Roboflow API - how to use filters | download list of images per project

🤝 Community Help
1

In general I have 5 projects under my private workspace. I want to download the list of files which were uploaded to Dataset and those in annotating process.

The problem is I can’t get Roboflow API to work.

For this example let’s use those VARs:

WORKSPACE=private-workspace-123
PROJECT=example-project
PROJECT_URL_1=https://app.roboflow.com/private-workspace-123/example-project
PROJECT_URL_2=app.roboflow.com/private-workspace-123/example-project
PROJECT_URL_3=private-workspace-123/example-project
PROJECT_URL_4=example-project

API_TOKEN=3j5hj534h5j

Based on documentation: Workspace Image Query | Roboflow Docs I tried to create my own query:

curl --location 'https://api.roboflow.com/private-workspace-123/search/v1?api_key= 3j5hj534h5j' --header 'Content-Type: application/json' --data '{"query": "project: PROJECT_URL_{no}", "pageSize": 100, "fields": ["filename", "tags"]}

Results in (I also tried other versions of that url like /browse or with http://):

Internal Server Error%                                                                                                                     

curl --location 'https://api.roboflow.com/private-workspace-123/search/v1?api_key= 3j5hj534h5j' --header 'Content-Type: application/json' --data '{"query": "project:example-project", "pageSize": 100, "fields": ["filename", "tags"]}

Results in always full list of images in all projects:

  ],
    "total": 10775,
    "continuationToken": "WzE3Mjg5NDEyODdd"
}%

And I cannot find any possible key added to fields that would return the name of project (so I could post process results for only project I’m interested in)

I found out also a second API:

curl --location 'https://api.roboflow.com/private-workspace-123/example-project/search?api_key= 3j5hj534h5j' --header 'Content-Type: application/json' --data '{"in_dataset": "true","pageSize": 500, "fields": ["name", "tags"]}'

it returns perfectly images from that project but ONLY 50 max and the response doesn’t have e.g. {"continuationToken": "WzE3Mjg5NDEyODdd"} so I cannot get to the second page.

In general, is it a bug or it’s a feature?

2

Hi, looking into it, will report back! :slight_smile:

3

Try removing the space between project: and PROJECT_URL_{no}

4

But I never had any space between key “project:project_url” in original requests.

Could you give me please example based on my real project URL?

curl --location 'https://api.roboflow.com/truebuilt-software-sxnvw/search/v1?api_key=3j5hj534h5j' --header 'Content-Type: application/json' --data '{"query": "project:tagsearch-detection", "pageSize": 100, "fields": ["filename", "tags"]}'

or

curl --location 'https://api.roboflow.com/truebuilt-software-sxnvw/search/v1?api_key=3j5hj534h5j' --header 'Content-Type: application/json' --data '{"query": "project:app.roboflow.com/truebuilt-software-sxnvw/tagsearch-detection", "pageSize": 100, "fields": ["filename", "tags"]}'

or

curl --location 'https://api.roboflow.com/truebuilt-software-sxnvw/search/v1?api_key=3j5hj534h5j' --header 'Content-Type: application/json' --data '{"query": "project:https://app.roboflow.com/truebuilt-software-sxnvw/tagsearch-detection", "pageSize": 100, "fields": ["filename", "tags"]}'

because I’m losing my mind and nothing is working. Always it return the full list of files