404 returned all the time when listing output batch part via REST-API

🤝 Community Help
1

404 status “{“error”:“Could not found requested batch.”}” is always return when listing the output batch parts via https://api.roboflow.com/data-staging/v1/external/{workspace}/batches/{batch_id}/parts :

Below is my curl commands : curl ‘https://api.roboflow.com/data-staging/v1/external/hostview/batches/test-1/parts?api_key=xxxxx’ where “test-1” is an existing batch id.

Also, is there a proper example on uploading and image, submitting a job and downloading the results both aggregated and individual files using the batch processing REST-API ? I find the current documentation has errors and not easy to follow.

Thx!

2

Hi @Clint_Suson ,

Script below works for me, could you try it out? It’s uploading a single image into the batch, and starts a batch processing job for it.

Thanks, Erik

import requests
import time

API_KEY = "MY_API_KEY"
WORKSPACE = "my-workspace"
WORKFLOW_ID = "my-workflow"
IMAGE_URL = "https://storage.googleapis.com/com-roboflow-marketing/docs/cars-highway.png"

BASE_STAGING = "https://api.roboflow.com/data-staging/v1/external"
BASE_JOBS = "https://api.roboflow.com/batch-processing/v1/external"

ts = int(time.time())
batch_id = f"my-batch-{ts}"
job_id = f"my-job-{ts}"

# 1. Upload an image to the batch
resp = requests.post(
    f"{BASE_STAGING}/{WORKSPACE}/batches/{batch_id}/upload/image",
    params={"api_key": API_KEY, "fileName": "cars-highway.png"},
    files={"file": ("image.png", requests.get(IMAGE_URL).content, "image/png")},
)
resp.raise_for_status()

time.sleep(3)

# 1.1 Check number of files in batch
resp = requests.get(f"{BASE_STAGING}/{WORKSPACE}/batches/{batch_id}/count", params={"api_key": API_KEY})
resp.raise_for_status()
print("Total files in batch: ", resp.json())

# 2. Create a job
resp = requests.post(
    f"{BASE_JOBS}/{WORKSPACE}/jobs/{job_id}",
    params={"api_key": API_KEY},
    json={
        "type": "simple-image-processing-v1",
        "jobInput": {
            "type": "staging-batch-input-v1",
            "batchId": batch_id,
        },
        "computeConfiguration": {
            "type": "compute-configuration-v2",
            "machineType": "gpu",
            "workersPerMachine": 4,
        },
        "processingSpecification": {
            "type": "workflows-processing-specification-v1",
            "workspace": WORKSPACE,
            "workflowId": WORKFLOW_ID,
            "aggregationFormat": "jsonl",
        },
        "processingTimeoutSeconds": 3600,
    },
)
resp.raise_for_status()

# 3. Wait for the job to complete
while True:
    resp = requests.get(f"{BASE_JOBS}/{WORKSPACE}/jobs/{job_id}", params={"api_key": API_KEY})
    resp.raise_for_status()
    data = resp.json()
    job = data.get("job", data)
    if job.get("isTerminal", False):
        break
    time.sleep(60)

# 4. Download results
notification = job.get("lastNotification", {})
results_batches = notification.get("resultsBatches", [])
for result_batch_id in results_batches:
    resp = requests.get(f"{BASE_STAGING}/{WORKSPACE}/batches/{result_batch_id}/parts", params={"api_key": API_KEY})
    parts = resp.json() if resp.ok else []
    if not parts:
        parts = [None]

    for part in parts:
        part_name = part if isinstance(part, str) else (part.get("partName") if isinstance(part, dict) else None)
        params = {"api_key": API_KEY}
        if part_name:
            params["partName"] = part_name
        resp = requests.get(f"{BASE_STAGING}/{WORKSPACE}/batches/{result_batch_id}/list", params=params)
        resp.raise_for_status()
        files = resp.json()
        items = files if isinstance(files, list) else files.get("filesMetadata", files.get("files", files.get("items", [])))
        for f in items:
            download_url = f.get("downloadURL") or f.get("url")
            name = f.get("fileName", "result")
            resp = requests.get(download_url)
            resp.raise_for_status()
            with open(name, "wb") as fh:
                fh.write(resp.content)
3

Hi Erik,

Thank you very much for this script, it worked for me. The issue is caused by the “batch-id” variable in the API is misleading. It should be named “outputBatches” or something similar to avoid confusion. Also, how do I enable the “Save Visualization” when submitting a job?

Thx!