Stream job logs from current offset

This commit is contained in:
2026-06-30 15:51:22 +08:00
parent e766e4ed26
commit 4b3d750df9
5 changed files with 73 additions and 10 deletions

View File

@@ -63,7 +63,10 @@ are parsed into lightweight training curves.
Job APIs and the SSE log stream also expose structured progress parsed from
YOLO, MMSeg/MMEngine, SegModel-style epoch logs, and generic tqdm percentages,
so the queue and live log panel can show stage, epoch/iteration, and percent
without changing the original training scripts.
without changing the original training scripts. Starting any web job or
dataset YOLO shortcut automatically opens its live log; the SSE stream resumes
from the current log size after the initial tail so existing lines are not
duplicated in the panel.
The coverage panel calls `GET /api/coverage` and verifies that the user-facing
scripts from the existing `Seg/` workspace are mapped to web jobs. MMSeg

View File

@@ -59,6 +59,11 @@ def evaluate_project() -> dict:
and "setResults(resultsNext)" in frontend_text
and "slice(0, 240)" not in frontend_text,
"job_progress_ui": "JobProgressBar" in frontend_text and "progressTrack" in frontend_text,
"live_log_stream_ui": "EventSource" in frontend_text
and "eventSourceRef" in frontend_text
and "log_size" in frontend_text
and "events?offset=" in frontend_text,
"live_log_offset_api": "log_size" in backend_text and "offset: int = Query(0, ge=0)" in backend_text,
"runtime_readiness_ui": "runtimeReadiness" in frontend_text and "环境就绪" in frontend_text,
"capability_matrix_ui": "capabilities" in frontend_text and "全功能矩阵" in frontend_text,
"dataset_api": "/api/datasets" in backend_text and "api_upload_dataset_files" in backend_text,

View File

@@ -39,6 +39,8 @@ app.add_middleware(
def _job_with_progress(job: dict, include_log_tail: bool = False) -> dict:
enriched = dict(job)
max_bytes = 65536 if include_log_tail else 32768
log_path = Path(enriched["log_path"])
enriched["log_size"] = log_path.stat().st_size if log_path.exists() else 0
enriched["progress"] = progress_from_log_path(enriched["log_path"], enriched["status"], max_bytes=max_bytes)
if include_log_tail:
enriched["log_tail"] = db.log_tail(enriched["log_path"], max_bytes=max_bytes)
@@ -183,9 +185,9 @@ def api_cancel_job(job_id: str) -> dict:
@app.get("/api/jobs/{job_id}/events")
async def api_job_events(job_id: str):
async def api_job_events(job_id: str, offset: int = Query(0, ge=0)):
async def stream():
last_size = 0
last_size = offset
while True:
job = db.get_job(job_id)
if not job:
@@ -196,6 +198,8 @@ async def api_job_events(job_id: str):
chunk = ""
if path.exists():
size = path.stat().st_size
if last_size > size:
last_size = size
if size > last_size:
with path.open("rb") as handle:
handle.seek(last_size)

View File

@@ -1,6 +1,38 @@
from fastapi.testclient import TestClient
from app.jobs import default_conda_env_for_job
from app.main import _job_with_progress, app
def test_mmseg_jobs_use_mmseg_conda_env_by_default():
assert default_conda_env_for_job("mmseg.train") == "seg_mmcv"
assert default_conda_env_for_job("segmodel.train") == "seg_smp"
def test_job_progress_reports_log_size(tmp_path):
log_path = tmp_path / "job.log"
log_path.write_text("line one\nline two\n", encoding="utf-8")
job = {"id": "job1", "status": "running", "log_path": str(log_path)}
enriched = _job_with_progress(job, include_log_tail=True)
assert enriched["log_size"] == log_path.stat().st_size
assert enriched["log_tail"] == "line one\nline two\n"
def test_job_events_respect_log_offset(tmp_path, monkeypatch):
from app import main
log_path = tmp_path / "job.log"
old_chunk = "old chunk\n"
log_path.write_text(old_chunk + "new chunk\n", encoding="utf-8")
def fake_get_job(job_id):
return {"id": job_id, "status": "success", "log_path": str(log_path)}
monkeypatch.setattr(main.db, "get_job", fake_get_job)
response = TestClient(app).get(f"/api/jobs/job1/events?offset={len(old_chunk.encode('utf-8'))}")
assert response.status_code == 200
assert "new chunk" in response.text
assert "old chunk" not in response.text

View File

@@ -1,4 +1,4 @@
import React, { useEffect, useMemo, useState } from "react";
import React, { useEffect, useMemo, useRef, useState } from "react";
import { createRoot } from "react-dom/client";
import {
Activity,
@@ -45,6 +45,7 @@ type Job = {
started_at?: string;
finished_at?: string;
log_tail?: string;
log_size?: number;
params: Record<string, unknown>;
progress?: JobProgress;
};
@@ -430,6 +431,11 @@ function App() {
const [uploadFiles, setUploadFiles] = useState<FileList | null>(null);
const [agentValidation, setAgentValidation] = useState<ValidationAgentPayload | null>(null);
const [agentBusy, setAgentBusy] = useState(false);
const eventSourceRef = useRef<EventSource | null>(null);
useEffect(() => () => {
eventSourceRef.current?.close();
}, []);
const runningCount = jobs.filter((job) => job.status === "running").length;
const successCount = jobs.filter((job) => job.status === "success").length;
@@ -515,10 +521,11 @@ function App() {
async function createJob() {
setBusy(true);
try {
await api<Job>("/api/jobs", {
const job = await api<Job>("/api/jobs", {
method: "POST",
body: JSON.stringify({ type: taskType, params: JSON.parse(params) })
});
await inspectJob(job);
await refresh();
} finally {
setBusy(false);
@@ -633,13 +640,14 @@ function App() {
try {
const generated = await createSelectedYoloYaml();
if (!generated) return;
await api<Job>("/api/jobs", {
const job = await api<Job>("/api/jobs", {
method: "POST",
body: JSON.stringify({
type: "yolo.train_custom",
params: { model: "YOLO11n-seg", data: generated.path, epochs: 10, imgsz: 640, batch: 1, workers: 0, device: "cpu", project: "var/custom_yolo_runs", name: selectedDataset.name, exist_ok: true }
})
});
await inspectJob(job);
window.location.hash = "jobs";
await refresh();
} finally {
@@ -651,13 +659,14 @@ function App() {
if (!selectedDataset?.absolute_layout) return;
setBusy(true);
try {
await api<Job>("/api/jobs", {
const job = await api<Job>("/api/jobs", {
method: "POST",
body: JSON.stringify({
type: "yolo.predict_custom",
params: { weights: customYoloWeightPath(selectedDataset), source: selectedDataset.absolute_layout.images, imgsz: 640, conf: 0.25, device: "cpu", project: "var/custom_yolo_runs", name: `${selectedDataset.name}_predict`, exist_ok: true }
})
});
await inspectJob(job);
window.location.hash = "jobs";
await refresh();
} finally {
@@ -669,13 +678,14 @@ function App() {
if (!selectedDataset?.absolute_layout) return;
setBusy(true);
try {
await api<Job>("/api/jobs", {
const job = await api<Job>("/api/jobs", {
method: "POST",
body: JSON.stringify({
type: "yolo.heatmap_custom",
params: { weights: customYoloWeightPath(selectedDataset), source: selectedDataset.absolute_layout.images, model_key: "YOLO11n-seg", cam_method: "GradCAM", target_layers: "model.model.model[9]", limit: 3, project: "var/custom_yolo_runs", name: `${selectedDataset.name}_heatmap` }
})
});
await inspectJob(job);
window.location.hash = "jobs";
await refresh();
} finally {
@@ -684,15 +694,24 @@ function App() {
}
async function inspectJob(job: Job) {
eventSourceRef.current?.close();
const detail = await api<Job>(`/api/jobs/${job.id}`);
setSelectedJob(detail);
setLog(detail.log_tail ?? "");
const source = new EventSource(`${API_BASE}/api/jobs/${job.id}/events`);
const source = new EventSource(`${API_BASE}/api/jobs/${job.id}/events?offset=${detail.log_size ?? 0}`);
eventSourceRef.current = source;
source.onmessage = (event) => {
const payload = JSON.parse(event.data);
if (payload.chunk) setLog((prev) => `${prev}${payload.chunk}`);
setSelectedJob(payload.job);
if (["success", "failed", "cancelled"].includes(payload.job.status)) source.close();
if (["success", "failed", "cancelled"].includes(payload.job.status)) {
source.close();
if (eventSourceRef.current === source) eventSourceRef.current = null;
}
};
source.onerror = () => {
source.close();
if (eventSourceRef.current === source) eventSourceRef.current = null;
};
}