async def test_condition_does_not_run_auto_start(self):
ran = []
@op(conditions=[CONDITION])
async def condition_test():
ran.append(True) # pragma: no cover
async with MemoryOrchestrator() as orchestrator:
async with orchestrator(DataFlow(condition_test)) as octx:
async for _ in octx.run([]):
pass
self.assertFalse(ran)
async def test_condition_does_not_run(self):
ran = []
@op(conditions=[CONDITION])
async def condition_test(hi: str):
ran.append(True)
async with MemoryOrchestrator() as orchestrator:
async with orchestrator(DataFlow(condition_test)) as octx:
async for _ in octx.run([
Input(
value=True,
definition=condition_test.op.inputs["hi"],
),
]):
pass
self.assertFalse(ran)
async def _multicomm_dataflow(self, config, request):
# Seed the network with inputs given by caller
# TODO(p0,security) allowlist of valid definitions to seed (set
# Input.origin to something other than seed)
inputs = []
# If data was sent add those inputs
if request.method == "POST":
# Accept a list of input data according to config.input_mode
if config.input_mode == "default":
# TODO validate that input data is dict of list of inputs each item
# has definition and value properties
for ctx, client_inputs in (await request.json()).items():
for input_data in client_inputs:
if (not input_data["definition"]
in config.dataflow.definitions):
return web.json_response(
{
"error":
f"Missing definition for {input_data['definition']} in dataflow"
},
status=HTTPStatus.NOT_FOUND,
)
inputs.append(
MemoryInputSet(
MemoryInputSetConfig(
ctx=StringInputSetContext(ctx),
inputs=[
Input(
value=input_data["value"],
definition=config.dataflow.definitions[
input_data["definition"]],
) for input_data in client_inputs
] + ([
Input(
value=request.headers,
definition=config.dataflow.definitions[
config.forward_headers],
)
] if config.forward_headers else []),
)))
elif ":" in config.input_mode:
preprocess_mode, *input_def = config.input_mode.split(":")
input_def = ":".join(input_def)
if input_def not in config.dataflow.definitions:
return web.json_response(
{
"error":
f"Missing definition for {input_def} in dataflow"
},
status=HTTPStatus.NOT_FOUND,
)
if preprocess_mode == "json":
value = await request.json()
elif preprocess_mode == "text":
value = await request.text()
elif preprocess_mode == "bytes":
value = await request.read()
elif preprocess_mode == "stream":
value = request.content
else:
return web.json_response(
{
"error":
f"preprocess tag must be one of {self.IO_MODES}, got {preprocess_mode}"
},
status=HTTPStatus.NOT_FOUND,
)
inputs.append(
MemoryInputSet(
MemoryInputSetConfig(
ctx=StringInputSetContext("post_input"),
inputs=[
Input(
value=value,
definition=config.dataflow.
definitions[input_def],
)
] + ([
Input(
value=request.headers,
definition=config.dataflow.definitions[
config.forward_headers],
)
] if config.forward_headers else []),
)))
else:
raise NotImplementedError(
"Input modes other than default,preprocess:definition_name not yet implemented"
)
# Run the operation in an orchestrator
# TODO(dfass) Create the orchestrator on startup of the HTTP API itself
async with MemoryOrchestrator() as orchestrator:
# TODO(dfass) Create octx on dataflow registration
async with orchestrator(config.dataflow) as octx:
results = {
str(ctx): result
async for ctx, result in octx.run(*inputs)
}
if config.output_mode == "json":
return web.json_response(results)
# content_info is a List[str] ([content_type,output_keys])
# in case of stream,bytes and string in others
postprocess_mode, *content_info = config.output_mode.split(":")
if postprocess_mode == "stream":
# stream:text/plain:get_single.beef
raise NotImplementedError(
"output mode not yet implemented")
elif postprocess_mode == "bytes":
content_type, output_keys = content_info
output_data = traverse_get(results, output_keys)
return web.Response(body=output_data)
elif postprocess_mode == "text":
output_data = traverse_get(results, content_info[0])
return web.Response(text=output_data)
else:
return web.json_response(
{"error": f"output mode not valid"},
status=HTTPStatus.NOT_FOUND,
)
async def test_run(self):
dataflow = DataFlow.auto(*OPIMPS)
# Instantiate inputs
repos = glob.glob(
os.path.join(
os.path.expanduser("~"),
"Documents",
"python",
"testrepos",
"*",
))
if not repos:
repos = glob.glob(
os.path.join(os.path.expanduser("~"), "Documents", "python",
"dffml"))
if not repos:
repos = [
"https://github.com/intel/dffml",
"https://github.com/pdxjohnny/dffml",
]
repos = repos[:2]
urls = [
Input(value=URL,
definition=dataflow.definitions["URL"],
parents=None) for URL in repos
]
no_git_branch_given = Input(
value=True,
definition=dataflow.definitions["no_git_branch_given"],
parents=None,
)
date_spec = Input(
value=datetime.now().strftime(TIME_FORMAT_MINTUE_RESOLUTION),
definition=dataflow.definitions["quarter_start_date"],
parents=None,
)
quarters = [
Input(
value=i,
definition=dataflow.definitions["quarter"],
parents=None,
) for i in range(0, 10)
]
group_by_spec = Input(
value={
"cloc": {
"group": "quarter",
"by": "language_to_comment_ratio",
"fill": 0,
},
"authors": {
"group": "quarter",
"by": "author_count",
"fill": 0,
},
"work": {
"group": "quarter",
"by": "work_spread",
"fill": 0
},
"release": {
"group": "quarter",
"by": "release_within_period",
"fill": False,
},
"commits": {
"group": "quarter",
"by": "commit_count",
"fill": 0,
},
},
definition=dataflow.definitions["group_by_spec"],
parents=None,
)
# Orchestrate the running of these operations
async with MemoryOrchestrator() as orchestrator:
async with orchestrator(dataflow) as octx:
# Add our inputs to the input network with the context being the URL
async for ctx, results in octx.run({
url.value: [
url,
no_git_branch_given,
date_spec,
group_by_spec,
*quarters,
]
for url in urls
}):
self.assertTrue(results)