Physical AI Event Video Generation

PAIDF Orchestration — Event Video Generation

Run the Event Video Generation DAG end to end: seed-image input preparation, Cosmos3 image-to-video anomaly augmentation, auto-labeling (detection and tracking, captioning, anomaly visual QA, person-attribute visual QA, person attribute search), anomaly dataset generation, and result retrieval.

DAG selection

The workflow builds one DAG per compute platform from airflow/dags/workflows/event_video_generation_dag/:

Kubernetes is the only platform whose manifest is checked in, so event_video_generation_dag_k8s is the only DAG this repository registers. A DAG is registered only if its manifest exists; a missing manifest means the DAG is absent from Airflow rather than broken. List the DAGs Airflow actually loaded before triggering, and never name a DAG ID that is not in that list.

There is a single end-to-end pipeline — there are no generation-only or labeling-only DAG variants. If a user asks for video generation without auto-labeling, tell them the checked-in DAG does not offer that flow rather than inventing a DAG ID.

Manual payload entry in the Airflow UI

If the user wants to enter their own payload directly in the Airflow UI rather than have you construct and trigger one, your job is limited to getting them to the UI: confirm controller readiness, ensure make port-forward is running (see airflow-direct-api.md), and report the reachable URL. Do not render a payload, run preflight, or trigger a run yourself in this case — the user is doing that from the UI. Resume monitoring (step 6 below) once they tell you a run has been triggered; you can find it via the Airflow API without needing the payload they used.

Scope

Before building any payload, collect all of the following from the user. Do not fall back to repository defaults, CI payloads, or any hardcoded endpoint URL or bucket path.

If the user does not provide a required value, ask for it explicitly before proceeding. Do not invent or reuse values from previous runs or checked-in files.

Always run the following readiness checks before triggering a run. The checks are short-circuiting — stop at the first failure and route to the environment-setup skill immediately.

Before any check, establish the cluster connection. The cluster is reached only through credentials the user supplies — they are never part of the repository. Check whether the cluster credential file path is already exported in the shell environment; if not, ask the user for the absolute path before running any cluster command. Never assume a path or fall back to any on-disk default — see setup-and-preflight.md for the full procedure.

The controller (Airflow) and DAG compute tasks run on the same cluster unless a different remote cluster connection was configured. GPU capacity is checked on this cluster.

1. Controller pods — check that the Airflow controller pods (not DAG task pods) are Running. DAG task pods in Pending or Failed state are normal and must not be mistaken for controller failures:

   kubectl get pods -n sdg-workflow -l "release=sdg-workflow-controller"
   

   All pods matching the release=sdg-workflow-controller label must be Running. If the namespace is absent, this is a first-install condition — route to the environment-setup skill, do not diagnose further.

2. Airflow API — reachable only if check 1 passes. First establish AIRFLOW_URL from the Kubernetes ClusterIP (always routable from the host, no port-forward required):

   AIRFLOW_URL="http://$(kubectl get svc -n sdg-workflow \
     sdg-workflow-controller-api-server \
     -o jsonpath='{.spec.clusterIP}'):8080"
   

   Then confirm the target DAG is loaded and is_paused: False. See airflow-direct-api.md for the full auth + check sequence. If the API is unreachable, route to the environment-setup skill.

3. Pools — only if check 2 passes. Required pools with open slots: k8s_gpu_1, default_pool, and the image2video pool for the chosen mode (external_image2video_service_pool for external, internal_image2video_service_pool for internal).

4. Compute-cluster GPUs — check the cluster (using the cluster connection established above):

   kubectl get nodes \
     -o custom-columns='NAME:.metadata.name,GPU_ALLOC:.status.allocatable.nvidia\.com/gpu'
   # Also check pods already consuming GPUs — capacity ≠ availability on a shared cluster
   kubectl get pods -n sdg-workflow \
     --field-selector=status.phase=Running -o wide
   

   The compute cluster is shared — other users' runs may be active. Report GPUs as free-versus-total, not just allocatable. Two independent GPU sources, only one of which is mode-dependent:

   • Task pods that run local model inference regardless of service mode: detection_and_tracking, captioning, and visual_qa all run on the k8s_gpu_task profile (1 GPU each) — this cost applies in both external and internal mode, since these do in-pod inference rather than calling an endpoint. event_and_person_attribute_search and augmentation run on CPU profiles and cost nothing. External mode therefore needs a minimum of three GPUs, not zero.
   • Internally deployed endpoints (external_services: false): one GPU per VLM/LLM replica, plus two GPUs per image2video replica (gpu_count: 2, host_ipc: true) — four GPUs for one replica of each service.
   • Internal mode total = both sources combined: the three task-pod GPUs plus the four endpoint GPUs — at minimum seven GPUs, not four.

5. Stale failed pods — before triggering, check for accumulated failed pods in the compute namespace and report them. They are retained by design and do not affect run correctness, but they consume namespace quota and clutter log searches:

   kubectl get pods -n sdg-workflow \
     --field-selector=status.phase=Failed \
     -o custom-columns='NAME:.metadata.name,AGE:.metadata.creationTimestamp,DAG:.metadata.labels.dag_id'
   

   Clean up only pods whose dag_id label matches a run you own, after confirming with the user.

Document each check result explicitly.

If any check fails: invoke the environment-setup skill automatically — do not wait for the user to say "set up" or ask them to name the skill.

If the user's request implies first-time or explicit deployment ("deploy", "install", "set up", "reinstall", "redeploy", "full setup"): invoke the environment-setup skill even if all checks pass, and confirm the planned commands first.

If all checks pass and the user only wants to run the workflow: proceed directly to payload and trigger.

Bundled tools

• scripts/upload_images.py: validate/upload a local seed image or flat directory of seed images.
• scripts/payload.py: render or validate a standalone EventVideoGenerationDagPayloadConfig-compatible JSON.
• scripts/summarize_results.py: summarize a downloaded anomaly_dataset/dataset.json dataset.

Run commands from this skill directory. Credentials must be inherited from the shell that launched the agent; never ask the user to paste secret values into the prompt.

Procedure

1. Determine the input source.

   • For local data, validate before upload:

     python scripts/upload_images.py --path /path/to/seed-images --validate-only
     

   • Then upload:

     python scripts/upload_images.py \
       --path /path/to/seed-images --destination-path event-video-generation/my-run
     

   • For an existing storage URL, use it unchanged after confirming it names either a single image or a flat directory of images (.jpg, .jpeg, .png, .bmp, .gif, .tiff, .webp). Unlike person-crop workflows, there is no subdirectory convention — every matching file directly under input_path is one input image. When input_path names a directory, the DAG sorts matching files and takes the first max_images of them.

2. Select service mode.

   • external requires explicit VLM, LLM, and image2video endpoint URLs.
   • internal lets the DAG's service lifecycle deploy all three services in-cluster.
   • Choose service mode independently from controller placement. A local controller may use external inference endpoints.
   • Keep nested service mode and output directory consistent with the top level.
   • On Kubernetes, VLM and LLM each claim one GPU from k8s_gpu_1, but image2video claims two GPUs per replica (gpu_count: 2, host_ipc: true) — four GPUs for the internally deployed endpoints alone. That's on top of, not instead of, the three GPUs detection_and_tracking/captioning/visual_qa always claim from k8s_gpu_task regardless of service mode (see the readiness-check GPU breakdown above): external mode needs a minimum of three allocatable GPUs, internal mode a minimum of seven — not zero and four.

3. Read payload-contract.md, then render a payload from the values collected above. Do not copy checked-in dev or CI payloads — they contain deployment- specific endpoint URLs and bucket paths that must not be inherited by user runs.

   External:

   python scripts/payload.py render \
     --input-path s3://bucket/input/seed-images/ \
     --output-directory s3://bucket/output/event-video-generation/ \
     --service-mode external \
     --vlm-url https://vlm.example/v1 \
     --llm-url https://llm.example/v1 \
     --image2video-url https://image2video.example/v1 \
     --max-images 10 --num-augmentation 3 \
     --variable-distribution assets/variable-distribution.json \
     --output /tmp/evg-payload.json
   

   Internal:

   python scripts/payload.py render \
     --input-path s3://bucket/input/seed-images/ \
     --output-directory s3://bucket/output/event-video-generation/ \
     --service-mode internal \
     --max-images 10 --num-augmentation 3 \
     --output /tmp/evg-payload.json
   

   Show the user the rendered payload (or its validated contents) and get explicit confirmation before proceeding. Only continue to preflight and triggering if they confirm; if they want changes, re-render and re-confirm.

4. Preflight the DAG through the Airflow API. Check that the DAG is loaded, required pools have slots, and controller pods are healthy — see airflow-direct-api.md#preflight-direct-path. Confirm presence only; never print credential values.

5. Submit exactly one DAG run. Pass the payload from step 3 as conf.payload — see airflow-direct-api.md#trigger-a-run for the full request shape. Record and return the dag_run_id, input path, output directory, and service mode.

6. Immediately after triggering — without waiting to be asked — monitor the run until it reaches a terminal state (success or failed). Poll the Airflow API every 60–120 seconds:

   # Poll run state
   RESPONSE=$(curl -s -H "Authorization: Bearer $TOKEN" \
     "$AIRFLOW_URL/api/v2/dags/$DAG_ID/dagRuns/$RUN_ID")
   RESPONSE="$RESPONSE" python3 -c "import json, os; print(json.loads(os.environ['RESPONSE'])['state'])"
   

   For a per-task breakdown when state is running or failed, see airflow-direct-api.md.

   Stop polling as soon as the run state is success or failed. Use the polling loop that fits your runtime — a shell while loop, a background process, or a tool-native scheduler. Do not block the user waiting for each poll; report state changes as they occur.

   Tell the user they can also watch progress live in the Airflow UI. make port-forward runs in the foreground and never exits, so start it as a background job — and prefer that the user runs it in their own terminal, since an agent-owned forward dies with the session. Resolve the host's real address rather than reporting a placeholder or localhost, which is meaningless from another machine:

   HOST_IP=$(hostname -I | awk '{print $1}')
   echo "Airflow UI: http://$HOST_IP:8080"
   

   Default credentials are admin/admin, defined in deploy/values.yaml under airflow.createUserJob.defaultUser (not webserver.defaultUser). Update them before production use.

   For a full per-task breakdown see airflow-direct-api.md.

   To stop an in-progress run: open the Airflow UI, find the active DagRun, locate the running task, and mark it Failed (task menu → Mark Failed). This triggers the DAG's shutdown path, cleaning up Deployments, Services, and GPU pods. Do not delete the DagRun or the DAG — that bypasses cleanup and leaves stale cluster resources.

7. After the run reaches success or failed, ask the user: "Would you like to download and analyze the results?" Do not download automatically — wait for confirmation.

   If the user confirms, use whatever AWS credentials are already available in the shell environment (standard AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY / AWS_DEFAULT_REGION, an AWS profile, or instance role). Never ask the user to paste credentials into the prompt. Run artifacts live under <output_directory>/<run_id>/, where <output_directory> is the payload value and <run_id> is the dag_run_id from step 5. The final dataset is in anomaly_dataset/:

   aws s3 sync "<output_directory>/<run_id>/anomaly_dataset/" /tmp/evg-results/
   python scripts/summarize_results.py --results-dir /tmp/evg-results
   

   To inspect intermediate generated videos instead, sync <output_directory>/<run_id>/cosmos/ and read metadata.json from each <video_key>/<augmentation_index>/ folder.

   Read outputs.md before interpreting files.

Guardrails

• Never use default endpoint URLs, bucket paths, or input paths from the codebase or checked-in payloads. Always ask the user for every deployment-specific value before building a payload. If a required value is missing, stop and ask — do not substitute a guess.
• Preserve explicit user inputs and endpoint/model selections throughout the session.
• Do not submit if payload validation, local dataset validation, or Airflow preflight fails.
• Do not show AWS credentials, Airflow bearer tokens, or S3 signed URLs.
• Do not start multiple runs unless the user explicitly requests them.
• Ask for a dataset location if none was supplied; this workflow has no implicit demo dataset.
• Do not invent generation-only or labeling-only DAG IDs — only the DAG listed above exists.
• Only offer a platform whose manifest exists and whose DAG is loaded in Airflow.
• Do not route person-crop clothing/attribute augmentation requests here — that is image-attribute-augmentation-workflow.

References

• Read setup-and-preflight.md for environment, storage, and policy requirements.
• Read payload-contract.md when creating or changing a payload.
• Read outputs.md when retrieving or interpreting results.
• Read troubleshooting.md after validation, API, or runtime errors.
• Read airflow-direct-api.md for all Airflow interactions: preflight, triggering, monitoring, per-task breakdown, and log retrieval.