):
+
+    def __init__(self, filename, **kwargs):
+        self._filename = filename
+        self._kwargs = kwargs
+        self._loop_count = 0
+
+        self._file_handler = codecs.open(filename)
+        self._csv_reader = csv.DictReader(self._file_handler, **kwargs)
+
+    def __next__(self):
+        for obj in self._csv_reader:
+            self._loop_count += 1
+            yield {t[0]: t[1] for t in obj.items()}
+
+    def __del__(self):
+        try:
+            self._file_handler.close()
+        except:
+            pass
+
+    def __copy__(self):
+        new_reader = CSVReader(self._filename, **self._kwargs)
+        while new_reader._loop_count < self._loop_count:
+            new_reader.__next__()
+        return new_reader

rltk/io/jsonlines_reader.py

@@ -0,0 +1,36 @@
+import json
+import codecs
+
+from rltk.io.reader import Reader
+
+
+class JsonLinesReader(Reader):
+
+    def __init__(self, filename, ignore_blank_line=True):
+        self._filename = filename
+        self._loop_count = 0
+
+        self._file_handler = codecs.open(filename)
+        self._ignore_blank_line = ignore_blank_line
+
+    def __next__(self):
+        for line in self._file_handler:
+            if line.strip() == '':
+                if self._ignore_blank_line:
+                    continue
+                else:
+                    raise ValueError('Blank line detected')
+            self._loop_count += 1
+            yield json.loads(line)
+
+    def __del__(self):
+        try:
+            self._file_handler.close()
+        except:
+            pass
+
+    def __copy__(self):
+        new_reader = JsonLinesReader(self._filename, self._ignore_blank_line)
+        while new_reader._loop_count < self._loop_count:
+            new_reader.__next__()
+        return new_reader

rltk/io/reader.py

@@ -0,0 +1,21 @@
+import itertools
+
+
+class Reader(object):
+    def __init__(self):
+        pass
+
+    def __iter__(self):
+        return self.__next__()
+
+    def __next__(self):
+        raise NotImplementedError
+
+    def __del__(self):
+        pass
+
+    def __copy__(self):
+        raise NotImplementedError
+
+    def copy(self):
+        return self.__copy__()