lizhilun/AI_NovelGenerator
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

分离提示词

Browse Source
This commit is contained in:
YILING0013
2025-02-06 15:54:25 +08:00
parent f0efb2947e
commit bf55e74ac9
2 changed files with 142 additions and 113 deletions
Download Patch File Download Diff File Expand all files Collapse all files
novel_generator.py
+96 -91
View File
@@ -5,7 +5,7 @@ import logging
import re
import time
import traceback
from typing import List, Optional
from typing import List, Optional, Tuple
# langchain 相关
from langchain_openai import ChatOpenAI, OpenAIEmbeddings
@@ -33,7 +33,8 @@ from prompt_definitions import (
chapter_blueprint_prompt,
summary_prompt,
update_character_state_prompt,
scene_dynamics_prompt
chapter_draft_prompt,
summarize_recent_chapters_prompt
)
# Ollama嵌入 (如使用Ollama时需要)
@@ -51,8 +52,12 @@ def remove_think_tags(text: str) -> str:
return re.sub(r'<think>.*?</think>', '', text, flags=re.DOTALL)
def debug_log(prompt: str, response_content: str):
logging.info(f"\n[######################################### Prompt #########################################]\n {prompt}\n")
logging.info(f"\n[######################################### Response #########################################]\n {response_content}\n")
logging.info(
f"\n[######################################### Prompt #########################################]\n{prompt}\n"
)
logging.info(
f"\n[######################################### Response #########################################]\n{response_content}\n"
)
def invoke_with_cleaning(model: ChatOpenAI, prompt: str) -> str:
"""通用封装：调用模型并移除 <think>...</think> 文本，记录日志后返回"""
@@ -132,9 +137,8 @@ def clear_vector_store(filepath: str) -> bool:
return False
try:
if os.path.exists(store_dir):
shutil.rmtree(store_dir)
logging.info(f"Vector store directory '{store_dir}' removed.")
shutil.rmtree(store_dir)
logging.info(f"Vector store directory '{store_dir}' removed.")
return True
except Exception as e:
logging.error(f"程序正在运行，无法删除，请在程序关闭后手动前往 {store_dir} 删除目录。\n {str(e)}")
@@ -222,7 +226,6 @@ def split_text_for_vectorstore(chapter_text: str,
return []
nltk.download('punkt', quiet=True)
nltk.download('punkt_tab', quiet=True)
sentences = nltk.sent_tokenize(chapter_text)
if not sentences:
return []
@@ -332,7 +335,7 @@ def get_relevant_context_from_vector_store(
return combined
# ========== 1) 生成总体架构 (Novel_architecture.txt) ==========
# ============ 1) 生成总体架构 (Novel_architecture.txt) ============
def Novel_architecture_generate(
api_key: str,
base_url: str,
@@ -408,7 +411,7 @@ def Novel_architecture_generate(
logging.info("Novel_architecture.txt has been generated successfully.")
# ========== 2) 生成章节蓝图 (Novel_directory.txt) ==========
# ============ 2) 生成章节蓝图 (Novel_directory.txt) ============
def Chapter_blueprint_generate(
api_key: str,
base_url: str,
@@ -467,31 +470,41 @@ def Chapter_blueprint_generate(
logging.info("Novel_directory.txt (chapter blueprint) has been generated successfully.")
# ============ 获取最近 N 章内容，生成短期摘要 ============
# ============ 工具：获取最近N章内容 ============
def get_last_n_chapters_text(chapters_dir: str, current_chapter_num: int, n: int = 3) -> List[str]:
"""
返回从 (current_chapter_num - n) 开始到 (current_chapter_num-1) 的章节文本列表。
若缺少文件，则对应位置为空字符串。
"""
texts = []
start_chap = max(1, current_chapter_num - n)
for c in range(start_chap, current_chapter_num):
chap_file = os.path.join(chapters_dir, f"chapter_{c}.txt")
if os.path.exists(chap_file):
text = read_file(chap_file).strip()
if text:
texts.append(text)
if len(texts) < n:
texts = [''] * (n - len(texts)) + texts
texts.append(text)
else:
texts.append("")
return texts
# ============ 新增函数：从合并文本中提炼「当前情节短期摘要」 & 「下一章关键字」 ============
def summarize_recent_chapters(
llm_model: str,
api_key: str,
base_url: str,
temperature: float,
chapters_text_list: List[str]
) -> str:
if not chapters_text_list:
return ""
if all(not txt.strip() for txt in chapters_text_list):
return "暂无摘要。"
) -> Tuple[str, str]:
"""
输入若干章节文本，合并后调用 summarize_recent_chapters_prompt，
返回 (short_summary, next_chapter_keywords)
如果解析失败，则返回(合并文本, "")
"""
combined_text = "\n".join(chapters_text_list).strip()
if not combined_text:
return ("", "")
model = ChatOpenAI(
model=llm_model,
@@ -500,57 +513,29 @@ def summarize_recent_chapters(
temperature=temperature
)
combined_text = "\n".join(chapters_text_list)
prompt = f"""你是一名资深长篇小说编辑，分析以下合并文本：\n\n {combined_text} \n\n
prompt = summarize_recent_chapters_prompt.format(combined_text=combined_text)
response_text = invoke_with_cleaning(model, prompt)
从中提取并预测下一章节的关键字[关键物品/人物/地点/事件/情节]
"""
# 简易解析
short_summary = ""
next_chapter_keywords = ""
summary_text = invoke_with_cleaning(model, prompt)
if not summary_text:
return (combined_text[:800] + "...") if len(combined_text) > 800 else combined_text
return summary_text
for line in response_text.splitlines():
line = line.strip()
if line.startswith("短期摘要:"):
short_summary = line.replace("短期摘要:", "").strip()
elif line.startswith("下一章关键字:"):
next_chapter_keywords = line.replace("下一章关键字:", "").strip()
# 如果解析失败，就把返回文本当作短期摘要
if not short_summary and not next_chapter_keywords:
short_summary = response_text
return (short_summary, next_chapter_keywords)
# ============ 剧情要点/冲突 ============
PLOT_ARCS_PROMPT = """\
下面是新生成的章节内容:
{chapter_text}
# ============ 3) 生成章节草稿（新版） ============
这里是已记录的剧情要点/未解决冲突(可能为空):
{old_plot_arcs}
请基于新的章节内容，提炼本章引入或延续的悬念、冲突、角色暗线等，将其合并到旧的剧情要点中。
若有新的冲突则添加，若有已解决/不再重要的冲突可标注或移除。
最终输出更新后的剧情要点列表，以帮助后续保持故事整体的一致性和悬念延续。
"""
def update_plot_arcs(
chapter_text: str,
old_plot_arcs: str,
api_key: str,
base_url: str,
model_name: str,
temperature: float
) -> str:
model = ChatOpenAI(
model=model_name,
api_key=api_key,
base_url=ensure_openai_base_url_has_v1(base_url),
temperature=temperature
)
prompt = PLOT_ARCS_PROMPT.format(
chapter_text=chapter_text,
old_plot_arcs=old_plot_arcs
)
arcs_text = invoke_with_cleaning(model, prompt)
if not arcs_text:
logging.warning("update_plot_arcs: No response or empty result.")
return old_plot_arcs
return arcs_text
# ========== 3) 生成章节草稿 ==========
def generate_chapter_draft(
api_key: str,
base_url: str,
@@ -571,12 +556,12 @@ def generate_chapter_draft(
embedding_retrieval_k: int = 2
) -> str:
"""
根据 scene_dynamics_prompt，生成本章草稿。
- novel_architecture 取自 Novel_architecture.txt
- blueprint 取自 Novel_directory.txt
- global_summary, character_state 分别取自全局摘要、角色状态文件
- 从向量库检索上下文（embedding_*参数）
- 用户还可以额外提供四个可选元素：核心人物、关键道具、空间坐标、时间压力
根据新的 chapter_draft_prompt，生成本章草稿。
- 首先获取最近3章文本 => 提炼短期摘要 & 下一章关键字
- 使用(短期摘要 + 下一章关键字) 拼成 query => 检索向量库
- 同时取上一章(或最后一个非空章节)末尾1500字作为 "前章片段"
- 组合所有信息后，调用模型生成章节草稿
- 最后保存到 chapters/chapter_{novel_number}.txt
"""
# 1) 读取相关文件
@@ -602,16 +587,37 @@ def generate_chapter_draft(
plot_twist_level = chapter_info["plot_twist_level"]
chapter_summary = chapter_info["chapter_summary"]
# 3) 取最近3章文本，拼成查询语句 => 用于向量库检索
chapters_dir = os.path.join(filepath, "chapters")
recent_3_texts = get_last_n_chapters_text(chapters_dir, novel_number, n=3)
merged_query_str = "回顾剧情：\n" + "\n".join(recent_3_texts) + "\n" + user_guidance
os.makedirs(chapters_dir, exist_ok=True)
# 4) 检索向量库上下文 (使用embedding_*参数)
# 3) 获取最近3章文本 => 提炼 (短期摘要 & 下一章关键字)
recent_3_texts = get_last_n_chapters_text(chapters_dir, novel_number, n=3)
short_summary, next_chapter_keywords = summarize_recent_chapters(
llm_model=model_name,
api_key=api_key,
base_url=base_url,
temperature=temperature,
chapters_text_list=recent_3_texts
)
# 4) 取上一章片段(或最后一个非空章节)的末尾1500字
previous_chapter_excerpt = ""
for text_block in reversed(recent_3_texts):
if text_block.strip():
# 找到最近一个非空章节
if len(text_block) > 1500:
previous_chapter_excerpt = text_block[-1500:]
else:
previous_chapter_excerpt = text_block
break
# 如果全为空，则 previous_chapter_excerpt 就是 ""
# 5) 构造向量检索查询： (短期摘要 + 下一章关键字)
retrieval_query = short_summary + " " + next_chapter_keywords
relevant_context = get_relevant_context_from_vector_store(
api_key=embedding_api_key,
base_url=embedding_url,
query=merged_query_str,
query=retrieval_query,
interface_format=embedding_interface_format,
embedding_model_name=embedding_model_name,
filepath=filepath,
@@ -620,9 +626,8 @@ def generate_chapter_draft(
if not relevant_context.strip():
relevant_context = "（无检索到的上下文）"
novel_setting_text = novel_architecture_text
prompt_text = scene_dynamics_prompt.format(
# 6) 组装 Prompt
prompt_text = chapter_draft_prompt.format(
novel_number=novel_number,
chapter_title=chapter_title,
chapter_role=chapter_role,
@@ -636,16 +641,17 @@ def generate_chapter_draft(
key_items=key_items,
scene_location=scene_location,
time_constraint=time_constraint,
user_guidance=user_guidance,
novel_setting=novel_setting_text,
novel_setting=novel_architecture_text,
global_summary=global_summary_text,
character_state=character_state_text
character_state=character_state_text,
previous_chapter_excerpt=previous_chapter_excerpt,
context_excerpt=relevant_context
)
# 合并检索到的上下文和用户指导
prompt_text += f"\n\n【检索到的上下文】\n{relevant_context}"
prompt_text += f"\n\n【章节额外指导】\n{user_guidance}\n"
# 7) 调用 LLM 生成章节正文
model = ChatOpenAI(
model=model_name,
api_key=api_key,
@@ -657,10 +663,8 @@ def generate_chapter_draft(
if not chapter_content.strip():
logging.warning("Generated chapter draft is empty.")
# 6) 写入 chapters
os.makedirs(chapters_dir, exist_ok=True)
# 8) 写入 chapters
chapter_file = os.path.join(chapters_dir, f"chapter_{novel_number}.txt")
clear_file_content(chapter_file)
save_string_to_txt(chapter_content, chapter_file)
@@ -668,7 +672,7 @@ def generate_chapter_draft(
return chapter_content
# ========== 4) 定稿章节 ==========
# ============ 4) 定稿章节 ============
def finalize_chapter(
novel_number: int,
word_number: int,
@@ -735,7 +739,7 @@ def finalize_chapter(
clear_file_content(character_state_file)
save_string_to_txt(new_char_state, character_state_file)
# 3) 更新向量库 (embedding相关)
# 3) 更新向量库
update_vector_store(
api_key=embedding_api_key,
base_url=embedding_url,
@@ -771,6 +775,7 @@ def enrich_chapter_text(
# ============ 导入外部知识文本到向量库 ============
def advanced_split_content(content: str,
similarity_threshold: float = 0.7,
max_length: int = 500) -> List[str]: