[PTDT-1867] Add llm data gen label types

[lgluszek]

It should be possible to run the code below to import pre-labels and ground truth

Prompt creation

import sys
import labelbox as lb
import labelbox.types as lb_types
import uuid

API_KEY=""
client = lb.Client(API_KEY, "http://localhost:8080/_gql")

prompt_annotation = lb_types.ClassificationAnnotation(
    name="Provide a prompt",
    value=lb_types.Text(answer="This is a MAL prompt 2"),
)


label = []
label.append(
  lb_types.Label(
    data=lb_types.LlmPromptCreationData(
        global_key="af3d406e-155f-451b-b828-a6f4744f2e2b"
    ),
    annotations=[
      prompt_annotation,
    ]
  )
)
  
upload_job_mal = lb.MALPredictionImport.create_from_objects(
    client = client, 
    project_id = "clptvgpf70193v2vu3xb02grt", 
    name="mal_import_job_llm_data_gen"+str(uuid.uuid4()), 
    predictions=label)
  
upload_job_mal.wait_until_done()

upload_job = lb.LabelImport.create_from_objects(
    client = client, 
    project_id = "clptvgpf70193v2vu3xb02grt", 
    name="label_import_job"+str(uuid.uuid4()),  
    labels=label)

upload_job.wait_until_done()

Prompt response creation

import sys
import labelbox as lb
import labelbox.types as lb_types
import uuid

API_KEY=""
client = lb.Client(API_KEY, "http://localhost:8080/_gql")


prompt_annotation = lb_types.ClassificationAnnotation(
    name="Prompt",
    value=lb_types.Text(answer="This is a MAL prompt"),
)

response_radio_annotation = lb_types.ClassificationAnnotation(
    name="Response Radio",
    value=lb_types.Radio(answer=lb_types.ClassificationAnswer(
        name="A",
        classifications=[
        lb_types.ClassificationAnnotation(
          name="Nested radio",
          value=lb_types.Radio(
            answer=lb_types.ClassificationAnswer(
              name="2"
            )
          )
        )
      ])))

response_checklist_annotation= lb_types.ClassificationAnnotation(
  name="Response Checklist", # must match your ontology feature's name
  value=lb_types.Checklist(
      answer = [
        lb_types.ClassificationAnswer(
        name="A",
        classifications=[
          lb_types.ClassificationAnnotation(
            name="Nested Checklist",
            value=lb_types.Checklist(
              answer=[
                  lb_types.ClassificationAnswer(
                    name="2"
                  )
              ]
            )
          )
        ]), 
        lb_types.ClassificationAnswer(
            name = "B"
        )
      ]
    )
 )

response_text_annotation = lb_types.ClassificationAnnotation(
    name="Response Text",
    value=lb_types.Text(answer="Look Morty, I turned myself into response text!"),
)


label = []
label.append(
  lb_types.Label(
    data=lb_types.LlmPromptResponseCreationData(
        global_key="c7a9ef3d-f417-479a-9352-629b354adbcc"
    ),
    annotations=[
        prompt_annotation,
        response_radio_annotation,
        response_checklist_annotation,
        response_text_annotation
    ]
  )
)
  
upload_job_mal = lb.MALPredictionImport.create_from_objects(
    client = client, 
    project_id = "clptu9e8c000cv2vu1nb7fii2", 
    name="mal_import_job_llm_data_gen"+str(uuid.uuid4()), 
    predictions=label)
  
upload_job_mal.wait_until_done()

upload_job = lb.LabelImport.create_from_objects(
    client = client, 
    project_id = "clptu9e8c000cv2vu1nb7fii2", 
    name="label_import_job"+str(uuid.uuid4()),  
    labels=label)

upload_job.wait_until_done()

Response creation

import sys
import labelbox as lb
import labelbox.types as lb_types
import uuid

API_KEY=""
client = lb.Client(API_KEY, "http://localhost:8080/_gql")

response_radio_annotation = lb_types.ClassificationAnnotation(
    name="Response Radio",
    value=lb_types.Radio(answer=lb_types.ClassificationAnswer(
        name="A",
        classifications=[
        lb_types.ClassificationAnnotation(
          name="Nested radio",
          value=lb_types.Radio(
            answer=lb_types.ClassificationAnswer(
              name="2"
            )
          )
        )
      ])))

response_checklist_annotation= lb_types.ClassificationAnnotation(
  name="Response Checklist", # must match your ontology feature's name
  value=lb_types.Checklist(
      answer = [
        lb_types.ClassificationAnswer(
        name="A",
        classifications=[
          lb_types.ClassificationAnnotation(
            name="Nested Checklist",
            value=lb_types.Checklist(
              answer=[
                  lb_types.ClassificationAnswer(
                    name="2"
                  )
              ]
            )
          )
        ]), 
        lb_types.ClassificationAnswer(
            name = "B"
        )
      ]
    )
 )

response_text_annotation = lb_types.ClassificationAnnotation(
    name="Response Text",
    value=lb_types.Text(answer="Look Morty, I turned myself into response text!"),
)


label = []
label.append(
  lb_types.Label(
    data=lb_types.LlmResponseCreationData(
        uid="cloqxnmw8007bb4vu22yf3966"
    ),
    annotations=[
        response_radio_annotation,
        response_checklist_annotation,
        response_text_annotation
    ]
  )
)
  
upload_job_mal = lb.MALPredictionImport.create_from_objects(
    client = client, 
    project_id = "clptw032p029hv2vu15su45ho", 
    name="mal_import_job_llm_data_gen"+str(uuid.uuid4()), 
    predictions=label)
  
upload_job_mal.wait_until_done()

upload_job = lb.LabelImport.create_from_objects(
    client = client, 
    project_id = "clptw032p029hv2vu15su45ho", 
    name="label_import_job"+str(uuid.uuid4()),  
    labels=label)

upload_job.wait_until_done()

[kkim-labelbox]

Do we have integration tests added for these new data types?

[lgluszek]

Do we have integration tests added for these new data types?

Added tests 6bf4073