DanielWalnut cd5bedaa74
feat: MiniMax provider for image/video/podcast skills + new music-generation skill (#3437)
* docs(spec): MiniMax integration for generation skills + new music skill

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* docs(plan): MiniMax generation providers implementation plan

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* test(skills): add importlib loader + FakeResp for skill tests

* test(skills): register loaded module in sys.modules; raise requests.HTTPError in FakeResp

* feat(image-generation): add MiniMax provider with env auto-detect

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* refactor(image-generation): guard unknown provider, derive ref MIME, strengthen tests

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(video-generation): add MiniMax provider with async poll/download

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* refactor(video-generation): surface base_resp errors while polling; add timeout test

* feat(podcast-generation): add MiniMax t2a_v2 provider with env auto-detect

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* refactor(podcast-generation): restore TTS credential guard; add volcengine + voice tests

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* feat(music-generation): new MiniMax music skill via skill-creator

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* refactor(music-generation): treat empty lyrics as absent; test no-audio-data path

* refactor(skills): add request timeouts to MiniMax network calls

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

* Potential fix for pull request finding 'Explicit returns mixed with implicit (fall through) returns'

Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>

* fix(models): strip inconsistent user-message names for MiniMax chat

DeerFlow middlewares tag user messages with provenance names (user-input, summary, loop_warning); langchain serializes them into the OpenAI-compatible payload and MiniMax rejects mismatched user-message names with "user name must be consistent (2013)". PatchedChatMiniMax now drops the per-message name from user-role messages. Point the config.example MiniMax models at PatchedChatMiniMax so they also get reasoning_content mapping.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(image-generation): MiniMax sends JSON prompt field, guard 1500-char limit

MiniMax image-01 takes one text string capped at 1500 chars, but the skill was sending the whole structured JSON. The MiniMax provider now extracts the JSON `prompt` field (relying on prompt_optimizer to expand it) and fails fast with a clear error before calling the API when that field exceeds 1500 chars. Authoring stays provider-agnostic; Gemini still receives the full JSON.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* feat(podcast-generation): per-provider TTS concurrency and retry/backoff

Each TTS provider owns its concurrency internally — MiniMax runs single-threaded to reduce rate-limit failures, Volcengine keeps 4 workers — with automatic retry and backoff on transient HTTP and base_resp errors. No caller-facing concurrency knob.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(skills): address Copilot review comments on generation skills

- video: add raise_for_status + timeout to the Gemini download/POST/poll calls so non-2xx responses surface as clear HTTP errors instead of JSON/KeyError or hangs
- video: check the task Fail status before the generic base_resp check so the failure keeps its task_id context
- video/image: create the output file parent directory before writing (matching music-generation) so nested output paths do not raise FileNotFoundError
- music: require a non-empty prompt and fail fast with ValueError instead of sending an empty prompt to the API

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(scripts): reclaim dev ports across worktrees in make stop/dev

All deer-flow worktrees (main checkout + linked worktrees) hardcode the same dev ports (8001/3000/2026), so a service started from any worktree must be reclaimable from another. stop_all now resolves the set of worktree roots (DEERFLOW_ROOTS) and treats a process as deer-flow-owned when its open files live under any of them. It also force-kills survivors on 2026 alongside 8001/3000, fixing `make dev` aborting on the nginx port preflight when a prior nginx lingered on 2026.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(view-image): hide the injected image-context message from the UI

ViewImageMiddleware injects a HumanMessage (text + base64 images) so the vision model can see viewed images, but it was the only internal injector that set neither hide_from_ui nor a hidden name, so it leaked into the chat UI (and IM channels) as a user bubble reading "Here are the images you've viewed:". Mark it with additional_kwargs={"hide_from_ui": True}, matching todo/dynamic_context injections, which the frontend isHiddenFromUIMessage and the channel sender already honor. The model still receives the full content.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

* fix(minimax): mark M2.7 models as text-only (no vision)

MiniMax M2.7 / M2.7-highspeed do not support vision; only M3 does. The
provider config asserted vision support for M2.7 in four places.

- config.example.yaml: 4 M2.7 entries -> supports_vision: false
- backend/docs/CONFIGURATION.md: M2.7 + highspeed -> supports_vision: false
- wizard: add LLMProvider.model_vision_overrides + extra_config_for() so
  selecting an M2.7 model writes supports_vision: false while M3 (default)
  keeps vision; wire it through setup_wizard.py
- tests: M2.7-highspeed fixture -> supports_vision=False; add
  test_minimax_vision_is_per_model

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Willem Jiang <willem.jiang@gmail.com>
Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
2026-06-08 22:04:38 +08:00

223 lines
8.0 KiB
Python

import base64
import os
import time
import requests
MINIMAX_DEFAULT_HOST = "https://api.minimaxi.com"
def _resolve_provider(override_env: str, existing_provider: str, has_existing_creds: bool) -> str:
"""Pick the provider: <SKILL>_PROVIDER override > existing creds > MiniMax fallback."""
override = os.getenv(override_env)
if override:
return override.strip().lower()
if has_existing_creds:
return existing_provider
if os.getenv("MINIMAX_API_KEY"):
return "minimax"
raise ValueError(
f"No credentials found. Set GEMINI_API_KEY for {existing_provider}, "
f"or MINIMAX_API_KEY for minimax (optionally force with {override_env})."
)
def _minimax_host() -> str:
return os.getenv("MINIMAX_API_HOST", MINIMAX_DEFAULT_HOST).rstrip("/")
def _ensure_output_dir(output_file: str) -> None:
"""Create the output file's parent directory so nested paths don't fail."""
output_dir = os.path.dirname(output_file)
if output_dir:
os.makedirs(output_dir, exist_ok=True)
def _check_base_resp(payload: dict) -> None:
base = payload.get("base_resp") or {}
if base.get("status_code", 0) != 0:
raise Exception(f"MiniMax error {base.get('status_code')}: {base.get('status_msg')}")
def _guess_mime(image_path: str) -> str:
ext = os.path.splitext(image_path)[1].lower()
return {
".png": "image/png",
".webp": "image/webp",
".gif": "image/gif",
".jpg": "image/jpeg",
".jpeg": "image/jpeg",
}.get(ext, "image/jpeg")
def _to_data_url(image_path: str) -> str:
with open(image_path, "rb") as f:
b64 = base64.b64encode(f.read()).decode("utf-8")
return f"data:{_guess_mime(image_path)};base64,{b64}"
def _poll_video_task(host: str, auth: str, task_id: str,
max_attempts: int = 120, interval: int = 3) -> str:
for _ in range(max_attempts):
response = requests.get(
f"{host}/v1/query/video_generation",
headers={"Authorization": auth},
params={"task_id": task_id},
timeout=30,
)
response.raise_for_status()
payload = response.json()
status = payload.get("status")
if status == "Success":
return payload["file_id"]
if status == "Fail":
base = payload.get("base_resp") or {}
raise Exception(
f"MiniMax video task {task_id} failed: "
f"{base.get('status_code')} {base.get('status_msg')}"
)
# Surface query-level errors (bad task_id, auth) that arrive as a non-zero
# base_resp without a terminal status, then keep polling.
_check_base_resp(payload)
time.sleep(interval)
raise Exception(f"MiniMax video task {task_id} timed out after {max_attempts} polls")
def _retrieve_file_url(host: str, auth: str, file_id: str) -> str:
response = requests.get(
f"{host}/v1/files/retrieve",
headers={"Authorization": auth},
params={"file_id": file_id},
timeout=30,
)
response.raise_for_status()
payload = response.json()
_check_base_resp(payload)
return payload["file"]["download_url"]
def _download(url: str, output_file: str) -> None:
response = requests.get(url, timeout=300)
response.raise_for_status()
_ensure_output_dir(output_file)
with open(output_file, "wb") as f:
f.write(response.content)
def _generate_video_minimax(
prompt: str, reference_images: list[str], output_file: str
) -> str:
api_key = os.getenv("MINIMAX_API_KEY")
if not api_key:
return "MINIMAX_API_KEY is not set"
host = _minimax_host()
auth = f"Bearer {api_key}"
body = {"model": os.getenv("MINIMAX_VIDEO_MODEL", "MiniMax-Hailuo-2.3"), "prompt": prompt}
if reference_images:
body["first_frame_image"] = _to_data_url(reference_images[0])
response = requests.post(
f"{host}/v1/video_generation",
headers={"Authorization": auth, "Content-Type": "application/json"},
json=body,
timeout=60,
)
response.raise_for_status()
payload = response.json()
_check_base_resp(payload)
task_id = payload["task_id"]
file_id = _poll_video_task(host, auth, task_id)
download_url = _retrieve_file_url(host, auth, file_id)
_download(download_url, output_file)
return f"The video has been generated successfully to {output_file}"
def download(url: str, output_file: str) -> None:
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
raise ValueError("GEMINI_API_KEY is not set")
response = requests.get(url, headers={"x-goog-api-key": api_key}, timeout=300)
response.raise_for_status()
_ensure_output_dir(output_file)
with open(output_file, "wb") as f:
f.write(response.content)
def _generate_video_gemini(
prompt: str, reference_images: list[str], output_file: str
) -> str:
reference_payload = []
request_json = {"instances": [{"prompt": prompt}]}
for reference_image in reference_images:
with open(reference_image, "rb") as f:
image_b64 = base64.b64encode(f.read()).decode("utf-8")
reference_payload.append(
{"image": {"mimeType": "image/jpeg", "bytesBase64Encoded": image_b64},
"referenceType": "asset"}
)
if reference_payload:
request_json["instances"][0]["referenceImages"] = reference_payload
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
return "GEMINI_API_KEY is not set"
response = requests.post(
"https://generativelanguage.googleapis.com/v1beta/models/veo-3.1-generate-preview:predictLongRunning",
headers={"x-goog-api-key": api_key, "Content-Type": "application/json"},
json=request_json,
timeout=60,
)
response.raise_for_status()
data = response.json()
operation_name = data["name"]
while True:
response = requests.get(
f"https://generativelanguage.googleapis.com/v1beta/{operation_name}",
headers={"x-goog-api-key": api_key},
timeout=30,
)
response.raise_for_status()
data = response.json()
if data.get("done", False):
sample = data["response"]["generateVideoResponse"]["generatedSamples"][0]
download(sample["video"]["uri"], output_file)
break
time.sleep(3)
return f"The video has been generated successfully to {output_file}"
def generate_video(
prompt_file: str,
reference_images: list[str],
output_file: str,
aspect_ratio: str = "16:9",
) -> str:
with open(prompt_file, "r", encoding="utf-8") as f:
prompt = f.read()
provider = _resolve_provider(
"VIDEO_GENERATION_PROVIDER", "gemini", bool(os.getenv("GEMINI_API_KEY"))
)
if provider == "minimax":
# MiniMax video uses resolution/duration, not aspect_ratio; aspect_ratio ignored.
return _generate_video_minimax(prompt, reference_images, output_file)
if provider in ("gemini", "google"):
return _generate_video_gemini(prompt, reference_images, output_file)
raise ValueError(f"Unknown video provider: {provider!r} (use 'gemini' or 'minimax')")
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(description="Generate videos using Gemini or MiniMax API")
parser.add_argument("--prompt-file", required=True, help="Absolute path to JSON prompt file")
parser.add_argument("--reference-images", nargs="*", default=[],
help="Absolute paths to reference images (space-separated)")
parser.add_argument("--output-file", required=True, help="Output path for generated video")
parser.add_argument("--aspect-ratio", required=False, default="16:9",
help="Aspect ratio of the generated video (Gemini only)")
args = parser.parse_args()
try:
print(generate_video(args.prompt_file, args.reference_images,
args.output_file, args.aspect_ratio))
except Exception as e:
print(f"Error while generating video: {e}")