Can't upload images with DataRequest method

[quaid281]

I'm attempting to upload images to my task. Why would this code '''

Create a TasksApi instance

    tasks_api = TasksApi(api_client=api_client)

    # Get the task
    task = tasks_api.retrieve(task_id)
    # Upload the images
    for image_path in image_paths:
        image_name = os.path.basename(image_path)
        if image_name.endswith(".jpg") or image_name.endswith(".png"):
            with open(os.path.join(data_path, image_name), "rb") as f:
                file_bytes = f.read()
            data_request = DataRequest(name=image_name,image_quality=75,chunk_size=0,image_data=file_bytes)
            tasks_api.create_data(id=task_id, data_request=data_request)
            name = task[0]['name']
        print(f'Uploaded {image_name} to task {name} ({task_id})')

''' Give this error: File "C:\Users\AXW0J9E\Downloads\FPLAir\automaticProjectCreationv5.py", line 55, in create_cvat_task data_request = DataRequest(name=image_name,image_quality=75,chunk_size=0,image_data=file_bytes) File "C:\Users\AXW0J9E\AppData\Local\miniconda3\envs\qaqc\lib\site-packages\cvat_sdk\api_client\model_utils.py", line 1465, in attempt_convert_item raise get_type_error(input_value, path_to_item, valid_classes, cvat_sdk.api_client.exceptions.ApiTypeError: Invalid type for variable 'image_data'. Required value type is one of [NoneType, bool, date, datetime, dict, float, int, list, str] and passed type was bytes at ['image_data']

I'm lost on how I could fix this. I could not include in the file_bytes as a parameter in DataRequest and use this instead: ''' data_request = DataRequest(name=image_name,image_quality=75) tasks_api.create_data(id=task_id, data_request=data_request) ''' However, by using the code doesn't crash but no images are uploaded in CVAT.

[zhiltsov-max]

Hi. Please use this code snippet for file uploads with the lower-level SDK API. There are some extra efforts needed to pass files via this API. Please also consider using higher level API, which makes this process much more simple.

[quaid281]

That's not easy for me to read and understand. Is there another way, or is there an example of how to use this snippet with a image_path from a local directory?

[zhiltsov-max]

Please check the documentation here.

If you need to pass other parameters than just files and image_quality, consider this sample:

```python from http import HTTPStatus from time import sleep from cvat_sdk import models from cvat_sdk.api_client import ApiClient, Configuration, apis def wait_until_task_is_created(api: apis.TasksApi, task_id: int) -> models.RqStatus: for _ in range(100): (status, _) = api.retrieve_status(task_id) if status.state.value in ["Finished", "Failed"]: return status sleep(1) raise Exception("Cannot create task") def create_task(api_client: ApiClient, spec: dict, data: dict, **kwargs) -> int: (task, response) = api_client.tasks_api.create(spec, **kwargs) assert response.status == HTTPStatus.CREATED if data.get("client_files"): # Can't encode binary files in json, use multipart/form-data instead (_, response) = api_client.tasks_api.create_data( task.id, data_request=models.DataRequest( client_files=data["client_files"], image_quality=data["image_quality"], ), upload_start=True, upload_multiple=True, _content_type="multipart/form-data", **kwargs, ) assert response.status == HTTPStatus.OK data = data.copy() del data["client_files"] # Pass the remaining parameters in this request and finish uploading (_, response) = api_client.tasks_api.create_data( task.id, data_request=models.DataRequest(**data), upload_finish=True, **kwargs ) assert response.status == HTTPStatus.ACCEPTED status = wait_until_task_is_created(api_client.tasks_api, task.id) assert status.state.value == "Finished" return task.id # check the docs about auth methods: # https://opencv.github.io/cvat/docs/api_sdk/sdk/lowlevel-api/#apiclient-and-configuration with ApiClient(Configuration(host="localhost", username="user", password="password")) as api_client: task_id = create_task(api_client, spec={ "name": "task with data", "labels": [{ "name": "cat" }] }, data={ "client_files": [ open('image1.jpg', 'rb') ], "image_quality": 70, "sorting_method": "lexicographical" } ) ```

Is there another way

Use the high-level SDK API:

from cvat_sdk import make_client
from cvat_sdk.core.proxies.tasks import ResourceType

with make_client("localhost", credentials=("user", "password")) as client:
    task = client.tasks.create_from_data(
        spec={
            "name": "task with local data",
            "labels": [ { "name": "car" } ],
        },
        data_params={
            "image_quality": 75,
        },
        resource_type=ResourceType.LOCAL,
        resources=["image.jpg"],
    )

[quaid281]

@zhiltsov-max wow man ...that worked perfectly. Thank you so much for the example

[alijawad07]

I am trying to upload the images by creating a task in an already existing project inside an organization, but this does not work. I am able to create the task without any media but when i try to attach media with it, im unable to do so.

[zhiltsov-max]

@alijawad07, do you receive any error? When working in an organization, please add client.organization_slug = "orgslug" before creating the task.