refactor abstract adapter and reader, add block reader

rltk/io/adapter/__init__.py

@@ -1,26 +1,2 @@
-from rltk.record import Record
-
-
-class Adapter(object):
-    def __init__(self):
-        pass
-
-    def __del__(self):
-        pass
-
-    def get(self, record_id) -> Record:
-        raise NotImplementedError
-
-    def set(self, record_id, record: Record):
-        raise NotImplementedError
-
-    def __iter__(self):
-        return self.__next__()
-
-    def __next__(self):
-        """iterator is not required in adapter"""
-        pass
-
-
-# subclasses
+from rltk.io.adapter.adapter import Adapter
 from rltk.io.adapter.memory_adapter import MemoryAdapter

rltk/io/adapter/adapter.py

@@ -0,0 +1,22 @@
+from rltk.record import Record
+
+
+class Adapter(object):
+    def __init__(self):
+        pass
+
+    def __del__(self):
+        pass
+
+    def get(self, record_id) -> Record:
+        raise NotImplementedError
+
+    def set(self, record_id, record: Record):
+        raise NotImplementedError
+
+    def __iter__(self):
+        return self.__next__()
+
+    def __next__(self):
+        """iterator is not required in adapter"""
+        pass

rltk/io/reader/__init__.py

@@ -1,18 +1,4 @@
-class Reader(object):
-    def __init__(self):
-        pass
-
-    def __iter__(self):
-        return self.__next__()
-
-    def __next__(self):
-        """return raw content of one record"""
-        raise NotImplementedError
-
-    def __del__(self):
-        pass
-
-# subclasses
+from rltk.io.reader.reader import Reader
 from rltk.io.reader.csv_reader import CSVReader
 from rltk.io.reader.jsonlines_reader import JsonLinesReader
-from rltk.io.reader.block_reader import FileBlockReader
+from rltk.io.reader.block_reader import BlockReader

rltk/io/reader/block_reader.py

@@ -0,0 +1,16 @@
+from rltk.io.reader import Reader
+
+
+class BlockReader(Reader):
+
+    def __init__(self, raw_blocks):
+        self._raw_blocks = raw_blocks
+
+    def __next__(self):
+        for block in self._raw_blocks: # iterate json objects
+            for id1, id2s in block.items(): # iterate on one json object
+                for id2 in id2s:
+                    yield id1, id2
+
+    def __del__(self):
+        pass

rltk/io/reader/reader.py

@@ -0,0 +1,13 @@
+class Reader(object):
+    def __init__(self):
+        pass
+
+    def __iter__(self):
+        return self.__next__()
+
+    def __next__(self):
+        """return raw content of one record"""
+        raise NotImplementedError
+
+    def __del__(self):
+        pass