[BUG] v4.7.10: praisonai videos CLI default sora invalid — sora-2 required
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 2/5
- Tempo stimato
- 1-3 ore
- Idoneità per principianti
- 70/100
Direzione di ricerca
Il valore predefinito è impostato da default="sora" in handle_videos nel file src/praisonai-code/praisonai_code/cli/features/capabilities.py, e il modello analizzato viene passato direttamente a video_generate. Inizia modificando quel valore predefinito e verifica praisonai models validate su di esso. È completato quando il valore predefinito viene analizzato in un id convalidato, un test verifica il valore predefinito analizzato senza -m e praisonai videos --help mostra il nuovo valore. Chiarisci la questione di alias e prefisso nel thread prima di aggiungere qualsiasi normalizzazione.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
[BUG] v4.7.10: CLI praisonai videos defaults to model sora — not in LiteLLM catalogue (sora-2 required)
Metadata
| Field | Value |
|---|---|
| Repository | https://github.com/MervinPraison/PraisonAI |
| Release tested | PraisonAI v4.7.10 (db166f427) |
| Package | praisonai-code CLI (CapabilitiesHandler.handle_videos) |
| Delegates to | praisonai.capabilities.videos.video_generate |
| Labels | bug, cli, video, sdk-contract, openai |
| Severity | High — primary CLI video entrypoint uses invalid default model id |
| Component | src/praisonai-code/praisonai_code/cli/features/capabilities.py |
| Discovered | 2026-09-28 OpenAI catalog + capabilities audit |
| OpenAI reference | https://developers.openai.com/api/docs/models/all (Sora / video models) |
| Duplicate search | gh search issues --repo MervinPraison/PraisonAI "praisonai videos" sora |
Executive summary
The legacy capabilities CLI exposes video generation via praisonai videos <prompt>. Argument parsing sets:
parser.add_argument("--model", "-m", default="sora", help="Model to use")
On v4.7.10, praisonai models validate sora fails and suggests sora-2. The canonical OpenAI video model id in LiteLLM’s catalogue is sora-2 (validation SUCCESS, video backend available via OpenAI provider).
Users who run the CLI without -m therefore pass an invalid model into video_generate(), compounding failures from the capabilities layer (wrong Runway default when users override incorrectly) and ensuring OpenAI-first CLI users fail validation before any API call.
Expected vs actual
| User action | Expected | Actual v4.7.10 |
|---|---|---|
praisonai videos "ocean waves" |
Uses current OpenAI Sora sku | Passes model=sora → invalid |
praisonai models validate sora |
N/A or alias redirect | ERROR, hint sora-2 |
praisonai models validate sora-2 |
SUCCESS | SUCCESS |
Help text --model default |
Matches validator | Shows default sora |
Root cause
CapabilitiesHandler.handle_videos in capabilities.py hardcodes default="sora". OpenAI renamed/shipped API ids as sora-2 (and related skus); bare sora is not registered in LiteLLM’s model cost map used by ModelCatalogue.
The handler forwards parsed.model directly:
result = video_generate(
prompt=parsed.prompt,
model=parsed.model,
)
No validation or normalization step runs before the capability call.
Architecture
flowchart LR
U[User CLI] --> H[handle_videos]
H --> P[argparse default sora]
P --> VG[video_generate]
VG --> LLM[litellm.video_generation]
CAT[ModelCatalogue] -.->|would reject sora| P
Minimal reproduction
pip install "praisonai==4.7.10"
python -m praisonai --version
python -m praisonai models validate sora
python -m praisonai models validate sora-2
# CLI path (requires entry through legacy capabilities router — same default in source)
python -m praisonai videos "a red ball on grass" --model sora
python -m praisonai videos "a red ball on grass" --model sora-2
Terminal evidence — v4.7.10 (2026-09-28)
$ python -m praisonai --version
PraisonAI version 4.7.10
$ python -m praisonai models validate sora
ERROR: ❌ 'sora' is not a valid model
INFO: Did you mean one of these?
• sora-2
$ python -m praisonai models validate sora-2
SUCCESS: ✅ 'sora-2' is a valid model
Capabilities: tool-calling
When --model is omitted, the CLI uses sora, which fails the same validation logic integrators rely on.
Terminal evidence — source inspection (v4.7.10 tag)
# src/praisonai-code/praisonai_code/cli/features/capabilities.py
@staticmethod
def handle_videos(args, unknown_args):
parser = argparse.ArgumentParser(prog="praisonai videos")
parser.add_argument("prompt", help="Video description")
parser.add_argument("--model", "-m", default="sora", help="Model to use")
Tag v4.7.10 contains this default unchanged.
Impact
| Audience | Effect |
|---|---|
| CLI users | Default command broken / confusing errors |
| Technical writers | Docs show sora shorthand incompatible with validator |
| OpenAI catalog alignment | Catalog documents Sora 2 class models, not bare sora |
| Support burden | Users file “Sora doesn’t work” when id string is wrong |
Suggested fix
- Change CLI default to
sora-2oropenai/sora-2(match LiteLLM provider string used elsewhere in repo). - On parse, optionally
ModelCatalogue.validate_model()with friendly error before calling API. - Add alias map:
sora→sora-2with deprecation warning (one release cycle). - Update
--helpand online docs to show validated id. - CLI integration test:
praisonai videos --helpshows new default; mockvideo_generate.
Acceptance criteria
- Default
--modelvalue passespraisonai models validate -
praisonai videos "test"without flags uses validated id - Deprecation warning if user passes
soraalias (optional) - Changelog entry for CLI default change
Workaround
Always specify model explicitly:
python -m praisonai videos "A cat playing piano" --model sora-2
Or use Python API with explicit id:
from praisonai.capabilities.videos import video_generate
video_generate("A cat playing piano", model="openai/sora-2", api_key=os.environ["OPENAI_API_KEY"])
(Exact provider prefix should match LiteLLM docs for your installed version.)
Relation to capabilities default bug
Separate but related issue: video_generate() Python default is runway/gen3a_turbo (invalid). CLI default is sora (invalid). Validators suggest runwayml/gen4_turbo and sora-2. A coordinated fix should publish one supported defaults table across CLI, capabilities, MCP, and VideoAgent docs.
Appendix — comparison table (v4.7.10 defaults)
| Entrypoint | Default model | models validate |
|---|---|---|
video_generate() |
runway/gen3a_turbo |
FAIL |
praisonai videos |
sora |
FAIL |
VideoAgent docs |
runwayml/gen4_turbo |
PASS |
| OpenAI catalog focus | sora-2 class |
PASS as sora-2 |
Appendix — UX mock (desired help output)
usage: praisonai videos [-h] [-m MODEL] [-o OUTPUT] prompt
--model, -m Model to use (default: sora-2)
Appendix — test plan
| Case | Expected |
|---|---|
Parse args with no -m |
model == "sora-2" |
| Validate default | exit 0 |
Mock video_generate |
called with sora-2 |
Appendix — severity rationale
High because CLI is the discoverable interface for non-Python users and the default is objectively invalid per the project’s own validator.
Appendix — OpenAI catalog note
OpenAI groups video models under image/video sections of the unified catalog. Integrators cross-checking https://developers.openai.com/api/docs/models/all expect sora-2-class ids, not legacy shorthand sora.
Appendix — mermaid (CLI → validate gap)
sequenceDiagram
participant User
participant CLI as praisonai videos
participant VG as video_generate
participant Val as models validate
User->>Val: validate sora
Val-->>User: ERROR (suggest sora-2)
User->>CLI: videos "prompt" (no -m)
CLI->>VG: model=sora
Note over CLI,Val: CLI ignores validator on defaults
Appendix — filing checklist
- Confirm legacy router still exposes
videossubcommand in v4.7.10 - Link release tag v4.7.10
- Cross-link capabilities video default issue in GitHub description
Appendix — extended validator transcript
$ python -m praisonai models validate sora
ERROR: ❌ 'sora' is not a valid model
INFO: Did you mean one of these?
• sora-2
$ python -m praisonai models validate sora-2
SUCCESS: ✅ 'sora-2' is a valid model
Capabilities: tool-calling
$ python -m praisonai models validate openai/sora-2
# Verify whether prefixed form also validates on your LiteLLM pin; document in fix PR
Appendix — documentation updates
Update any README / docs site references:
- Replace bare
sorawithsora-2oropenai/sora-2 - Add note that
praisonai models validateis the source of truth for CLI defaults
Appendix — post-fix verification
python -m praisonai models validate sora-2
python -m praisonai videos "test prompt" # mock API or dry-run flag if added
Appendix — historical context
Earlier PraisonAI releases introduced Sora support via LiteLLM as OpenAI video backend matured. CLI default sora likely predates sora-2 catalogue entry and was never updated when LiteLLM aligned ids.
Appendix — duplicate search commands
gh search issues --repo MervinPraison/PraisonAI "default=\"sora\"" --limit 20
gh search issues --repo MervinPraison/PraisonAI "handle_videos" --limit 20
Appendix — argparse behavior detail
When users pass -m sora, argparse does not consult ModelCatalogue. A fix should centralize normalization:
def normalize_video_model(name: str) -> str:
aliases = {"sora": "sora-2"}
return aliases.get(name, name)
Call from both CLI and optional video_generate wrapper.
Appendix — exit codes
Today CLI returns 1 on generation exceptions. After fix, distinguish:
| Code | Meaning |
|---|---|
| 0 | success |
| 1 | generation failed (provider) |
| 2 | invalid model before API |
Appendix — shell completion (future)
If CLI completion suggests models from catalogue, sora must not appear unless aliased.
Appendix — Windows vs Unix
Issue reproduced on Windows; argparse defaults are platform-independent — expect same invalid default on Linux/macOS.
Appendix — integration with VideoAgent docs
Recipe creator and SDK knowledge modules mention Sora strings — grep repo for bare \bsora\b defaults outside CLI.
Appendix — telemetry (optional product)
If CLI telemetry exists, tag video_model_invalid_default when users omit -m pre-fix to measure adoption after fix.
Appendix — full CLI error path (expected after using invalid default)
User runs:
praisonai videos "ocean sunset"
Internal call:
video_generate(..., model="sora")
Downstream LiteLLM may treat sora as unknown provider/model — error message varies. Normalizing to sora-2 aligns errors with OpenAI auth/quota only.
Appendix — maintainer grep commands
git grep -n 'default="sora"' src/praisonai-code
git grep -n "default='sora'" src/praisonai-code
Appendix — user-facing release note (sample)
Breaking:
praisonai videosnow defaults tosora-2instead ofsora. Pass-m sora-2explicitly in scripts.
Appendix — cross-provider CLI defaults policy
Document principle: every capabilities CLI default must pass praisonai models validate on release branch CI.
Appendix — Typer vs legacy argparse
Capabilities video command lives in legacy CapabilitiesHandler. Confirm whether modern Typer CLI exposes duplicate video commands — both must share one default constant exported from praisonai.constants or similar.
Appendix — constant extraction pattern
# proposed: src/praisonai/praisonai/constants.py
DEFAULT_OPENAI_VIDEO_MODEL = "sora-2"
DEFAULT_RUNWAY_VIDEO_MODEL = "runwayml/gen4_turbo"
CLI and capabilities import same symbols — prevents third drift.
Appendix — help text audit
Run:
python -m praisonai videos --help
Ensure -m help lists validated examples: sora-2, runwayml/gen4_turbo, gemini/veo-3.1-lite-generate-preview.
Appendix — nightly CI job (proposal)
- run: python -m praisonai models validate sora-2
- run: python -c "import praisonai_code.cli.features.capabilities as c; ..."
Fail build if CLI default ≠ validated constant.
Appendix — GitHub issue title (copy-paste)
[BUG] praisonai videos CLI default model sora invalid — use sora-2 on v4.7.10
Appendix — labels
bug, cli, video, sdk-contract, openai
Appendix — user story
As a CLI user, I want praisonai videos "prompt" to use a model id that passes praisonai models validate, so I only debug API keys and quotas — not SDK defaults.
Appendix — owner routing
| Team | Action |
|---|---|
| CLI | Change argparse default |
| SDK | Export shared video default constant |
| Docs | Update CLI reference |
| QA | CLI snapshot test for --help default |
Appendix — severity escalation path
If OpenAI removes alias handling entirely, invalid CLI defaults become release blockers — treat as High until fixed.
End of issue document.
- Lingua principale
- Python
- Stelle
- 9.3k
- Fork
- 1.5k
- Merge medio
- 1g 5h
- PR unite (30g)
- 520
Preparare l'ambiente
Questo progetto non fornisce container di sviluppo, Dockerfile né guida per i contributori, quindi l'ambiente è a tuo carico: parti dal suo README e consulta la nostra guida al primo contributo per i passaggi generali.
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di MervinPraison/PraisonAI
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
MervinPraison/PraisonAI#6710 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
claude
Difficoltà 2/5 1-3 ore Idoneità per principianti 74/100
MervinPraison/PraisonAI#6707 ·
I maintainer di solito rispondono entro 1 giorno
-
bug claude
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 90/100
MervinPraison/PraisonAI#6671 ·
I maintainer di solito rispondono entro 1 giorno
-
Compaction budget regression fixture remains below budget with accurate tokenizationForse di nuovo libera Una pull request per questa issue è stata chiusa senza essere unita. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
MervinPraison/PraisonAI#5418 · 3 commenti ·
I maintainer di solito rispondono entro 1 giorno
-
bug claude
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
MervinPraison/PraisonAI#5096 · 2 commenti ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di MervinPraison/PraisonAI
Issue simili
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 76/100
I maintainer di solito rispondono entro 2 giorni
-
[Bug]: LANGSMITH_SAMPLING_RATE ignores fractional values (isdigit check), traces 100% of requestsAperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 84/100
I maintainer di solito rispondono entro 1 giorno
-
platform/backend tests
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
Significant-Gravitas/AutoGPT#17100 ·
I maintainer di solito rispondono entro 1 giorno
-
bug good first issue
Difficoltà 2/5 1-3 ore Idoneità per principianti 64/100
-
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 61/100
scverse/infercnvpy#238 ·
I maintainer di solito rispondono entro 1 giorno