File size: 11,116 Bytes
a4583f2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
"""Higgsfield.ai cloud provider — access to Kling 3.0, Sora 2, Veo 3.1, and more.

Higgsfield provides a unified API for multiple AI video and image generation models
including Kling 3.0 (Kuaishou), Sora 2 (OpenAI), Veo 3.1 (Google), WAN 2.5 (Alibaba),
and their own Higgsfield Soul for character consistency.

SDK: pip install higgsfield-client
Docs: https://cloud.higgsfield.ai/
"""

from __future__ import annotations

import asyncio
import base64
import logging
import os
import time
import uuid
from typing import Any

import httpx

from content_engine.services.cloud_providers.base import CloudGenerationResult, CloudProvider

logger = logging.getLogger(__name__)

# Model IDs for Higgsfield platform
# Format: provider/model/task
TEXT_TO_IMAGE_MODELS = {
    "seedream-4": "bytedance/seedream/v4/text-to-image",
    "seedream-4.5": "bytedance/seedream/v4.5/text-to-image",
    "nano-banana-pro": "google/nano-banana-pro/text-to-image",
    "flux-2": "black-forest-labs/flux-2/text-to-image",
    "gpt-image": "openai/gpt-image/text-to-image",
    "default": "bytedance/seedream/v4/text-to-image",
}

TEXT_TO_VIDEO_MODELS = {
    "kling-3.0": "kuaishou/kling/v3.0/text-to-video",
    "kling-3.0-pro": "kuaishou/kling/v3.0-pro/text-to-video",
    "sora-2": "openai/sora-2/text-to-video",
    "veo-3.1": "google/veo-3.1/text-to-video",
    "wan-2.5": "alibaba/wan-2.5/text-to-video",
    "seedance-pro": "bytedance/seedance/pro/text-to-video",
    "default": "kuaishou/kling/v3.0/text-to-video",
}

IMAGE_TO_VIDEO_MODELS = {
    "kling-3.0": "kuaishou/kling/v3.0/image-to-video",
    "kling-3.0-pro": "kuaishou/kling/v3.0-pro/image-to-video",
    "sora-2": "openai/sora-2/image-to-video",
    "veo-3.1": "google/veo-3.1/image-to-video",
    "wan-2.5": "alibaba/wan-2.5/image-to-video",
    "higgsfield-dop": "higgsfield/dop/image-to-video",
    "default": "kuaishou/kling/v3.0/image-to-video",
}

IMAGE_EDIT_MODELS = {
    "higgsfield-soul": "higgsfield/soul/image-to-image",
    "seedream-4-edit": "bytedance/seedream/v4/edit",
    "default": "higgsfield/soul/image-to-image",
}


class HiggsFieldProvider(CloudProvider):
    """Cloud provider using Higgsfield.ai for Kling 3.0, Sora 2, Veo 3.1, etc."""

    def __init__(self, api_key: str = None, api_secret: str = None):
        """Initialize with Higgsfield credentials.

        Can use either:
        - Combined key (HF_KEY env var)
        - Separate key/secret (HF_API_KEY/HF_API_SECRET env vars)
        """
        self._api_key = api_key or os.getenv("HIGGSFIELD_API_KEY") or os.getenv("HF_API_KEY")
        self._api_secret = api_secret or os.getenv("HIGGSFIELD_API_SECRET") or os.getenv("HF_API_SECRET")
        self._combined_key = os.getenv("HF_KEY")

        self._http_client = httpx.AsyncClient(timeout=300)
        self._client = None

        # Try to initialize SDK if available
        try:
            from higgsfield_client import HiggsFieldClient
            if self._combined_key:
                self._client = HiggsFieldClient()
            elif self._api_key and self._api_secret:
                self._client = HiggsFieldClient(api_key=self._api_key, api_secret=self._api_secret)
            logger.info("Higgsfield SDK initialized")
        except ImportError:
            logger.warning("higgsfield-client not installed, using direct API")
        except Exception as e:
            logger.warning("Failed to init Higgsfield SDK: %s", e)

    @property
    def name(self) -> str:
        return "higgsfield"

    async def is_available(self) -> bool:
        """Check if Higgsfield API is configured."""
        return bool(self._client or self._api_key)

    def _resolve_t2i_model(self, model_name: str | None) -> str:
        """Resolve friendly name to Higgsfield model ID for text-to-image."""
        if model_name and model_name in TEXT_TO_IMAGE_MODELS:
            return TEXT_TO_IMAGE_MODELS[model_name]
        if model_name:
            return model_name
        return TEXT_TO_IMAGE_MODELS["default"]

    def _resolve_t2v_model(self, model_name: str | None) -> str:
        """Resolve friendly name to Higgsfield model ID for text-to-video."""
        if model_name and model_name in TEXT_TO_VIDEO_MODELS:
            return TEXT_TO_VIDEO_MODELS[model_name]
        if model_name:
            return model_name
        return TEXT_TO_VIDEO_MODELS["default"]

    def _resolve_i2v_model(self, model_name: str | None) -> str:
        """Resolve friendly name to Higgsfield model ID for image-to-video."""
        if model_name and model_name in IMAGE_TO_VIDEO_MODELS:
            return IMAGE_TO_VIDEO_MODELS[model_name]
        if model_name:
            return model_name
        return IMAGE_TO_VIDEO_MODELS["default"]

    def _resolve_edit_model(self, model_name: str | None) -> str:
        """Resolve friendly name to Higgsfield model ID for image editing."""
        if model_name and model_name in IMAGE_EDIT_MODELS:
            return IMAGE_EDIT_MODELS[model_name]
        if model_name:
            return model_name
        return IMAGE_EDIT_MODELS["default"]

    async def generate_image(
        self,
        prompt: str,
        model: str | None = None,
        resolution: str = "2K",
        aspect_ratio: str = "16:9",
        **kwargs,
    ) -> CloudGenerationResult:
        """Generate an image using Higgsfield text-to-image models."""
        start = time.time()
        model_id = self._resolve_t2i_model(model)

        if self._client:
            # Use SDK
            try:
                result = self._client.subscribe(
                    model_id,
                    {
                        "prompt": prompt,
                        "resolution": resolution,
                        "aspect_ratio": aspect_ratio,
                        **kwargs,
                    }
                )

                # Extract image URL
                images = result.get("images", [])
                if not images:
                    raise RuntimeError(f"No images in Higgsfield response: {result}")

                image_url = images[0].get("url") if isinstance(images[0], dict) else images[0]

                # Download image
                resp = await self._http_client.get(image_url)
                resp.raise_for_status()

                return CloudGenerationResult(
                    job_id=str(uuid.uuid4()),
                    image_bytes=resp.content,
                    generation_time_seconds=time.time() - start,
                )
            except Exception as e:
                logger.error("Higgsfield image generation failed: %s", e)
                raise
        else:
            raise RuntimeError("Higgsfield SDK not initialized")

    async def generate_video(
        self,
        prompt: str,
        model: str | None = None,
        duration: int = 5,
        resolution: str = "720p",
        aspect_ratio: str = "16:9",
        enable_audio: bool = False,
        image_url: str | None = None,
        **kwargs,
    ) -> dict:
        """Generate a video using Higgsfield models (Kling 3.0, Sora 2, Veo 3.1, etc.).

        Args:
            prompt: Text description of desired video
            model: Model to use (kling-3.0, sora-2, veo-3.1, etc.)
            duration: Video duration in seconds (3-15 for Kling 3.0)
            resolution: Output resolution (720p, 1080p)
            aspect_ratio: Aspect ratio (16:9, 9:16, 1:1)
            enable_audio: Enable audio generation (Kling 3.0 supports this)
            image_url: Reference image URL for image-to-video

        Returns:
            Dict with job_id, status, and video_url when complete
        """
        start = time.time()

        # Choose model based on whether we have an image
        if image_url:
            model_id = self._resolve_i2v_model(model)
        else:
            model_id = self._resolve_t2v_model(model)

        if self._client:
            try:
                payload = {
                    "prompt": prompt,
                    "resolution": resolution,
                    "aspect_ratio": aspect_ratio,
                    "duration": duration,
                }

                if enable_audio:
                    payload["enable_audio"] = True

                if image_url:
                    payload["image"] = image_url

                payload.update(kwargs)

                # Submit and wait for result
                result = self._client.subscribe(model_id, payload)

                # Extract video URL
                video_url = None
                if "video" in result:
                    video_url = result["video"]
                elif "outputs" in result and result["outputs"]:
                    video_url = result["outputs"][0]
                elif "output" in result:
                    video_url = result["output"]

                if not video_url:
                    raise RuntimeError(f"No video URL in Higgsfield response: {result}")

                return {
                    "job_id": str(uuid.uuid4()),
                    "status": "completed",
                    "video_url": video_url,
                    "generation_time": time.time() - start,
                }

            except Exception as e:
                logger.error("Higgsfield video generation failed: %s", e)
                raise
        else:
            raise RuntimeError("Higgsfield SDK not initialized")

    async def submit_generation(
        self,
        positive_prompt: str,
        negative_prompt: str = "",
        model: str | None = None,
        width: int = 1024,
        height: int = 1024,
        seed: int = -1,
        **kwargs,
    ) -> str:
        """Submit an image generation job. Returns job ID."""
        # For now, use synchronous generation
        result = await self.generate_image(
            prompt=positive_prompt,
            model=model,
            **kwargs,
        )

        # Cache result for get_result
        self._last_result = {
            "job_id": result.job_id,
            "result": result,
            "timestamp": time.time(),
        }

        return result.job_id

    async def check_status(self, job_id: str) -> str:
        """Check job status."""
        if hasattr(self, '_last_result') and self._last_result.get("job_id") == job_id:
            return "completed"
        return "unknown"

    async def get_result(self, job_id: str) -> CloudGenerationResult:
        """Get the generation result."""
        if not hasattr(self, '_last_result') or self._last_result.get("job_id") != job_id:
            raise RuntimeError(f"No cached result for job {job_id}")
        return self._last_result["result"]

    async def generate(
        self,
        positive_prompt: str,
        negative_prompt: str = "",
        model: str | None = None,
        width: int = 1024,
        height: int = 1024,
        seed: int = -1,
        **kwargs,
    ) -> CloudGenerationResult:
        """Convenience method: generate image in one call."""
        return await self.generate_image(
            prompt=positive_prompt,
            model=model,
            **kwargs,
        )