[core] Refactoring IOFormats so there is one and only obvious way to send it.

This is the commit where I admit that having more than one input/output
format for readers and writers was complicating the code too much for a
very small gain, and that it would be easier to only have one way to do
it.

So such way is now:

- Returning (or yielding) a dict if you have key-value type collections.
- Returning (or yielding) a tuple if you have a list-type collection.
- Returning (or yielding) something else otherwise, which will continue
  to work like the old "arg0" format.

IOFORMAT options has been removed in favour of a RemovedOption, which
will complain if you're still trying to set it to anything else than the
one value allowed.

bonobo/commands/convert.py

@@ -120,7 +120,8 @@ def register(parser):
     parser.add_argument(
         '--' + WRITER,
         '-w',
-        help='Choose the writer factory if it cannot be detected from extension, or if detection is wrong (use - for console pretty print).'
+        help=
+        'Choose the writer factory if it cannot be detected from extension, or if detection is wrong (use - for console pretty print).'
     )
     parser.add_argument(
         '--filter',

bonobo/config/options.py

@@ -66,10 +66,10 @@ class Option:
         self._creation_counter = Option._creation_counter
         Option._creation_counter += 1
 
-    def __get__(self, inst, typ):
+    def __get__(self, inst, type_):
         # XXX If we call this on the type, then either return overriden value or ... ???
         if inst is None:
-            return vars(type).get(self.name, self)
+            return vars(type_).get(self.name, self)
 
         if not self.name in inst._options_values:
             inst._options_values[self.name] = self.get_default()
@@ -96,6 +96,24 @@ class Option:
         return self.default() if callable(self.default) else self.default
 
 
+class RemovedOption(Option):
+    def __init__(self, *args, value, **kwargs):
+        kwargs['required'] = False
+        super(RemovedOption, self).__init__(*args, **kwargs)
+        self.value = value
+
+    def clean(self, value):
+        if value != self.value:
+            raise ValueError(
+                'Removed options cannot change value, {!r} must now be {!r} (and you should remove setting the value explicitely, as it is deprecated and will be removed quite soon.'.
+                format(self.name, self.value)
+            )
+        return self.value
+
+    def get_default(self):
+        return self.value
+
+
 class Method(Option):
     """
     A Method is a special callable-valued option, that can be used in three different ways (but for same purpose).

bonobo/config/processors.py

@@ -1,9 +1,8 @@
 from collections import Iterable
 from contextlib import contextmanager
 
-from bonobo.config.options import Option
+from bonobo.config import Option
-from bonobo.util.compat import deprecated_alias
+from bonobo.util import deprecated_alias, ensure_tuple
-from bonobo.util.iterators import ensure_tuple
 
 _CONTEXT_PROCESSORS_ATTR = '__processors__'
 
@@ -24,7 +23,7 @@ class ContextProcessor(Option):
     Example:
 
         >>> from bonobo.config import Configurable
-        >>> from bonobo.util.objects import ValueHolder
+        >>> from bonobo.util import ValueHolder
 
         >>> class Counter(Configurable):
         ...     @ContextProcessor
@@ -91,7 +90,10 @@ class ContextCurrifier:
         self._stack, self._stack_values = list(), list()
         for processor in resolve_processors(self.wrapped):
             _processed = processor(self.wrapped, *context, *self.context)
+            try:
                 _append_to_context = next(_processed)
+            except TypeError as exc:
+                raise TypeError('Context processor should be generators (using yield).') from exc
             self._stack_values.append(_append_to_context)
             if _append_to_context is not None:
                 self.context += ensure_tuple(_append_to_context)

bonobo/execution/base.py

@@ -5,6 +5,7 @@ from time import sleep
 from bonobo.config import create_container
 from bonobo.config.processors import ContextCurrifier
 from bonobo.plugins import get_enhancers
+from bonobo.util import inspect_node, isconfigurabletype
 from bonobo.util.errors import print_error
 from bonobo.util.objects import Wrapper, get_name
 
@@ -72,6 +73,15 @@ class LoopingExecutionContext(Wrapper):
         self._started = True
 
         self._stack = ContextCurrifier(self.wrapped, *self._get_initial_context())
+        if isconfigurabletype(self.wrapped):
+            # Not normal to have a partially configured object here, so let's warn the user instead of having get into
+            # the hard trouble of understanding that by himself.
+            raise TypeError(
+                'The Configurable should be fully instanciated by now, unfortunately I got a PartiallyConfigured object...'
+            )
+            # XXX enhance that, maybe giving hints on what's missing.
+            # print(inspect_node(self.wrapped))
+
         self._stack.setup(self)
 
         for enhancer in self._enhancers:

bonobo/execution/graph.py

@@ -1,3 +1,4 @@
+import time
 from functools import partial
 
 from bonobo.config import create_container
@@ -7,6 +8,9 @@ from bonobo.execution.plugin import PluginExecutionContext
 
 
 class GraphExecutionContext:
+    NodeExecutionContextType = NodeExecutionContext
+    PluginExecutionContextType = PluginExecutionContext
+
     @property
     def started(self):
         return any(node.started for node in self.nodes)
@@ -21,15 +25,17 @@ class GraphExecutionContext:
 
     def __init__(self, graph, plugins=None, services=None):
         self.graph = graph
-        self.nodes = [NodeExecutionContext(node, parent=self) for node in self.graph]
+        self.nodes = [self.create_node_execution_context_for(node) for node in self.graph]
-        self.plugins = [PluginExecutionContext(plugin, parent=self) for plugin in plugins or ()]
+        self.plugins = [self.create_plugin_execution_context_for(plugin) for plugin in plugins or ()]
         self.services = create_container(services)
 
         # Probably not a good idea to use it unless you really know what you're doing. But you can access the context.
         self.services['__graph_context'] = self
 
         for i, node_context in enumerate(self):
-            node_context.outputs = [self[j].input for j in self.graph.outputs_of(i)]
+            outputs = self.graph.outputs_of(i)
+            if len(outputs):
+                node_context.outputs = [self[j].input for j in outputs]
             node_context.input.on_begin = partial(node_context.send, BEGIN, _control=True)
             node_context.input.on_end = partial(node_context.send, END, _control=True)
             node_context.input.on_finalize = partial(node_context.stop)
@@ -43,6 +49,12 @@ class GraphExecutionContext:
     def __iter__(self):
         yield from self.nodes
 
+    def create_node_execution_context_for(self, node):
+        return self.NodeExecutionContextType(node, parent=self)
+
+    def create_plugin_execution_context_for(self, plugin):
+        return self.PluginExecutionContextType(plugin, parent=self)
+
     def write(self, *messages):
         """Push a list of messages in the inputs of this graph's inputs, matching the output of special node "BEGIN" in
         our graph."""
@@ -51,17 +63,23 @@ class GraphExecutionContext:
             for message in messages:
                 self[i].write(message)
 
-    def start(self):
+    def start(self, starter=None):
-        # todo use strategy
         for node in self.nodes:
+            if starter is None:
                 node.start()
+            else:
+                starter(node)
 
-    def stop(self):
+    def start_plugins(self, starter=None):
-        # todo use strategy
+        for plugin in self.plugins:
+            if starter is None:
+                plugin.start()
+            else:
+                starter(plugin)
+
+    def stop(self, stopper=None):
         for node in self.nodes:
+            if stopper is None:
                 node.stop()
-
+            else:
-    def loop(self):
+                stopper(node)
-        # todo use strategy
-        for node in self.nodes:
-            node.loop()

bonobo/execution/node.py

@@ -2,15 +2,15 @@ import traceback
 from queue import Empty
 from time import sleep
 
-from bonobo.constants import INHERIT_INPUT, NOT_MODIFIED
+from bonobo import settings
+from bonobo.constants import INHERIT_INPUT, NOT_MODIFIED, BEGIN, END
 from bonobo.errors import InactiveReadableError, UnrecoverableError
 from bonobo.execution.base import LoopingExecutionContext
 from bonobo.structs.bags import Bag
 from bonobo.structs.inputs import Input
+from bonobo.util import get_name, iserrorbag, isloopbackbag, isdict, istuple
 from bonobo.util.compat import deprecated_alias
-from bonobo.util.inspect import iserrorbag, isloopbackbag
 from bonobo.util.iterators import iter_if_not_sequence
-from bonobo.util.objects import get_name
 from bonobo.util.statistics import WithStatistics
 
 
@@ -28,12 +28,12 @@ class NodeExecutionContext(WithStatistics, LoopingExecutionContext):
     def alive_str(self):
         return '+' if self.alive else '-'
 
-    def __init__(self, wrapped, parent=None, services=None):
+    def __init__(self, wrapped, parent=None, services=None, _input=None, _outputs=None):
         LoopingExecutionContext.__init__(self, wrapped, parent=parent, services=services)
         WithStatistics.__init__(self, 'in', 'out', 'err')
 
-        self.input = Input()
+        self.input = _input or Input()
-        self.outputs = []
+        self.outputs = _outputs or []
 
     def __str__(self):
         return self.alive_str + ' ' + self.__name__ + self.get_statistics_as_string(prefix=' ')
@@ -51,6 +51,11 @@ class NodeExecutionContext(WithStatistics, LoopingExecutionContext):
         for message in messages:
             self.input.put(message)
 
+    def write_sync(self, *messages):
+        self.write(BEGIN, *messages, END)
+        for _ in messages:
+            self.step()
+
     # XXX deprecated alias
     recv = deprecated_alias('recv', write)
 
@@ -143,12 +148,18 @@ def _resolve(input_bag, output):
         return output
 
     # If it does not look like a bag, let's create one for easier manipulation
-    if hasattr(output, 'apply'):
+    if hasattr(output, 'apply'):  # XXX TODO use isbag() ?
         # Already a bag? Check if we need to set parent.
         if INHERIT_INPUT in output.flags:
             output.set_parent(input_bag)
-    else:
-        # Not a bag? Let's encapsulate it.
-        output = Bag(output)
-
         return output
+
+    # If we're using kwargs ioformat, then a dict means kwargs.
+    if settings.IOFORMAT == settings.IOFORMAT_KWARGS and isdict(output):
+        return Bag(**output)
+
+    if istuple(output):
+        return Bag(*output)
+
+    # Either we use arg0 format, either it's "just" a value.
+    return Bag(output)

bonobo/nodes/io/base.py

@@ -1,39 +1,4 @@
-from bonobo import settings
 from bonobo.config import Configurable, ContextProcessor, Option, Service
-from bonobo.errors import UnrecoverableValueError, UnrecoverableNotImplementedError
-from bonobo.structs.bags import Bag
-
-
-class IOFormatEnabled(Configurable):
-    ioformat = Option(default=settings.IOFORMAT.get)
-
-    def get_input(self, *args, **kwargs):
-        if self.ioformat == settings.IOFORMAT_ARG0:
-            if len(args) != 1 or len(kwargs):
-                raise UnrecoverableValueError(
-                    'Wrong input formating: IOFORMAT=ARG0 implies one arg and no kwargs, got args={!r} and kwargs={!r}.'.
-                    format(args, kwargs)
-                )
-            return args[0]
-
-        if self.ioformat == settings.IOFORMAT_KWARGS:
-            if len(args) or not len(kwargs):
-                raise UnrecoverableValueError(
-                    'Wrong input formating: IOFORMAT=KWARGS ioformat implies no arg, got args={!r} and kwargs={!r}.'.
-                    format(args, kwargs)
-                )
-            return kwargs
-
-        raise UnrecoverableNotImplementedError('Unsupported format.')
-
-    def get_output(self, row):
-        if self.ioformat == settings.IOFORMAT_ARG0:
-            return row
-
-        if self.ioformat == settings.IOFORMAT_KWARGS:
-            return Bag(**row)
-
-        raise UnrecoverableNotImplementedError('Unsupported format.')
 
 
 class FileHandler(Configurable):

bonobo/nodes/io/csv.py

@@ -1,10 +1,11 @@
 import csv
 
 from bonobo.config import Option
+from bonobo.config.options import RemovedOption
 from bonobo.config.processors import ContextProcessor
 from bonobo.constants import NOT_MODIFIED
+from bonobo.nodes.io.base import FileHandler
 from bonobo.nodes.io.file import FileReader, FileWriter
-from bonobo.nodes.io.base import FileHandler, IOFormatEnabled
 from bonobo.util.objects import ValueHolder
 
 
@@ -27,9 +28,10 @@ class CsvHandler(FileHandler):
     delimiter = Option(str, default=';')
     quotechar = Option(str, default='"')
     headers = Option(tuple, required=False)
+    ioformat = RemovedOption(positional=False, value='kwargs')
 
 
-class CsvReader(IOFormatEnabled, FileReader, CsvHandler):
+class CsvReader(FileReader, CsvHandler):
     """
     Reads a CSV and yield the values as dicts.
 
@@ -62,18 +64,17 @@ class CsvReader(IOFormatEnabled, FileReader, CsvHandler):
             if len(row) != field_count:
                 raise ValueError('Got a line with %d fields, expecting %d.' % (len(row), field_count, ))
 
-            yield self.get_output(dict(zip(_headers, row)))
+            yield dict(zip(_headers, row))
 
 
-class CsvWriter(IOFormatEnabled, FileWriter, CsvHandler):
+class CsvWriter(FileWriter, CsvHandler):
     @ContextProcessor
     def writer(self, context, fs, file, lineno):
         writer = csv.writer(file, delimiter=self.delimiter, quotechar=self.quotechar, lineterminator=self.eol)
         headers = ValueHolder(list(self.headers) if self.headers else None)
         yield writer, headers
 
-    def write(self, fs, file, lineno, writer, headers, *args, **kwargs):
+    def write(self, fs, file, lineno, writer, headers, **row):
-        row = self.get_input(*args, **kwargs)
         if not lineno:
             headers.set(headers.value or row.keys())
             writer.writerow(headers.get())

bonobo/nodes/io/json.py

@@ -1,8 +1,9 @@
 import json
 
+from bonobo.config.options import RemovedOption
 from bonobo.config.processors import ContextProcessor
 from bonobo.constants import NOT_MODIFIED
-from bonobo.nodes.io.base import FileHandler, IOFormatEnabled
+from bonobo.nodes.io.base import FileHandler
 from bonobo.nodes.io.file import FileReader, FileWriter
 from bonobo.structs.bags import Bag
 
@@ -10,14 +11,15 @@ from bonobo.structs.bags import Bag
 class JsonHandler(FileHandler):
     eol = ',\n'
     prefix, suffix = '[', ']'
+    ioformat = RemovedOption(positional=False, value='kwargs')
 
 
-class JsonReader(IOFormatEnabled, FileReader, JsonHandler):
+class JsonReader(FileReader, JsonHandler):
     loader = staticmethod(json.load)
 
     def read(self, fs, file):
         for line in self.loader(file):
-            yield self.get_output(line)
+            yield line
 
 
 class JsonDictItemsReader(JsonReader):
@@ -26,21 +28,20 @@ class JsonDictItemsReader(JsonReader):
             yield Bag(*line)
 
 
-class JsonWriter(IOFormatEnabled, FileWriter, JsonHandler):
+class JsonWriter(FileWriter, JsonHandler):
     @ContextProcessor
     def envelope(self, context, fs, file, lineno):
         file.write(self.prefix)
         yield
         file.write(self.suffix)
 
-    def write(self, fs, file, lineno, *args, **kwargs):
+    def write(self, fs, file, lineno, **row):
         """
         Write a json row on the next line of file pointed by ctx.file.
 
         :param ctx:
         :param row:
         """
-        row = self.get_input(*args, **kwargs)
         self._write_line(file, (self.eol if lineno.value else '') + json.dumps(row))
         lineno += 1
         return NOT_MODIFIED

bonobo/nodes/io/pickle.py

@@ -1,9 +1,10 @@
 import pickle
 
 from bonobo.config import Option
+from bonobo.config.options import RemovedOption
 from bonobo.config.processors import ContextProcessor
 from bonobo.constants import NOT_MODIFIED
-from bonobo.nodes.io.base import FileHandler, IOFormatEnabled
+from bonobo.nodes.io.base import FileHandler
 from bonobo.nodes.io.file import FileReader, FileWriter
 from bonobo.util.objects import ValueHolder
 
@@ -20,7 +21,7 @@ class PickleHandler(FileHandler):
     item_names = Option(tuple, required=False)
 
 
-class PickleReader(IOFormatEnabled, FileReader, PickleHandler):
+class PickleReader(FileReader, PickleHandler):
     """
     Reads a Python pickle object and yields the items in dicts.
     """
@@ -54,10 +55,10 @@ class PickleReader(IOFormatEnabled, FileReader, PickleHandler):
             if len(i) != item_count:
                 raise ValueError('Received an object with %d items, expecting %d.' % (len(i), item_count, ))
 
-            yield self.get_output(dict(zip(i)) if is_dict else dict(zip(pickle_headers.value, i)))
+            yield dict(zip(i)) if is_dict else dict(zip(pickle_headers.value, i))
 
 
-class PickleWriter(IOFormatEnabled, FileWriter, PickleHandler):
+class PickleWriter(FileWriter, PickleHandler):
     mode = Option(str, default='wb')
 
     def write(self, fs, file, lineno, item):

bonobo/settings.py

@@ -42,6 +42,9 @@ class Setting:
     def __repr__(self):
         return '<Setting {}={!r}>'.format(self.name, self.get())
 
+    def __eq__(self, other):
+        return self.get() == other
+
     def set(self, value):
         value = self.formatter(value) if self.formatter else value
         if self.validator and not self.validator(value):

bonobo/strategies/base.py

@@ -6,10 +6,13 @@ class Strategy:
     Base class for execution strategies.
 
     """
-    graph_execution_context_factory = GraphExecutionContext
+    GraphExecutionContextType = GraphExecutionContext
 
-    def create_graph_execution_context(self, graph, *args, **kwargs):
+    def __init__(self, GraphExecutionContextType=None):
-        return self.graph_execution_context_factory(graph, *args, **kwargs)
+        self.GraphExecutionContextType = GraphExecutionContextType or self.GraphExecutionContextType
+
+    def create_graph_execution_context(self, graph, *args, GraphExecutionContextType=None, **kwargs):
+        return (GraphExecutionContextType or self.GraphExecutionContextType)(graph, *args, **kwargs)
 
     def execute(self, graph, *args, **kwargs):
         raise NotImplementedError

bonobo/strategies/executor.py

@@ -19,42 +19,16 @@ class ExecutorStrategy(Strategy):
     def create_executor(self):
         return self.executor_factory()
 
-    def execute(self, graph, *args, plugins=None, services=None, **kwargs):
+    def execute(self, graph, **kwargs):
-        context = self.create_graph_execution_context(graph, plugins=plugins, services=services)
+        context = self.create_graph_execution_context(graph, **kwargs)
         context.write(BEGIN, Bag(), END)
 
         executor = self.create_executor()
 
         futures = []
 
-        for plugin_context in context.plugins:
+        context.start_plugins(self.get_plugin_starter(executor, futures))
-
+        context.start(self.get_starter(executor, futures))
-            def _runner(plugin_context=plugin_context):
-                with plugin_context:
-                    try:
-                        plugin_context.loop()
-                    except Exception as exc:
-                        print_error(exc, traceback.format_exc(), context=plugin_context)
-
-            futures.append(executor.submit(_runner))
-
-        for node_context in context.nodes:
-
-            def _runner(node_context=node_context):
-                try:
-                    node_context.start()
-                except Exception as exc:
-                    print_error(exc, traceback.format_exc(), context=node_context, method='start')
-                    node_context.input.on_end()
-                else:
-                    node_context.loop()
-
-                try:
-                    node_context.stop()
-                except Exception as exc:
-                    print_error(exc, traceback.format_exc(), context=node_context, method='stop')
-
-            futures.append(executor.submit(_runner))
 
         while context.alive:
             time.sleep(0.1)
@@ -62,10 +36,45 @@ class ExecutorStrategy(Strategy):
         for plugin_context in context.plugins:
             plugin_context.shutdown()
 
+        context.stop()
+
         executor.shutdown()
 
         return context
 
+    def get_starter(self, executor, futures):
+        def starter(node):
+            def _runner():
+                try:
+                    node.start()
+                except Exception as exc:
+                    print_error(exc, traceback.format_exc(), context=node, method='start')
+                    node.input.on_end()
+                else:
+                    node.loop()
+
+                try:
+                    node.stop()
+                except Exception as exc:
+                    print_error(exc, traceback.format_exc(), context=node, method='stop')
+
+            futures.append(executor.submit(_runner))
+
+        return starter
+
+    def get_plugin_starter(self, executor, futures):
+        def plugin_starter(plugin):
+            def _runner():
+                with plugin:
+                    try:
+                        plugin.loop()
+                    except Exception as exc:
+                        print_error(exc, traceback.format_exc(), context=plugin)
+
+            futures.append(executor.submit(_runner))
+
+        return plugin_starter
+
 
 class ThreadPoolExecutorStrategy(ExecutorStrategy):
     executor_factory = ThreadPoolExecutor

bonobo/strategies/naive.py

@@ -4,13 +4,23 @@ from bonobo.structs.bags import Bag
 
 
 class NaiveStrategy(Strategy):
-    def execute(self, graph, *args, plugins=None, **kwargs):
+    # TODO: how to run plugins in "naive" mode ?
-        context = self.create_graph_execution_context(graph, plugins=plugins)
+
+    def execute(self, graph, **kwargs):
+        context = self.create_graph_execution_context(graph, **kwargs)
         context.write(BEGIN, Bag(), END)
 
-        # TODO: how to run plugins in "naive" mode ?
+        # start
         context.start()
-        context.loop()
+
+        # loop
+        nodes = list(context.nodes)
+        while len(nodes):
+            for node in nodes:
+                node.loop()
+            nodes = list(node for node in nodes if node.alive)
+
+        # stop
         context.stop()
 
         return context

bonobo/structs/bags.py

@@ -96,7 +96,29 @@ class Bag:
         return cls(*args, _flags=(INHERIT_INPUT, ), **kwargs)
 
     def __eq__(self, other):
-        return isinstance(other, Bag) and other.args == self.args and other.kwargs == self.kwargs
+        # XXX there are overlapping cases, but this is very handy for now. Let's think about it later.
+
+        # bag
+        if isinstance(other, Bag) and other.args == self.args and other.kwargs == self.kwargs:
+            return True
+
+        # tuple of (tuple, dict)
+        if isinstance(other, tuple) and len(other) == 2 and other[0] == self.args and other[1] == self.kwargs:
+            return True
+
+        # tuple (aka args)
+        if isinstance(other, tuple) and other == self.args:
+            return True
+
+        # dict (aka kwargs)
+        if isinstance(other, dict) and not self.args and other == self.kwargs:
+            return True
+
+        # arg0
+        if len(self.args) == 1 and not len(self.kwargs) and self.args[0] == other:
+            return True
+
+        return False
 
     def __repr__(self):
         return '<{} ({})>'.format(

bonobo/util/__init__.py

@@ -1,14 +1,18 @@
 from bonobo.util.collections import sortedlist
+from bonobo.util.iterators import ensure_tuple
+from bonobo.util.compat import deprecated, deprecated_alias
 from bonobo.util.inspect import (
     inspect_node,
     isbag,
     isconfigurable,
     isconfigurabletype,
     iscontextprocessor,
+    isdict,
     iserrorbag,
     isloopbackbag,
     ismethod,
     isoption,
+    istuple,
     istype,
 )
 from bonobo.util.objects import (get_name, get_attribute_or_create, ValueHolder)
@@ -17,6 +21,8 @@ from bonobo.util.python import require
 # Bonobo's util API
 __all__ = [
     'ValueHolder',
+    'deprecated',
+    'deprecated_alias',
     'get_attribute_or_create',
     'get_name',
     'inspect_node',
@@ -24,6 +30,7 @@ __all__ = [
     'isconfigurable',
     'isconfigurabletype',
     'iscontextprocessor',
+    'isdict',
     'iserrorbag',
     'isloopbackbag',
     'ismethod',

bonobo/util/inspect.py

@@ -68,6 +68,26 @@ def istype(mixed):
     return isinstance(mixed, type)
 
 
+def isdict(mixed):
+    """
+    Check if the given argument is a dict.
+
+    :param mixed:
+    :return: bool
+    """
+    return isinstance(mixed, dict)
+
+
+def istuple(mixed):
+    """
+    Check if the given argument is a tuple.
+
+    :param mixed:
+    :return: bool
+    """
+    return isinstance(mixed, tuple)
+
+
 def isbag(mixed):
     """
     Check if the given argument is an instance of a :class:`bonobo.Bag`.

bonobo/util/iterators.py

@@ -38,6 +38,6 @@ def tuplize(generator):
 
 
 def iter_if_not_sequence(mixed):
-    if isinstance(mixed, (dict, list, str)):
+    if isinstance(mixed, (dict, list, str, bytes, )):
         raise TypeError(type(mixed).__name__)
     return iter(mixed)

bonobo/util/testing.py

@@ -1,16 +1,10 @@
 from contextlib import contextmanager
-from unittest.mock import MagicMock
 
-from bonobo import open_fs
+from bonobo import open_fs, Token
+from bonobo.execution import GraphExecutionContext
 from bonobo.execution.node import NodeExecutionContext
 
 
-class CapturingNodeExecutionContext(NodeExecutionContext):
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.send = MagicMock()
-
-
 @contextmanager
 def optional_contextmanager(cm, *, ignore=False):
     if cm is None or ignore:
@@ -35,3 +29,38 @@ class FilesystemTester:
     def get_services_for_writer(self, tmpdir):
         fs, filename = open_fs(tmpdir), 'output.' + self.extension
         return fs, filename, {'fs': fs}
+
+
+class QueueList(list):
+    def append(self, item):
+        if not isinstance(item, Token):
+            super(QueueList, self).append(item)
+
+    put = append
+
+
+class BufferingContext:
+    def __init__(self, buffer=None):
+        if buffer is None:
+            buffer = QueueList()
+        self.buffer = buffer
+
+    def get_buffer(self):
+        return self.buffer
+
+
+class BufferingNodeExecutionContext(BufferingContext, NodeExecutionContext):
+    def __init__(self, *args, buffer=None, **kwargs):
+        BufferingContext.__init__(self, buffer)
+        NodeExecutionContext.__init__(self, *args, **kwargs, _outputs=[self.buffer])
+
+
+class BufferingGraphExecutionContext(BufferingContext, GraphExecutionContext):
+    NodeExecutionContextType = BufferingNodeExecutionContext
+
+    def __init__(self, *args, buffer=None, **kwargs):
+        BufferingContext.__init__(self, buffer)
+        GraphExecutionContext.__init__(self, *args, **kwargs)
+
+    def create_node_execution_context_for(self, node):
+        return self.NodeExecutionContextType(node, parent=self, buffer=self.buffer)

tests/execution/test_node.py

@@ -0,0 +1,104 @@
+from bonobo import Bag, Graph
+from bonobo.strategies import NaiveStrategy
+from bonobo.util.testing import BufferingNodeExecutionContext, BufferingGraphExecutionContext
+
+
+def test_node_string():
+    def f():
+        return 'foo'
+
+    with BufferingNodeExecutionContext(f) as context:
+        context.write_sync(Bag())
+        output = context.get_buffer()
+
+        assert len(output) == 1
+        assert output[0] == (('foo', ), {})
+
+    def g():
+        yield 'foo'
+        yield 'bar'
+
+    with BufferingNodeExecutionContext(g) as context:
+        context.write_sync(Bag())
+        output = context.get_buffer()
+
+        assert len(output) == 2
+        assert output[0] == (('foo', ), {})
+        assert output[1] == (('bar', ), {})
+
+
+def test_node_bytes():
+    def f():
+        return b'foo'
+
+    with BufferingNodeExecutionContext(f) as context:
+        context.write_sync(Bag())
+
+        output = context.get_buffer()
+        assert len(output) == 1
+        assert output[0] == ((b'foo', ), {})
+
+    def g():
+        yield b'foo'
+        yield b'bar'
+
+    with BufferingNodeExecutionContext(g) as context:
+        context.write_sync(Bag())
+        output = context.get_buffer()
+
+        assert len(output) == 2
+        assert output[0] == ((b'foo', ), {})
+        assert output[1] == ((b'bar', ), {})
+
+
+def test_node_dict():
+    def f():
+        return {'id': 1, 'name': 'foo'}
+
+    with BufferingNodeExecutionContext(f) as context:
+        context.write_sync(Bag())
+        output = context.get_buffer()
+
+        assert len(output) == 1
+        assert output[0] == {'id': 1, 'name': 'foo'}
+
+    def g():
+        yield {'id': 1, 'name': 'foo'}
+        yield {'id': 2, 'name': 'bar'}
+
+    with BufferingNodeExecutionContext(g) as context:
+        context.write_sync(Bag())
+        output = context.get_buffer()
+
+        assert len(output) == 2
+        assert output[0] == {'id': 1, 'name': 'foo'}
+        assert output[1] == {'id': 2, 'name': 'bar'}
+
+
+def test_node_dict_chained():
+    strategy = NaiveStrategy(GraphExecutionContextType=BufferingGraphExecutionContext)
+
+    def uppercase_name(**kwargs):
+        return {**kwargs, 'name': kwargs['name'].upper()}
+
+    def f():
+        return {'id': 1, 'name': 'foo'}
+
+    graph = Graph(f, uppercase_name)
+    context = strategy.execute(graph)
+    output = context.get_buffer()
+
+    assert len(output) == 1
+    assert output[0] == {'id': 1, 'name': 'FOO'}
+
+    def g():
+        yield {'id': 1, 'name': 'foo'}
+        yield {'id': 2, 'name': 'bar'}
+
+    graph = Graph(g, uppercase_name)
+    context = strategy.execute(graph)
+    output = context.get_buffer()
+
+    assert len(output) == 2
+    assert output[0] == {'id': 1, 'name': 'FOO'}
+    assert output[1] == {'id': 2, 'name': 'BAR'}

tests/io/test_csv.py

@@ -3,25 +3,19 @@ import pytest
 from bonobo import Bag, CsvReader, CsvWriter, settings
 from bonobo.constants import BEGIN, END
 from bonobo.execution.node import NodeExecutionContext
-from bonobo.util.testing import CapturingNodeExecutionContext, FilesystemTester
+from bonobo.util.testing import FilesystemTester, BufferingNodeExecutionContext
 
 csv_tester = FilesystemTester('csv')
 csv_tester.input_data = 'a,b,c\na foo,b foo,c foo\na bar,b bar,c bar'
 
 
-def test_write_csv_to_file_arg0(tmpdir):
+def test_write_csv_ioformat_arg0(tmpdir):
     fs, filename, services = csv_tester.get_services_for_writer(tmpdir)
+    with pytest.raises(ValueError):
+        CsvWriter(path=filename, ioformat=settings.IOFORMAT_ARG0)
 
-    with NodeExecutionContext(CsvWriter(path=filename, ioformat=settings.IOFORMAT_ARG0), services=services) as context:
+    with pytest.raises(ValueError):
-        context.write(BEGIN, Bag({'foo': 'bar'}), Bag({'foo': 'baz', 'ignore': 'this'}), END)
+        CsvReader(path=filename, delimiter=',', ioformat=settings.IOFORMAT_ARG0),
-        context.step()
-        context.step()
-
-    with fs.open(filename) as fp:
-        assert fp.read() == 'foo\nbar\nbaz\n'
-
-    with pytest.raises(AttributeError):
-        getattr(context, 'file')
 
 
 @pytest.mark.parametrize('add_kwargs', ({}, {
@@ -30,7 +24,7 @@ def test_write_csv_to_file_arg0(tmpdir):
 def test_write_csv_to_file_kwargs(tmpdir, add_kwargs):
     fs, filename, services = csv_tester.get_services_for_writer(tmpdir)
 
-    with NodeExecutionContext(CsvWriter(path=filename, **add_kwargs), services=services) as context:
+    with NodeExecutionContext(CsvWriter(filename, **add_kwargs), services=services) as context:
         context.write(BEGIN, Bag(**{'foo': 'bar'}), Bag(**{'foo': 'baz', 'ignore': 'this'}), END)
         context.step()
         context.step()
@@ -42,61 +36,24 @@ def test_write_csv_to_file_kwargs(tmpdir, add_kwargs):
         getattr(context, 'file')
 
 
-def test_read_csv_from_file_arg0(tmpdir):
-    fs, filename, services = csv_tester.get_services_for_reader(tmpdir)
-
-    with CapturingNodeExecutionContext(
-        CsvReader(path=filename, delimiter=',', ioformat=settings.IOFORMAT_ARG0),
-        services=services,
-    ) as context:
-        context.write(BEGIN, Bag(), END)
-        context.step()
-
-    assert len(context.send.mock_calls) == 2
-
-    args0, kwargs0 = context.send.call_args_list[0]
-    assert len(args0) == 1 and not len(kwargs0)
-    args1, kwargs1 = context.send.call_args_list[1]
-    assert len(args1) == 1 and not len(kwargs1)
-
-    assert args0[0].args[0] == {
-        'a': 'a foo',
-        'b': 'b foo',
-        'c': 'c foo',
-    }
-    assert args1[0].args[0] == {
-        'a': 'a bar',
-        'b': 'b bar',
-        'c': 'c bar',
-    }
-
-
 def test_read_csv_from_file_kwargs(tmpdir):
     fs, filename, services = csv_tester.get_services_for_reader(tmpdir)
 
-    with CapturingNodeExecutionContext(
+    with BufferingNodeExecutionContext(
         CsvReader(path=filename, delimiter=','),
         services=services,
     ) as context:
         context.write(BEGIN, Bag(), END)
         context.step()
+        output = context.get_buffer()
 
-    assert len(context.send.mock_calls) == 2
+    assert len(output) == 2
-
+    assert output[0] == {
-    args0, kwargs0 = context.send.call_args_list[0]
-    assert len(args0) == 1 and not len(kwargs0)
-    args1, kwargs1 = context.send.call_args_list[1]
-    assert len(args1) == 1 and not len(kwargs1)
-
-    _args, _kwargs = args0[0].get()
-    assert not len(_args) and _kwargs == {
         'a': 'a foo',
         'b': 'b foo',
         'c': 'c foo',
     }
-
+    assert output[1] == {
-    _args, _kwargs = args1[0].get()
-    assert not len(_args) and _kwargs == {
         'a': 'a bar',
         'b': 'b bar',
         'c': 'c bar',

tests/io/test_file.py

@@ -3,7 +3,7 @@ import pytest
 from bonobo import Bag, FileReader, FileWriter
 from bonobo.constants import BEGIN, END
 from bonobo.execution.node import NodeExecutionContext
-from bonobo.util.testing import CapturingNodeExecutionContext, FilesystemTester
+from bonobo.util.testing import BufferingNodeExecutionContext, FilesystemTester
 
 txt_tester = FilesystemTester('txt')
 txt_tester.input_data = 'Hello\nWorld\n'
@@ -41,16 +41,10 @@ def test_file_writer_in_context(tmpdir, lines, output):
 def test_file_reader(tmpdir):
     fs, filename, services = txt_tester.get_services_for_reader(tmpdir)
 
-    with CapturingNodeExecutionContext(FileReader(path=filename), services=services) as context:
+    with BufferingNodeExecutionContext(FileReader(path=filename), services=services) as context:
-        context.write(BEGIN, Bag(), END)
+        context.write_sync(Bag())
-        context.step()
+        output = context.get_buffer()
 
-    assert len(context.send.mock_calls) == 2
+    assert len(output) == 2
-
+    assert output[0] == 'Hello'
-    args0, kwargs0 = context.send.call_args_list[0]
+    assert output[1] == 'World'
-    assert len(args0) == 1 and not len(kwargs0)
-    args1, kwargs1 = context.send.call_args_list[1]
-    assert len(args1) == 1 and not len(kwargs1)
-
-    assert args0[0].args[0] == 'Hello'
-    assert args1[0].args[0] == 'World'

tests/io/test_json.py

@@ -3,21 +3,20 @@ import pytest
 from bonobo import Bag, JsonReader, JsonWriter, settings
 from bonobo.constants import BEGIN, END
 from bonobo.execution.node import NodeExecutionContext
-from bonobo.util.testing import CapturingNodeExecutionContext, FilesystemTester
+from bonobo.util.testing import FilesystemTester
 
 json_tester = FilesystemTester('json')
 json_tester.input_data = '''[{"x": "foo"},{"x": "bar"}]'''
 
 
-def test_write_json_arg0(tmpdir):
+def test_write_json_ioformat_arg0(tmpdir):
     fs, filename, services = json_tester.get_services_for_writer(tmpdir)
 
-    with NodeExecutionContext(JsonWriter(filename, ioformat=settings.IOFORMAT_ARG0), services=services) as context:
+    with pytest.raises(ValueError):
-        context.write(BEGIN, Bag({'foo': 'bar'}), END)
+        JsonWriter(filename, ioformat=settings.IOFORMAT_ARG0)
-        context.step()
 
-    with fs.open(filename) as fp:
+    with pytest.raises(ValueError):
-        assert fp.read() == '[{"foo": "bar"}]'
+        JsonReader(filename, ioformat=settings.IOFORMAT_ARG0),
 
 
 @pytest.mark.parametrize('add_kwargs', ({}, {
@@ -32,24 +31,3 @@ def test_write_json_kwargs(tmpdir, add_kwargs):
 
     with fs.open(filename) as fp:
         assert fp.read() == '[{"foo": "bar"}]'
-
-
-def test_read_json_arg0(tmpdir):
-    fs, filename, services = json_tester.get_services_for_reader(tmpdir)
-
-    with CapturingNodeExecutionContext(
-        JsonReader(filename, ioformat=settings.IOFORMAT_ARG0),
-        services=services,
-    ) as context:
-        context.write(BEGIN, Bag(), END)
-        context.step()
-
-    assert len(context.send.mock_calls) == 2
-
-    args0, kwargs0 = context.send.call_args_list[0]
-    assert len(args0) == 1 and not len(kwargs0)
-    args1, kwargs1 = context.send.call_args_list[1]
-    assert len(args1) == 1 and not len(kwargs1)
-
-    assert args0[0].args[0] == {'x': 'foo'}
-    assert args1[0].args[0] == {'x': 'bar'}

tests/io/test_pickle.py

@@ -2,10 +2,9 @@ import pickle
 
 import pytest
 
-from bonobo import Bag, PickleReader, PickleWriter, settings
+from bonobo import Bag, PickleReader, PickleWriter
-from bonobo.constants import BEGIN, END
 from bonobo.execution.node import NodeExecutionContext
-from bonobo.util.testing import CapturingNodeExecutionContext, FilesystemTester
+from bonobo.util.testing import BufferingNodeExecutionContext, FilesystemTester
 
 pickle_tester = FilesystemTester('pkl', mode='wb')
 pickle_tester.input_data = pickle.dumps([['a', 'b', 'c'], ['a foo', 'b foo', 'c foo'], ['a bar', 'b bar', 'c bar']])
@@ -14,10 +13,8 @@ pickle_tester.input_data = pickle.dumps([['a', 'b', 'c'], ['a foo', 'b foo', 'c
 def test_write_pickled_dict_to_file(tmpdir):
     fs, filename, services = pickle_tester.get_services_for_writer(tmpdir)
 
-    with NodeExecutionContext(PickleWriter(filename, ioformat=settings.IOFORMAT_ARG0), services=services) as context:
+    with NodeExecutionContext(PickleWriter(filename), services=services) as context:
-        context.write(BEGIN, Bag({'foo': 'bar'}), Bag({'foo': 'baz', 'ignore': 'this'}), END)
+        context.write_sync(Bag({'foo': 'bar'}), Bag({'foo': 'baz', 'ignore': 'this'}))
-        context.step()
-        context.step()
 
     with fs.open(filename, 'rb') as fp:
         assert pickle.loads(fp.read()) == {'foo': 'bar'}
@@ -29,25 +26,17 @@ def test_write_pickled_dict_to_file(tmpdir):
 def test_read_pickled_list_from_file(tmpdir):
     fs, filename, services = pickle_tester.get_services_for_reader(tmpdir)
 
-    with CapturingNodeExecutionContext(
+    with BufferingNodeExecutionContext(PickleReader(filename), services=services) as context:
-        PickleReader(filename, ioformat=settings.IOFORMAT_ARG0), services=services
+        context.write_sync(Bag())
-    ) as context:
+        output = context.get_buffer()
-        context.write(BEGIN, Bag(), END)
-        context.step()
 
-    assert len(context.send.mock_calls) == 2
+    assert len(output) == 2
-
+    assert output[0] == {
-    args0, kwargs0 = context.send.call_args_list[0]
-    assert len(args0) == 1 and not len(kwargs0)
-    args1, kwargs1 = context.send.call_args_list[1]
-    assert len(args1) == 1 and not len(kwargs1)
-
-    assert args0[0].args[0] == {
         'a': 'a foo',
         'b': 'b foo',
         'c': 'c foo',
     }
-    assert args1[0].args[0] == {
+    assert output[1] == {
         'a': 'a bar',
         'b': 'b bar',
         'c': 'c bar',

tests/test_basicusage.py

@@ -12,4 +12,5 @@ def test_run_graph_noop():
 
     with patch('bonobo._api._is_interactive_console', side_effect=lambda: False):
         result = bonobo.run(graph)
+
     assert isinstance(result, GraphExecutionContext)

tests/test_commands.py

@@ -30,9 +30,13 @@ def test_entrypoint():
     for command in pkg_resources.iter_entry_points('bonobo.commands'):
         commands[command.name] = command
 
-    assert 'init' in commands
+    assert not {
-    assert 'run' in commands
+        'convert',
-    assert 'version' in commands
+        'init',
+        'inspect',
+        'run',
+        'version',
+    }.difference(set(commands))
 
 
 @all_runners

tests/test_execution.py

@@ -51,31 +51,31 @@ def test_simple_execution_context():
     graph = Graph()
     graph.add_chain(*chain)
 
-    ctx = GraphExecutionContext(graph)
+    context = GraphExecutionContext(graph)
-    assert len(ctx.nodes) == len(chain)
+    assert len(context.nodes) == len(chain)
-    assert not len(ctx.plugins)
+    assert not len(context.plugins)
 
     for i, node in enumerate(chain):
-        assert ctx[i].wrapped is node
+        assert context[i].wrapped is node
 
-    assert not ctx.alive
+    assert not context.alive
-    assert not ctx.started
+    assert not context.started
-    assert not ctx.stopped
+    assert not context.stopped
 
-    ctx.write(BEGIN, Bag(), END)
+    context.write(BEGIN, Bag(), END)
 
-    assert not ctx.alive
+    assert not context.alive
-    assert not ctx.started
+    assert not context.started
-    assert not ctx.stopped
+    assert not context.stopped
 
-    ctx.start()
+    context.start()
 
-    assert ctx.alive
+    assert context.alive
-    assert ctx.started
+    assert context.started
-    assert not ctx.stopped
+    assert not context.stopped
 
-    ctx.stop()
+    context.stop()
 
-    assert not ctx.alive
+    assert not context.alive
-    assert ctx.started
+    assert context.started
-    assert ctx.stopped
+    assert context.stopped