Skip to content

feat(executors): Add support for grpc remote execution via CLI #526

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
AndreyMarkinPPC wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

feat(executors): Add support for grpc remote execution via CLI #526

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 11 additions & 7 deletions libs/executors/garf/executors/entrypoints/grpc_server.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,8 +45,11 @@
from opentelemetry.instrumentation.grpc import GrpcInstrumentorServer
from opentelemetry.instrumentation.logging import LoggingInstrumentor

OTEL_SERVICE_NAME = 'garf'
OTEL_SERVICE_NAME = 'garf-grpc'
LoggingInstrumentor().instrument(set_logging_format=False)
grpc_server_instrumentor = GrpcInstrumentorServer()
grpc_server_instrumentor.instrument()


server_start_time = time.time()

Expand Down Expand Up @@ -174,17 +177,18 @@ def ListExecutors(self, request, context):
'--parallel-threshold', dest='parallel_threshold', default=10, type=int
)
args, _ = parser.parse_known_args()
initialize_tracer()
meter = initialize_meter()
otel_service_name = os.getenv('OTEL_SERVICE_NAME', OTEL_SERVICE_NAME)
initialize_tracer(otel_service_name)
meter = initialize_meter(otel_service_name)
logger = utils.init_logging(
loglevel='INFO', logger_type='local', name=OTEL_SERVICE_NAME
loglevel='INFO',
logger_type='local',
name=otel_service_name,
)
logger.addHandler(initialize_logger())

grpc_server_instrumentor = GrpcInstrumentorServer()
grpc_server_instrumentor.instrument()
server = grpc.server(
futures.ThreadPoolExecutor(max_workers=args.parallel_threshold)
futures.ThreadPoolExecutor(max_workers=args.parallel_threshold),
)

service = GarfService()
Expand Down
214 changes: 161 additions & 53 deletions libs/executors/garf/executors/entrypoints/typer_cli.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,10 +21,12 @@
from typing import Optional

import garf.executors
import garf.executors.garf_pb2 as pb
import grpc
import requests
import typer
from garf.core import cache
from garf.executors import exceptions, setup
from garf.executors import exceptions, garf_pb2_grpc, setup
from garf.executors.config import Config
from garf.executors.entrypoints import utils
from garf.executors.entrypoints.tracer import (
Expand All @@ -34,7 +36,9 @@
from garf.executors.telemetry import tracer
from garf.executors.workflows import workflow, workflow_runner
from garf.io import reader, writer
from google.protobuf.json_format import ParseDict
from opentelemetry import trace
from opentelemetry.instrumentation.grpc import GrpcInstrumentorClient
from opentelemetry.instrumentation.logging import LoggingInstrumentor
from opentelemetry.trace.propagation.tracecontext import (
TraceContextTextMapPropagator,
Expand All @@ -44,6 +48,8 @@
from typing_extensions import Annotated

LoggingInstrumentor().instrument(set_logging_format=False)
client_instrumentor = GrpcInstrumentorClient()
client_instrumentor.instrument()
console = Console()


Expand All @@ -52,6 +58,10 @@
((f, f) for f in sorted(setup.find_executors())),
)

ServerTypeEnum = enum.Enum(
'ServerTypeEnum', (('http', 'http'), ('grpc', 'grpc'))
)

initialize_tracer()
typer_app = typer.Typer(
help='Garf\n\nCall APIs with SQL in your terminal', rich_markup_mode='rich'
Expand Down Expand Up @@ -154,16 +164,21 @@ def prune(time_days: int):
]


def _init_runner(
file, context=None, config=None
) -> workflow_runner.WorkflowRunner:
wf_parent = pathlib.Path.cwd() / pathlib.Path(file).parent
def _init_workflow(file, context=None, config=None) -> workflow.Workflow:
execution_workflow = workflow.Workflow.from_file(
path=file, context=context, config_file=config
)
garf.executors.version.validate_version(
execution_workflow.metadata.required_garf_version
)
return execution_workflow


def _init_runner(
file, context=None, config=None
) -> workflow_runner.WorkflowRunner:
execution_workflow = _init_workflow(file, context, config)
wf_parent = pathlib.Path.cwd() / pathlib.Path(file).parent
return workflow_runner.WorkflowRunner(
execution_workflow=execution_workflow, wf_parent=wf_parent
)
Expand Down Expand Up @@ -203,6 +218,9 @@ def execute(
server_url: Annotated[
str | None, typer.Option(help='Address of garf server in HOST:PORT format')
] = None,
server_type: Annotated[
ServerTypeEnum, typer.Option(help='Type of server')
] = 'http',
additional_output: Annotated[
Optional[str],
typer.Option(
Expand Down Expand Up @@ -275,46 +293,10 @@ def execute(
with tracer.start_as_current_span('read_queries'):
batch = {query: reader_client.read(query) for query in found_queries}
if server_url:
rest_context = context.model_dump()
if rest_context.get('writer') == ['console']:
del rest_context['writer']
headers = {}
TraceContextTextMapPropagator().inject(headers)
if parallel_queries and len(batch) > 1:
endpoint = f'{server_url}/api/execute:batch'
request = {
'batch': batch,
'source': source,
'context': rest_context,
}
try:
response = requests.post(url=endpoint, json=request, headers=headers)
response.raise_for_status()
except requests.exceptions.HTTPError as e:
raise exceptions.GarfExecutorError(
f'Server error: {e.response.status_code} - {e.response.json()}'
) from e

with tracer.start_as_current_span('parse_results batch'):
typer.secho(response.json().get('results'))
else:
endpoint = f'{server_url}/api/execute'
for title, text in batch.items():
request = {
'source': source,
'title': title,
'query': text,
'context': rest_context,
}
try:
response = requests.post(url=endpoint, json=request, headers=headers)
response.raise_for_status()
except requests.exceptions.HTTPError as e:
raise exceptions.GarfExecutorError(
f'Server error: {e.response.status_code} - {e.response.json()}'
) from e
with tracer.start_as_current_span(f'parse_results {title}'):
typer.secho(response.json().get('results'))
if server_type == ServerTypeEnum.http:
_send_http(context, parallel_queries, batch, server_url, source)
elif server_type == ServerTypeEnum.grpc:
_send_grpc(context, parallel_queries, batch, server_url, source)
else:
query_executor = setup.setup_executor(
source=source,
Expand Down Expand Up @@ -363,6 +345,12 @@ def run(
enable_cache: EnableCache = False,
cache_ttl_seconds: CacheTTL = 3600,
simulate: Simulate = False,
server_url: Annotated[
str | None, typer.Option(help='Address of garf server in HOST:PORT format')
] = None,
server_type: Annotated[
ServerTypeEnum, typer.Option(help='Type of server')
] = 'http',
):
"""Runs workflow from a file."""
span = trace.get_current_span()
Expand All @@ -373,14 +361,21 @@ def run(
)
garf_logger.addHandler(initialize_logger())
context = utils.ParamsParser().parse_all(ctx.args)
runner = _init_runner(file, context, config)
runner.run(
enable_cache=enable_cache,
cache_ttl_seconds=cache_ttl_seconds,
selected_aliases=include,
skipped_aliases=exclude,
simulate=simulate,
)
if server_url:
if server_type == ServerTypeEnum.http:
_run_worklow_http(server_url, file)
elif server_type == ServerTypeEnum.grpc:
execution_workflow = _init_workflow(file, context, config)
_run_worklow_grpc(server_url, execution_workflow)
else:
runner = _init_runner(file, context, config)
runner.run(
enable_cache=enable_cache,
cache_ttl_seconds=cache_ttl_seconds,
selected_aliases=include,
skipped_aliases=exclude,
simulate=simulate,
)


@workflow_app.command(
Expand Down Expand Up @@ -457,6 +452,119 @@ def version() -> str:
raise typer.Exit()


def _send_grpc(context, parallel_queries, batch, server_url, source):
channel = grpc.insecure_channel(server_url)
stub = garf_pb2_grpc.GarfServiceStub(channel)
rest_context = context.model_dump()
if rest_context.get('writer') == ['console']:
del rest_context['writer']
else:
rest_context['writer'] = rest_context['writer'][0]
if parallel_queries and len(batch) > 1:
batch = [
pb.QueryDefinition(title=title, text=text)
for title, text in batch.items()
]
request = pb.ExecuteBatchRequest(
source=source, batch=batch, context=pb.ExecutionContext(**rest_context)
)
response = stub.ExecuteBatch(request)
with tracer.start_as_current_span('parse_results batch'):
typer.secho(response.results)
else:
for title, text in batch.items():
request = pb.ExecuteRequest(
source=source,
title=title,
query=text,
context=pb.ExecutionContext(**rest_context),
)
response = stub.Execute(request)
with tracer.start_as_current_span('parse_results'):
typer.secho(response.results)


def _send_http(context, parallel_queries, batch, server_url, source):
rest_context = context.model_dump()
if rest_context.get('writer') == ['console']:
del rest_context['writer']
headers = {}
TraceContextTextMapPropagator().inject(headers)
if parallel_queries and len(batch) > 1:
endpoint = f'{server_url}/api/execute:batch'
request = {
'batch': batch,
'source': source,
'context': rest_context,
}
try:
response = requests.post(url=endpoint, json=request, headers=headers)
response.raise_for_status()
except requests.exceptions.HTTPError as e:
raise exceptions.GarfExecutorError(
f'Server error: {e.response.status_code} - {e.response.json()}'
) from e

with tracer.start_as_current_span('parse_results batch'):
typer.secho(response.json().get('results'))
else:
endpoint = f'{server_url}/api/execute'
for title, text in batch.items():
request = {
'source': source,
'title': title,
'query': text,
'context': rest_context,
}
try:
response = requests.post(url=endpoint, json=request, headers=headers)
response.raise_for_status()
except requests.exceptions.HTTPError as e:
raise exceptions.GarfExecutorError(
f'Server error: {e.response.status_code} - {e.response.json()}'
) from e
with tracer.start_as_current_span(f'parse_results {title}'):
typer.secho(response.json().get('results'))


def _run_worklow_grpc(server_url, workflow_data):
workflow_data.compile()

channel = grpc.insecure_channel(server_url)
stub = garf_pb2_grpc.GarfServiceStub(channel)
workflow_pb = pb.Workflow()
workflow_dict = workflow_data.model_dump()
config_pb = pb.Config()
if config := workflow_dict.pop('execution_config'):
ParseDict(config, config_pb)
context_pb = pb.ExecutionContext()
if context := workflow_dict.pop('context'):
ParseDict(context, context_pb)
ParseDict(workflow_dict, workflow_pb, ignore_unknown_fields=True)
request = pb.ExecuteWorkflowRequest(
workflow=workflow_pb,
config=config_pb,
context=context_pb,
)
response = stub.ExecuteWorkflow(request)
with tracer.start_as_current_span('parse_results workflow'):
typer.secho(response.results)


def _run_worklow_http(server_url, workflow_file):
headers = {}
TraceContextTextMapPropagator().inject(headers)
endpoint = f'{server_url}/api/execute:workflow'
with open(workflow_file, mode='rb') as f:
files = {'workflow_file': f}
response = requests.post(
url=endpoint,
files=files,
headers=headers,
)
response.raise_for_status()


@typer_app.callback(
invoke_without_command=True,
context_settings={'allow_extra_args': True, 'ignore_unknown_options': True},
Expand Down
Loading
Loading