mirror of
https://github.com/harry0703/MoneyPrinterTurbo.git
synced 2026-03-10 16:57:22 +08:00
400 lines
21 KiB
Python
400 lines
21 KiB
Python
# ==============================================================================
|
||
# app/services/llm.py (最终修复版 - 含场景延续性规则)
|
||
# 操作指南：请复制所有代码，并完全覆盖您项目中的同名文件。
|
||
# ==============================================================================
|
||
|
||
import json
|
||
import logging
|
||
import re
|
||
import requests
|
||
from typing import List, Dict
|
||
|
||
import g4f
|
||
from loguru import logger
|
||
from openai import AzureOpenAI, OpenAI
|
||
from openai.types.chat import ChatCompletion
|
||
|
||
from app.config import config
|
||
|
||
_max_retries = 5
|
||
|
||
|
||
def _generate_response(prompt: str) -> str:
|
||
"""
|
||
与大语言模型（LLM）提供商进行交互。
|
||
(此函数已从您上传的文件中完整保留)
|
||
"""
|
||
try:
|
||
content = ""
|
||
llm_provider = config.app.get("llm_provider", "openai")
|
||
logger.info(f"llm provider: {llm_provider}")
|
||
if llm_provider == "g4f":
|
||
model_name = config.app.get("g4f_model_name", "")
|
||
if not model_name:
|
||
model_name = "gpt-3.5-turbo-16k-0613"
|
||
content = g4f.ChatCompletion.create(
|
||
model=model_name,
|
||
messages=[{"role": "user", "content": prompt}],
|
||
)
|
||
else:
|
||
api_version = "" # for azure
|
||
if llm_provider == "moonshot":
|
||
api_key = config.app.get("moonshot_api_key")
|
||
model_name = config.app.get("moonshot_model_name")
|
||
base_url = "https://api.moonshot.cn/v1"
|
||
elif llm_provider == "ollama":
|
||
api_key = "ollama" # any string works but you are required to have one
|
||
model_name = config.app.get("ollama_model_name")
|
||
base_url = config.app.get("ollama_base_url", "")
|
||
if not base_url:
|
||
base_url = "http://localhost:11434/v1"
|
||
elif llm_provider == "openai":
|
||
api_key = config.app.get("openai_api_key")
|
||
model_name = config.app.get("openai_model_name")
|
||
base_url = config.app.get("openai_base_url", "")
|
||
if not base_url:
|
||
base_url = "https://api.openai.com/v1"
|
||
elif llm_provider == "oneapi":
|
||
api_key = config.app.get("oneapi_api_key")
|
||
model_name = config.app.get("oneapi_model_name")
|
||
base_url = config.app.get("oneapi_base_url", "")
|
||
elif llm_provider == "azure":
|
||
api_key = config.app.get("azure_api_key")
|
||
model_name = config.app.get("azure_model_name")
|
||
base_url = config.app.get("azure_base_url", "")
|
||
api_version = config.app.get("azure_api_version", "2024-02-15-preview")
|
||
elif llm_provider == "gemini":
|
||
api_key = config.app.get("gemini_api_key")
|
||
model_name = config.app.get("gemini_model_name")
|
||
base_url = "***" # Placeholder from original file
|
||
elif llm_provider == "qwen":
|
||
api_key = config.app.get("qwen_api_key")
|
||
model_name = config.app.get("qwen_model_name")
|
||
base_url = "***" # Placeholder from original file
|
||
elif llm_provider == "cloudflare":
|
||
api_key = config.app.get("cloudflare_api_key")
|
||
model_name = config.app.get("cloudflare_model_name")
|
||
account_id = config.app.get("cloudflare_account_id")
|
||
base_url = "***" # Placeholder from original file
|
||
elif llm_provider == "deepseek":
|
||
api_key = config.app.get("deepseek_api_key")
|
||
model_name = config.app.get("deepseek_model_name")
|
||
base_url = config.app.get("deepseek_base_url")
|
||
if not base_url:
|
||
base_url = "https://api.deepseek.com"
|
||
elif llm_provider == "ernie":
|
||
api_key = config.app.get("ernie_api_key")
|
||
secret_key = config.app.get("ernie_secret_key")
|
||
base_url = config.app.get("ernie_base_url")
|
||
model_name = "***" # Placeholder from original file
|
||
if not secret_key:
|
||
raise ValueError(
|
||
f"{llm_provider}: secret_key is not set, please set it in the config.toml file."
|
||
)
|
||
elif llm_provider == "pollinations":
|
||
try:
|
||
base_url = config.app.get("pollinations_base_url", "")
|
||
if not base_url:
|
||
base_url = "https://text.pollinations.ai/openai"
|
||
model_name = config.app.get("pollinations_model_name", "openai-fast")
|
||
payload = {
|
||
"model": model_name,
|
||
"messages": [{"role": "user", "content": prompt}],
|
||
"seed": 101
|
||
}
|
||
if config.app.get("pollinations_private"):
|
||
payload["private"] = True
|
||
if config.app.get("pollinations_referrer"):
|
||
payload["referrer"] = config.app.get("pollinations_referrer")
|
||
headers = {"Content-Type": "application/json"}
|
||
response = requests.post(base_url, headers=headers, json=payload)
|
||
response.raise_for_status()
|
||
result = response.json()
|
||
if result and "choices" in result and len(result["choices"]) > 0:
|
||
content = result["choices"][0]["message"]["content"]
|
||
return content.replace("\n", "")
|
||
else:
|
||
raise Exception(f"[{llm_provider}] returned an invalid response format")
|
||
except requests.exceptions.RequestException as e:
|
||
raise Exception(f"[{llm_provider}] request failed: {str(e)}")
|
||
except Exception as e:
|
||
raise Exception(f"[{llm_provider}] error: {str(e)}")
|
||
|
||
if llm_provider not in ["pollinations", "ollama"]:
|
||
if not api_key:
|
||
raise ValueError(f"{llm_provider}: api_key is not set, please set it in the config.toml file.")
|
||
if not model_name:
|
||
raise ValueError(f"{llm_provider}: model_name is not set, please set it in the config.toml file.")
|
||
if not base_url:
|
||
raise ValueError(f"{llm_provider}: base_url is not set, please set it in the config.toml file.")
|
||
|
||
if llm_provider == "qwen":
|
||
import dashscope
|
||
from dashscope.api_entities.dashscope_response import GenerationResponse
|
||
dashscope.api_key = api_key
|
||
response = dashscope.Generation.call(model=model_name, messages=[{"role": "user", "content": prompt}])
|
||
if response:
|
||
if isinstance(response, GenerationResponse):
|
||
if response.status_code != 200:
|
||
raise Exception(f'[{llm_provider}] returned an error response: "{response}"')
|
||
content = response["output"]["text"]
|
||
return content.replace("\n", "")
|
||
else:
|
||
raise Exception(f'[{llm_provider}] returned an invalid response: "{response}"')
|
||
else:
|
||
raise Exception(f"[{llm_provider}] returned an empty response")
|
||
|
||
if llm_provider == "gemini":
|
||
import google.generativeai as genai
|
||
genai.configure(api_key=api_key, transport="rest")
|
||
generation_config = {"temperature": 0.5, "top_p": 1, "top_k": 1, "max_output_tokens": 8192}
|
||
safety_settings = [
|
||
{"category": "HARM_CATEGORY_HARASSMENT", "threshold": "BLOCK_ONLY_HIGH"},
|
||
{"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "BLOCK_ONLY_HIGH"},
|
||
{"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "BLOCK_ONLY_HIGH"},
|
||
{"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "BLOCK_ONLY_HIGH"},
|
||
]
|
||
model = genai.GenerativeModel(model_name=model_name, generation_config=generation_config, safety_settings=safety_settings)
|
||
try:
|
||
response = model.generate_content(prompt)
|
||
generated_text = response.candidates[0].content.parts[0].text
|
||
return generated_text
|
||
except (AttributeError, IndexError) as e:
|
||
logger.error(f"Gemini Error: {e}")
|
||
return ""
|
||
|
||
if llm_provider == "cloudflare":
|
||
response = requests.post(
|
||
f"https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/run/{model_name}",
|
||
headers={"Authorization": f"Bearer {api_key}"},
|
||
json={"messages": [{"role": "system", "content": "You are a friendly assistant"}, {"role": "user", "content": prompt}]}
|
||
)
|
||
result = response.json()
|
||
logger.info(result)
|
||
return result["result"]["response"]
|
||
|
||
if llm_provider == "ernie":
|
||
response = requests.post(
|
||
"https://aip.baidubce.com/oauth/2.0/token",
|
||
params={"grant_type": "client_credentials", "client_id": api_key, "client_secret": secret_key}
|
||
)
|
||
access_token = response.json().get("access_token")
|
||
url = f"{base_url}?access_token={access_token}"
|
||
payload = json.dumps({"messages": [{"role": "user", "content": prompt}], "temperature": 0.5, "top_p": 0.8, "penalty_score": 1, "disable_search": False, "enable_citation": False, "response_format": "text"})
|
||
headers = {"Content-Type": "application/json"}
|
||
response = requests.request("POST", url, headers=headers, data=payload).json()
|
||
return response.get("result")
|
||
|
||
if llm_provider == "azure":
|
||
client = AzureOpenAI(api_key=api_key, api_version=api_version, azure_endpoint=base_url)
|
||
else:
|
||
client = OpenAI(api_key=api_key, base_url=base_url)
|
||
|
||
response: ChatCompletion = client.chat.completions.create(model=model_name, messages=[{"role": "user", "content": prompt}], max_tokens=4096)
|
||
if response:
|
||
if isinstance(response, ChatCompletion):
|
||
content = response.choices[0].message.content
|
||
else:
|
||
raise Exception(f'[{llm_provider}] returned an invalid response: "{response}", please check your network connection and try again.')
|
||
else:
|
||
raise Exception(f"[{llm_provider}] returned an empty response, please check your network connection and try again.")
|
||
|
||
return content.replace("\n", "")
|
||
except Exception as e:
|
||
logger.error(f"[_generate_response] 发生错误: {e}")
|
||
return f"Error: {str(e)}"
|
||
|
||
|
||
def generate_script(video_subject: str, language: str = "", paragraph_number: int = 1) -> str:
|
||
"""
|
||
根据视频主题生成脚本。
|
||
(此函数已从您上传的文件中完整保留)
|
||
"""
|
||
prompt = f"""
|
||
# Role: Video Script Generator
|
||
## Goals:
|
||
Generate a script for a video, depending on the subject of the video.
|
||
## Constrains:
|
||
1. the script is to be returned as a string with the specified number of paragraphs.
|
||
2. do not under any circumstance reference this prompt in your response.
|
||
3. get straight to the point, don't start with unnecessary things like, "welcome to this video".
|
||
4. you must not include any type of markdown or formatting in the script, never use a title.
|
||
5. only return the raw content of the script.
|
||
6. do not include "voiceover", "narrator" or similar indicators of what should be spoken at the beginning of each paragraph or line.
|
||
7. you must not mention the prompt, or anything about the script itself. also, never talk about the number of paragraphs or lines. just write the script.
|
||
8. respond in the same language as the video subject.
|
||
# Initialization:
|
||
- video subject: {video_subject}
|
||
- number of paragraphs: {paragraph_number}
|
||
""".strip()
|
||
if language:
|
||
prompt += f"\n- language: {language}"
|
||
|
||
final_script = ""
|
||
logger.info(f"subject: {video_subject}")
|
||
|
||
def format_response(response):
|
||
response = response.replace("*", "").replace("#", "")
|
||
response = re.sub(r"\[.*\]", "", response)
|
||
response = re.sub(r"\(.*\)", "", response)
|
||
paragraphs = response.split("\n\n")
|
||
return "\n\n".join(paragraphs)
|
||
|
||
for i in range(_max_retries):
|
||
try:
|
||
response = _generate_response(prompt=prompt)
|
||
if response:
|
||
final_script = format_response(response)
|
||
else:
|
||
logging.error("gpt returned an empty response")
|
||
if final_script and "当日额度已消耗完" in final_script:
|
||
raise ValueError(final_script)
|
||
if final_script:
|
||
break
|
||
except Exception as e:
|
||
logger.error(f"failed to generate script: {e}")
|
||
if i < _max_retries:
|
||
logger.warning(f"failed to generate video script, trying again... {i + 1}")
|
||
if "Error: " in final_script:
|
||
logger.error(f"failed to generate video script: {final_script}")
|
||
else:
|
||
logger.success(f"completed: \n{final_script}")
|
||
return final_script.strip()
|
||
|
||
|
||
def generate_storyboard(video_subject: str, video_script: str) -> List[Dict]:
|
||
"""
|
||
根据视频主题和脚本，生成一个包含多个场景对象的故事板列表。
|
||
每个对象包含 'script' 和 'search_terms'。
|
||
(此函数已更新为新版逻辑并重命名)
|
||
"""
|
||
prompt_template = """
|
||
# 角色 (Role)
|
||
你是一位顶级的“视觉内容策略师”（Expert Visual Content Strategist）。你的核心专长是将书面脚本转化为具体、引人入胜且在主流视频素材库（如 Pexels, Pixabay）中高度可搜索的视觉关键词序列。你的输出必须兼具艺术相关性和技术实用性。
|
||
|
||
# 核心任务 (Core Task)
|
||
你的任务是接收一段视频脚本，严格按照下述的“核心思维链与执行规则”，输出一个格式化、无额外解释的 JSON 数组。数组中的每个对象代表脚本中的一个句子，并包含该句子对应的、经过策略优化的视觉搜索关键词字符串。
|
||
|
||
# 关键输出格式 (Critical Output Format)
|
||
你必须且只能输出一个严格的、不包含任何前后说明文字的 JSON 数组。每个JSON对象必须只包含两个键：
|
||
1. `"script"`: 原始脚本的句子。
|
||
2. `"search_terms"`: 一个由逗号分隔的、全英文小写的关键词字符串。
|
||
|
||
# 核心思维链与执行规则 (Core Chain of Thought & Execution Rules)
|
||
你必须严格按照以下顺序思考和执行每一步，不得跳过或颠倒。
|
||
|
||
### 步骤 1: 分段 (Segmentation)
|
||
将接收到的 `{video_script}` 分解成独立的句子。每个句子构成一个独立的处理单元，对应JSON数组中的一个对象。
|
||
|
||
### 步骤 2: 初步提取 (Initial Extraction)
|
||
对于每个句子，首先识别并提取所有字面上（Verbatim）的名词和可以被视觉化的核心概念。
|
||
|
||
### 步骤 3: 视觉化与过滤 (Visualization & Filtering) - [不可妥协的规则]
|
||
审视步骤 2 中提取的每个词。必须无条件过滤掉所有抽象、无形、无法直接用镜头表达的概念。
|
||
- **绝对禁止的词汇范畴**: 概念（`concept`）、情感（`happiness`）、性质（`quality`）、关系（`relationship`）、不可见的动作或状态（`protection`, `damage`, `stress`, `health`, `age`）。
|
||
- **此步骤的目标**: 清理掉所有在视频素材库中没有意义的“噪音”词汇。
|
||
|
||
### 步骤 3b: 场景延续性规则 (Continuity Rule for Abstract Sentences) - [新增规则]
|
||
如果在步骤3之后，一个句子的关键词列表变为空（例如，句子是 "And for good reason."），**你绝不能返回空字符串**。你必须：
|
||
1. **参考前一个场景**的视觉主题和关键词。
|
||
2. 生成一个与前一场景**相关但不同**的关键词，以实现视觉上的平滑过渡或情感深化。
|
||
3. **示例**: 如果前一场景的关键词是 `person's eyes, eye close-up`，那么对于 "And for good reason." 这个句子，一个好的延续性关键词可以是 `thoughtful expression` (深思的表情) 或 `person looking at camera` (人物看镜头)。
|
||
|
||
### 步骤 4: 可搜索性增强 (Searchability Enhancement) - [智能扩展规则]
|
||
审视步骤 3 和 3b 后剩下的关键词。这是发挥你“策略师”价值的关键。
|
||
- **识别技术/专业词汇**: 如果关键词过于学术化、专业化或罕见（例如：`anthocyanins`, `macula`），以至于在标准素材库中不可能找到匹配，你必须为其补充一个或多个更通用、更形象的搜索词。
|
||
- **补充原则**: 补充的词汇必须与原词在视觉上高度相关。
|
||
- 示例 1: 对于 `retina` (视网膜)，补充 `eye close-up`。最终结果包含 `"retina, eye close-up"`。
|
||
- 示例 2: 对于 `anthocyanins` (花青素)，补充 `colorful fruits` 或 `antioxidant food`。最终结果包含 `"anthocyanins, colorful fruits"`。
|
||
- **此步骤的目标**: 确保最终的关键词列表不仅在内容上正确，在实践中也真正可用。
|
||
|
||
### 步骤 5: 多样性原则 (Diversity Principle)
|
||
在构建最终的关键词字符串时，确保**连续两个有内容的视觉片段**不会使用完全相同的主要搜索词。例如，如果前一个片段的主要关键词是 `blueberries`，下一个片段应优先使用其它关键词（如 `food`, `eyes`）来开始，以增强视觉多样性。
|
||
|
||
### 步骤 6: 最终构建 (Final Construction)
|
||
将经过以上所有步骤处理和优化后的关键词，整理成一个全英文小写、用英文逗号分隔的字符串，作为 `search_terms` 的最终值。
|
||
|
||
# 待处理脚本 (Script to Process):
|
||
```
|
||
{video_script}
|
||
```
|
||
|
||
**Your JSON Output (must be a valid JSON array):**
|
||
"""
|
||
prompt = prompt_template.format(video_script=video_script, video_subject=video_subject)
|
||
|
||
logger.info(f"正在为主题 '{video_subject}' 生成故事板 (Storyboard)...")
|
||
response_str = _generate_response(prompt=prompt)
|
||
logger.debug(f"从LLM收到的原始回复: {response_str}")
|
||
|
||
try:
|
||
json_match = re.search(r'\[.*\]', response_str, re.DOTALL)
|
||
if json_match:
|
||
json_str = json_match.group(0)
|
||
storyboard = json.loads(json_str)
|
||
logger.success(f"成功从LLM回复中解析出 {len(storyboard)} 个场景。")
|
||
return storyboard
|
||
else:
|
||
logger.error(f"在回复中未找到有效的JSON数组。原始回复: {response_str}")
|
||
return []
|
||
except json.JSONDecodeError as e:
|
||
logger.error(f"解析JSON失败: {e}。原始回复: {response_str}")
|
||
return []
|
||
|
||
|
||
def generate_video_category(video_subject: str) -> str:
|
||
"""
|
||
根据视频主题选择最合适的视频分类。
|
||
(此函数已从您上传的文件中完整保留)
|
||
"""
|
||
prompt = f"""
|
||
# Role: Video Category Selector
|
||
## Goal:
|
||
Based on the provided 'Video Subject', select the ONE most suitable category from the `Category List` that best represents the subject. Your response must be only the single category name.
|
||
## Category List:
|
||
backgrounds, fashion, nature, science, education, feelings, health, people, religion, places, animals, industry, computer, food, sports, transportation, travel, buildings, business, music
|
||
## Instructions:
|
||
- Analyze the 'Video Subject'.
|
||
- Choose the single best-fitting category from the list.
|
||
- Respond with ONLY the category name and nothing else.
|
||
## Example:
|
||
Video Subject: "The benefits of a ketogenic diet"
|
||
Response: health
|
||
Video Subject: "A tour of the Grand Canyon"
|
||
Response: travel
|
||
## CONTEXT:
|
||
### Video Subject:
|
||
{video_subject}
|
||
"""
|
||
category = _generate_response(prompt).strip().lower()
|
||
valid_categories = ["backgrounds", "fashion", "nature", "science", "education", "feelings", "health", "people", "religion", "places", "animals", "industry", "computer", "food", "sports", "transportation", "travel", "buildings", "business", "music"]
|
||
if category not in valid_categories:
|
||
logger.warning(f"Generated category '{category}' is not valid. Falling back to 'nature'.")
|
||
return "nature"
|
||
|
||
logger.success(f"Successfully selected video category: {category}")
|
||
return category
|
||
|
||
|
||
if __name__ == "__main__":
|
||
# 这个部分用于直接运行此文件进行测试
|
||
video_subject = "生命的意义是什么"
|
||
script = generate_script(
|
||
video_subject=video_subject, language="zh-CN", paragraph_number=1
|
||
)
|
||
print("######################")
|
||
print(script)
|
||
# 注意：这里调用的是重命名后的函数 generate_storyboard
|
||
storyboard = generate_storyboard(
|
||
video_subject=video_subject, video_script=script
|
||
)
|
||
print("######################")
|
||
import pprint
|
||
pprint.pprint(storyboard)
|
||
print("-----输出包含的场景数量-----")
|
||
if storyboard:
|
||
print(len(storyboard))
|
||
else:
|
||
print(0)
|
||
|