Skip to content
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

Allow setting min_time in Python API #635

Merged
merged 3 commits into from
Nov 2, 2023
Merged

Allow setting min_time in Python API #635

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
f875876
Add task info JSON output
Kobzol Oct 31, 2023
c2644cc
Allow setting `min_time` resource request in Python API
Kobzol Oct 31, 2023
7811f1f
Update changelog
Kobzol Oct 31, 2023
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
9 changes: 9 additions & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,12 @@
# Dev

## Breaking change
* The output format of the `job info` command with JSON output mode has been changed. Note that
the JSON output mode is still unstable.

## New features
* Allow setting minimum duration for a task (`min_time` resource value) using the Python API.

# v0.17.0

## Breaking change
Expand Down
203 changes: 118 additions & 85 deletions crates/hyperqueue/src/client/output/json.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,90 +80,63 @@ impl Output for JsonOutput {
self.print(json!(statuses))
}
fn print_job_detail(&self, jobs: Vec<JobDetail>, _worker_map: WorkerMap, server_uid: &str) {
let job_details: Vec<_> = jobs.into_iter().map(|job| {
let task_paths = resolve_task_paths(&job, server_uid);

let JobDetail {
info,
job_desc,
tasks,
tasks_not_found: _,
max_fails,
submission_date,
completion_date_or_now,
submit_dir,
} = job;

let finished_at = if info.counters.is_terminated(info.n_tasks) {
Some(completion_date_or_now)
} else {
None
};

let mut json = json!({
"info": format_job_info(info),
"max_fails": max_fails,
"started_at": format_datetime(submission_date),
"finished_at": finished_at.map(format_datetime),
"submit_dir": submit_dir
});

if let JobDescription::Array {
task_desc:
TaskDescription {
program:
ProgramDefinition {
args,
env,
stdout,
stderr,
cwd,
stdin: _,
},
resources,
pin_mode,
time_limit,
priority,
task_dir,
crash_limit,
},
..
} = job_desc
{
json["program"] = json!({
"args": args.into_iter().map(|args| args.to_string()).collect::<Vec<_>>(),
"env": env.into_iter().map(|(key, value)| (key.to_string(), value.to_string())).collect::<Map<String, String>>(),
"cwd": cwd,
"stderr": format_stdio_def(&stderr),
"stdout": format_stdio_def(&stdout),
});
json["resources"] = resources.variants.into_iter().map(|v| {
let ResourceRequest { n_nodes, resources, min_time } = v;
json!({
"n_nodes": n_nodes,
"resources": resources.into_iter().map(|res| {
json!({
"resource": res.resource,
"request": res.policy
})
}).collect::<Vec<_>>(),
"min_time": format_duration(min_time)
})}
).collect();
json["pin_mode"] = json!(match pin_mode {
PinMode::None => None,
PinMode::OpenMP => Some("openmp"),
PinMode::TaskSet => Some("taskset"),
let job_details: Vec<_> = jobs
.into_iter()
.map(|job| {
let task_paths = resolve_task_paths(&job, server_uid);

let JobDetail {
info,
job_desc,
tasks,
tasks_not_found: _,
max_fails,
submission_date,
completion_date_or_now,
submit_dir,
} = job;

let finished_at = if info.counters.is_terminated(info.n_tasks) {
Some(completion_date_or_now)
} else {
None
};

let mut json = json!({
"info": format_job_info(info),
"max_fails": max_fails,
"started_at": format_datetime(submission_date),
"finished_at": finished_at.map(format_datetime),
"submit_dir": submit_dir
});
json["priority"] = json!(priority);
json["time_limit"] = json!(time_limit.map(format_duration));
json["task_dir"] = json!(task_dir);
json["crash_limit"] = json!(crash_limit);
}

json["tasks"] = format_tasks(tasks, task_paths);
json
}).collect();
let task_description = match job_desc {
JobDescription::Array { task_desc, .. } => {
let mut task = json!({});
format_task_description(task_desc, &mut task);
json!({
"array": task
})
}
JobDescription::Graph { tasks } => {
let tasks: Vec<Value> = tasks
.into_iter()
.map(|task| {
let mut json = json!({});
format_task_description(task.task_desc, &mut json);
json
})
.collect();
json!({
"graph": tasks
})
}
};
json["task-desc"] = task_description;
json["tasks"] = format_tasks(tasks, task_paths);
json
})
.collect();
self.print(Value::Array(job_details));
}

Expand Down Expand Up @@ -216,12 +189,14 @@ impl Output for JsonOutput {

fn print_task_info(
&self,
_job: (JobId, JobDetail),
_tasks: Vec<JobTaskInfo>,
job: (JobId, JobDetail),
tasks: Vec<JobTaskInfo>,
_worker_map: WorkerMap,
_server_uid: &str,
server_uid: &str,
_verbosity: Verbosity,
) {
let map = resolve_task_paths(&job.1, server_uid);
self.print(format_tasks(tasks, map));
}

fn print_summary(&self, filename: &Path, summary: Summary) {
Expand Down Expand Up @@ -257,6 +232,64 @@ impl Output for JsonOutput {
}
}

fn format_task_description(task_desc: TaskDescription, json: &mut Value) {
let TaskDescription {
program:
ProgramDefinition {
args,
env,
stdout,
stderr,
cwd,
stdin: _,
},
resources,
pin_mode,
time_limit,
priority,
task_dir,
crash_limit,
} = task_desc;

json["program"] = json!({
"args": args.into_iter().map(|args| args.to_string()).collect::<Vec<_>>(),
"env": env.into_iter().map(|(key, value)| (key.to_string(), value.to_string())).collect::<Map<String, String>>(),
"cwd": cwd,
"stderr": format_stdio_def(&stderr),
"stdout": format_stdio_def(&stdout),
});
json["resources"] = resources
.variants
.into_iter()
.map(|v| {
let ResourceRequest {
n_nodes,
resources,
min_time,
} = v;
json!({
"n_nodes": n_nodes,
"resources": resources.into_iter().map(|res| {
json!({
"resource": res.resource,
"request": res.policy
})
}).collect::<Vec<_>>(),
"min_time": format_duration(min_time)
})
})
.collect();
json["pin_mode"] = json!(match pin_mode {
PinMode::None => None,
PinMode::OpenMP => Some("openmp"),
PinMode::TaskSet => Some("taskset"),
});
json["priority"] = json!(priority);
json["time_limit"] = json!(time_limit.map(format_duration));
json["task_dir"] = json!(task_dir);
json["crash_limit"] = json!(crash_limit);
}

fn fill_task_started_data(dict: &mut Value, data: StartedTaskData) {
dict["started_at"] = format_datetime(data.start_date);
if data.worker_ids.len() == 1 {
Expand Down
6 changes: 5 additions & 1 deletion crates/pyhq/python/hyperqueue/ffi/protocol.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,5 @@
import dataclasses
import datetime
from typing import Dict, List, Optional, Sequence, Union

from ..output import StdioDef
Expand All @@ -7,22 +8,25 @@
class ResourceRequest:
n_nodes: int = 0
resources: Dict[str, Union[int, float, str]] = dataclasses.field(default_factory=dict)
min_time: Optional[float] = None

def __init__(
self,
*,
n_nodes=0,
cpus: Union[int, float, str] = 1,
resources: Optional[Dict[str, Union[int, float, str]]] = None,
min_time: Optional[datetime.timedelta] = None,
):
self.n_nodes = n_nodes
if resources is None:
resources = {}
resources["cpus"] = cpus
self.resources = resources
self.min_time = min_time.total_seconds() if min_time is not None else None

def __repr__(self):
return f"<ResourceRequest n_nodes={self.n_nodes} resources={self.resources}>"
return f"<ResourceRequest n_nodes={self.n_nodes} resources={self.resources} min_time={self.min_time}>"


@dataclasses.dataclass()
Expand Down
5 changes: 3 additions & 2 deletions crates/pyhq/src/client/job.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@ use tako::gateway::{ResourceRequestEntries, ResourceRequestEntry, ResourceReques
use tako::program::{FileOnCloseBehavior, ProgramDefinition, StdioDef};
use tako::resources::{AllocationRequest, NumOfNodes, ResourceAmount};

use crate::marshal::FromPy;
use crate::marshal::{FromPy, WrappedDuration};
use crate::utils::error::ToPyResult;
use crate::{borrow_mut, run_future, ClientContextPtr, FromPyObject, PyJobId, PyTaskId};

Expand All @@ -38,6 +38,7 @@ enum AllocationValue {
pub struct ResourceRequestDescription {
n_nodes: NumOfNodes,
resources: HashMap<String, AllocationValue>,
min_time: Option<WrappedDuration>,
}

#[derive(Debug, FromPyObject)]
Expand Down Expand Up @@ -182,7 +183,7 @@ fn build_task_desc(desc: TaskDescription, submit_dir: &Path) -> anyhow::Result<H
})
})
.collect::<anyhow::Result<ResourceRequestEntries>>()?,
min_time: Default::default(),
min_time: rq.min_time.map(|v| v.into()).unwrap_or_default(),
})
})
.collect::<anyhow::Result<_>>()?,
Expand Down
6 changes: 6 additions & 0 deletions crates/pyhq/src/marshal.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,3 +59,9 @@ impl<'a> FromPyObject<'a> for WrappedDuration {
Ok(WrappedDuration(Duration::from_nanos(nanoseconds as u64)))
}
}

impl From<WrappedDuration> for Duration {
fn from(value: WrappedDuration) -> Self {
value.0
}
}
Loading
Loading