diff --git a/.gitignore b/.gitignore index 47b5f9f..f2bce34 100644 --- a/.gitignore +++ b/.gitignore @@ -4,5 +4,5 @@ /build /dist /.vscode -__pycache__ -config.json \ No newline at end of file +/__pycache__ +config.json diff --git a/README.md b/README.md index e758021..9699b88 100644 --- a/README.md +++ b/README.md @@ -20,10 +20,22 @@ --- ## **2. 安装依赖** -**手动安装以下依赖**： + +**进入项目目录，执行**： ```bash -pip install openai langchain chromadb langchain_openai langchain_chroma langgraph typing_extensions langchain-community +pip install -r requirements.txt ``` +### 安装语句切分模型punkt（可选，默认程序运行后会自动加载） +**Python环境终端输入**: +```bash +python +import nltk +nltk.download('punkt') +``` + +等待下载完成(很小，下载很快的) + +**至此，环境配置完成。** --- @@ -82,6 +94,11 @@ pyinstaller --onefile --windowed main.py ``` 这样会在 `dist/` 目录下生成 `main.exe`（Windows）或 `main`（Linux/macOS）。 +或者使用提供的`main.spec`，执行以下打包指令： +```bash +pyinstaller main.spec +``` + --- ## **6. 使用指南** diff --git a/__pycache__/config_manager.cpython-310.pyc b/__pycache__/config_manager.cpython-310.pyc index 62b3d07..575cf01 100644 Binary files a/__pycache__/config_manager.cpython-310.pyc and b/__pycache__/config_manager.cpython-310.pyc differ diff --git a/__pycache__/consistency_checker.cpython-310.pyc b/__pycache__/consistency_checker.cpython-310.pyc index ab16c36..792d4c9 100644 Binary files a/__pycache__/consistency_checker.cpython-310.pyc and b/__pycache__/consistency_checker.cpython-310.pyc differ diff --git a/__pycache__/novel_generator.cpython-310.pyc b/__pycache__/novel_generator.cpython-310.pyc index b8aa27c..0c40b7e 100644 Binary files a/__pycache__/novel_generator.cpython-310.pyc and b/__pycache__/novel_generator.cpython-310.pyc differ diff --git a/__pycache__/prompt_definitions.cpython-310.pyc b/__pycache__/prompt_definitions.cpython-310.pyc index 3e98f89..44af9b9 100644 Binary files a/__pycache__/prompt_definitions.cpython-310.pyc and b/__pycache__/prompt_definitions.cpython-310.pyc differ diff --git a/__pycache__/ui.cpython-310.pyc b/__pycache__/ui.cpython-310.pyc index f3470fd..af2051e 100644 Binary files a/__pycache__/ui.cpython-310.pyc and b/__pycache__/ui.cpython-310.pyc differ diff --git a/__pycache__/utils.cpython-310.pyc b/__pycache__/utils.cpython-310.pyc index e08e35c..1304809 100644 Binary files a/__pycache__/utils.cpython-310.pyc and b/__pycache__/utils.cpython-310.pyc differ diff --git a/config_manager.py b/config_manager.py index 892fa90..c3b45f3 100644 --- a/config_manager.py +++ b/config_manager.py @@ -1,3 +1,5 @@ +# config_manager.py +# -*- coding: utf-8 -*- import json import os diff --git a/consistency_checker.py b/consistency_checker.py index 2cf6afe..cf34349 100644 --- a/consistency_checker.py +++ b/consistency_checker.py @@ -1,7 +1,5 @@ -""" -演示多Agent思路中的“审校Agent”，对最新章节进行简单的一致性或逻辑冲突检查。 -可根据需要进行扩展。 -""" +# consistency_checker.py +# -*- coding: utf-8 -*- from langchain_openai import ChatOpenAI CONSISTENCY_PROMPT = """\ @@ -28,7 +26,8 @@ def check_consistency( chapter_text: str, api_key: str, base_url: str, - model_name: str + model_name: str, + temperature: float = 0.3 ) -> str: """ 调用模型做简单的一致性检查。可扩展更多提示或校验规则。 @@ -43,9 +42,16 @@ def check_consistency( model=model_name, api_key=api_key, base_url=base_url, - temperature=0.3 + temperature=temperature ) + # 调试日志 + print("\n[ConsistencyChecker] Prompt >>>", prompt) + response = model.invoke(prompt) if not response: return "审校Agent无回复" + + # 调试日志 + print("[ConsistencyChecker] Response <<<", response.content.strip()) + return response.content.strip() diff --git a/main.py b/main.py index 2639413..52e3da3 100644 --- a/main.py +++ b/main.py @@ -1,3 +1,5 @@ +# main.py +# -*- coding: utf-8 -*- import tkinter as tk from ui import NovelGeneratorGUI diff --git a/main.spec b/main.spec index bfd34cb..57e1776 100644 --- a/main.spec +++ b/main.spec @@ -6,7 +6,19 @@ a = Analysis( pathex=[], binaries=[], datas=[], - hiddenimports=['typing_extensions', 'langchain-openai', 'langgraph', 'openai', 'chromadb','langchain-community','pydantic','pydantic.deprecated.decorator'], + hiddenimports=['typing_extensions', + 'langchain-openai', + 'langgraph', + 'openai', + 'chromadb', + 'nltk', + 'sentence_transformers', + 'scikit-learn', + 'langchain-community', + 'pydantic', + 'pydantic.deprecated.decorator', + 'chromadb.utils.embedding_functions.onnx_mini_lm_l6_v2' + ], hookspath=[], hooksconfig={}, runtime_hooks=[], @@ -21,7 +33,7 @@ exe = EXE( a.scripts, [], exclude_binaries=True, - name='AI_NovelGenerator_V1.0', + name='AI_NovelGenerator_V1.1', debug=False, bootloader_ignore_signals=False, strip=False, @@ -41,5 +53,5 @@ coll = COLLECT( strip=False, upx=True, upx_exclude=[], - name='AI_NovelGenerator_V1.0', + name='AI_NovelGenerator_V1.1', ) diff --git a/novel_generator.py b/novel_generator.py index 2ac9aee..35d07b6 100644 --- a/novel_generator.py +++ b/novel_generator.py @@ -1,5 +1,8 @@ +# novel_generator.py +# -*- coding: utf-8 -*- import os import logging +import re from typing import Dict, List, Optional try: from typing import TypedDict # Python 3.8+ 直接可用；若是3.7可改用 typing_extensions @@ -12,6 +15,12 @@ from langchain_openai import OpenAIEmbeddings from langchain_community.vectorstores import Chroma from langchain.docstore.document import Document +# +import nltk +import math +from sentence_transformers import SentenceTransformer +from sklearn.metrics.pairwise import cosine_similarity + from utils import ( read_file, append_text_to_file, clear_file_content, save_string_to_txt @@ -23,7 +32,7 @@ from prompt_definitions import ( chapter_outline_prompt, chapter_write_prompt ) -# ============ 日志配置（可选） ============ +# ============ 日志配置 ============ logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s") # ============ 向量检索相关函数（Chroma） ============ @@ -37,7 +46,7 @@ def init_vector_store(api_key: str, base_url: str, texts: List[str]) -> Chroma: """ embeddings = OpenAIEmbeddings( openai_api_key=api_key, - openai_api_base=base_url # <-- 这里用传进来的 base_url + openai_api_base=base_url ) documents = [Document(page_content=t) for t in texts] vectorstore = Chroma.from_documents( @@ -48,18 +57,16 @@ def init_vector_store(api_key: str, base_url: str, texts: List[str]) -> Chroma: vectorstore.persist() return vectorstore - def load_vector_store(api_key: str, base_url: str) -> Optional[Chroma]: """读取已存在的向量库。若不存在则返回 None。""" if not os.path.exists(VECTOR_STORE_DIR): return None embeddings = OpenAIEmbeddings( openai_api_key=api_key, - openai_api_base=base_url # <-- 使用 base_url + openai_api_base=base_url ) return Chroma(persist_directory=VECTOR_STORE_DIR, embedding_function=embeddings) - def update_vector_store(api_key: str, base_url: str, new_chapter: str) -> None: """将最新章节文本插入到向量库里，用于后续检索参考。若库不存在则初始化。""" store = load_vector_store(api_key, base_url) @@ -72,7 +79,6 @@ def update_vector_store(api_key: str, base_url: str, new_chapter: str) -> None: store.add_documents([new_doc]) store.persist() - def get_relevant_context_from_vector_store(api_key: str, base_url: str, query: str, k: int = 2) -> str: """ 从向量库中检索与 query 最相关的 k 条文本，拼接后返回。 @@ -86,7 +92,6 @@ def get_relevant_context_from_vector_store(api_key: str, base_url: str, query: s combined = "\n".join([d.page_content for d in docs]) return combined - # ============ 多步生成：设置 & 目录 ============ class OverallState(TypedDict): @@ -100,7 +105,6 @@ class OverallState(TypedDict): final_novel_setting: str novel_directory: str - def Novel_novel_directory_generate( api_key: str, base_url: str, @@ -109,7 +113,8 @@ def Novel_novel_directory_generate( genre: str, number_of_chapters: int, word_number: int, - filepath: str + filepath: str, + temperature: float = 0.7 ) -> None: """ 使用多步流程，生成 Novel_setting.txt 与 Novel_directory.txt 并保存到 filepath。 @@ -122,6 +127,7 @@ def Novel_novel_directory_generate( :param number_of_chapters: 章节数 :param word_number: 单章目标字数 :param filepath: 存放生成文件的目录路径 + :param temperature: 生成温度 """ # 确保文件夹存在 os.makedirs(filepath, exist_ok=True) @@ -129,9 +135,15 @@ def Novel_novel_directory_generate( model = ChatOpenAI( model=llm_model, api_key=api_key, - base_url=base_url + base_url=base_url, + temperature=temperature ) + def debug_log(prompt: str, response_content: str): + """在控制台打印或记录下每次Prompt与Response，[调试]""" + logging.info(f"\n[Prompt >>>] {prompt}\n") + logging.info(f"[Response <<<] {response_content}\n") + def generate_base_setting(state: OverallState) -> Dict[str, str]: prompt = set_prompt.format( topic=state["topic"], @@ -143,6 +155,7 @@ def Novel_novel_directory_generate( if not response: logging.warning("generate_base_setting: No response.") return {"novel_setting_base": ""} + debug_log(prompt, response.content) return {"novel_setting_base": response.content.strip()} def generate_character_setting(state: OverallState) -> Dict[str, str]: @@ -153,6 +166,7 @@ def Novel_novel_directory_generate( if not response: logging.warning("generate_character_setting: No response.") return {"character_setting": ""} + debug_log(prompt, response.content) return {"character_setting": response.content.strip()} def generate_dark_lines(state: OverallState) -> Dict[str, str]: @@ -163,6 +177,7 @@ def Novel_novel_directory_generate( if not response: logging.warning("generate_dark_lines: No response.") return {"dark_lines": ""} + debug_log(prompt, response.content) return {"dark_lines": response.content.strip()} def finalize_novel_setting(state: OverallState) -> Dict[str, str]: @@ -175,6 +190,7 @@ def Novel_novel_directory_generate( if not response: logging.warning("finalize_novel_setting: No response.") return {"final_novel_setting": ""} + debug_log(prompt, response.content) return {"final_novel_setting": response.content.strip()} def generate_novel_directory(state: OverallState) -> Dict[str, str]: @@ -186,6 +202,7 @@ def Novel_novel_directory_generate( if not response: logging.warning("generate_novel_directory: No response.") return {"novel_directory": ""} + debug_log(prompt, response.content) return {"novel_directory": response.content.strip()} # 构建状态图 @@ -196,7 +213,6 @@ def Novel_novel_directory_generate( graph.add_node("finalize_novel_setting", finalize_novel_setting) graph.add_node("generate_novel_directory", generate_novel_directory) - # 注意修正此处节点名称 graph.add_edge(START, "generate_base_setting") graph.add_edge("generate_base_setting", "generate_character_setting") graph.add_edge("generate_character_setting", "generate_dark_lines") @@ -229,22 +245,132 @@ def Novel_novel_directory_generate( filename_set = os.path.join(filepath, "Novel_setting.txt") filename_novel_directory = os.path.join(filepath, "Novel_directory.txt") - # 清理文本（去除多余 # 或 * 等） + # 清理文本（可根据需要去除多余字符） def clean_text(txt: str) -> str: return txt.replace('#', '').replace('*', '') final_novel_setting_cleaned = clean_text(final_novel_setting) final_novel_directory_cleaned = clean_text(final_novel_directory) - # 以追加方式保存；如果希望覆盖可改为 save_string_to_txt() append_text_to_file(final_novel_setting_cleaned, filename_set) append_text_to_file(final_novel_directory_cleaned, filename_novel_directory) logging.info("Novel settings and directory generated successfully.") - # ============ 生成章节（每章独立文件） ============ +CHINESE_NUM_MAP = { + '零': 0, '○': 0, '〇': 0, + '一': 1, '二': 2, '三': 3, '四': 4, '五': 5, + '六': 6, '七': 7, '八': 8, '九': 9, + '十': 10, '百': 100, '千': 1000, '万': 10000 +} + +def chinese_to_arabic(chinese_str: str) -> int: + """ + 只能处理到万(10000)以内的中文数字，正常小说章节应该够用了 + """ + total = 0 + current_unit = 1 # 记录当前单位 + tmp_val = 0 # 暂存本轮数字 + + for char in reversed(chinese_str): + if char in CHINESE_NUM_MAP: + val = CHINESE_NUM_MAP[char] + if val >= 10: + if val > current_unit: + # 如 100, 1000, 10000 + current_unit = val + else: + # 比如 “十二” -> 2 * 10 + 1 + # 如果 val <= current_unit, 那么相当于在这个单位下加 + total += tmp_val * val + tmp_val = 0 + else: + # 0~9 + tmp_val = tmp_val + val * current_unit + else: + # 非中文数字字符，视情况决定怎么处理，这里直接跳过 + pass + + total += tmp_val + return total + +def parse_chapter_title_from_directory(novel_directory_text: str, + novel_number: int, + range_size: int = 1) -> str: + """ + 从小说目录文本中，提取指定章节（以及前后几章）的目录信息。 + range_size=1，表示获取当前章节、前一章和后一章的目录信息(若存在)。 + 支持多种常见的章节格式。 + """ + + lines = novel_directory_text.splitlines() + + # 可以根据需求自行扩展，这里列举了几种常见的章节标题格式，如果模型实在不听话，可以适当调整 + # 每个pattern都应该捕获两个组： + # 1. chapter_num_str：章节数字(可能是中文也可能是阿拉伯数字) + # 2. chapter_title ：章节标题（.*） + patterns = [ + # 1) 第12章 标题 + r"^第\s*([\d]+)\s*章[:：]?\s*(.*)$", + # 2) 第十二章 标题（中文数字） + r"^第\s*([零○〇一二三四五六七八九十百千万]+)\s*章[:：]?\s*(.*)$", + # 3) Chapter 12 标题 + r"^Chapter\s+(\d+)\s*[:：]?\s*(.*)$", + # 4) Ch 12 标题 + r"^Ch\s+(\d+)\s*[:：]?\s*(.*)$", + # 5) 第12节 标题 + r"^第\s*([\d]+)\s*节[:：]?\s*(.*)$", + # 6) 第12话 标题 + r"^第\s*([\d]+)\s*话[:：]?\s*(.*)$", + # ... 更多模式 ... + ] + + # 用来存储匹配结果： chapter_num -> title + directory_map = {} + + for line in lines: + line = line.strip() + if not line: + continue + + # 依次尝试每一种pattern + matched = False + for pat in patterns: + match = re.match(pat, line, flags=re.IGNORECASE) + if match: + chapter_num_str = match.group(1) + chapter_title = match.group(2).strip() + + # 如果是中文数字，需要转换 + # 如果是阿拉伯数字，直接转 int 即可 + if re.match(r"^[零○〇一二三四五六七八九十百千万]+$", chapter_num_str): + chapter_num = chinese_to_arabic(chapter_num_str) + else: + chapter_num = int(chapter_num_str) + + directory_map[chapter_num] = chapter_title + matched = True + break + + # 如果已经匹配到其中一个pattern，就不需要继续匹配剩余pattern + if matched: + continue + + # 收集需要的章节范围 + chapters_info = [] + for cnum in range(novel_number - range_size, novel_number + range_size + 1): + if cnum in directory_map: + if cnum == novel_number: + chapters_info.append(f"【当前】第{cnum}章：{directory_map[cnum]}") + else: + chapters_info.append(f"第{cnum}章：{directory_map[cnum]}") + + if chapters_info: + return "\n".join(chapters_info) + return "" + def generate_chapter_with_state( novel_settings: str, novel_novel_directory: str, @@ -254,19 +380,22 @@ def generate_chapter_with_state( novel_number: int, filepath: str, word_number: int, - lastchapter: str + lastchapter: str, + user_guidance: str = "", + temperature: float = 0.7 ) -> str: """ 多步流程: 1) 更新/创建全局摘要 2) 更新/生成角色状态文档 3) 向量检索获取往期上下文 - 4) 大纲 -> 正文 - 5) 写入 chapter_{novel_number}.txt, 更新 last_chapter.txt - 6) 更新向量库 + 4) 从Novel_directory.txt中获取当前(和前后几章)的目录信息 + 5) 大纲 -> 正文（可结合用户给出的额外指导） + 6) 写入 chapter_{novel_number}.txt, 更新 last_chapter.txt + 7) 更新向量库 :param novel_settings: 最终的作品设定（字符串） - :param novel_novel_directory: 小说目录信息（此处暂时未使用，可根据需求做扩展） + :param novel_novel_directory: 小说目录信息 :param api_key: OpenAI API Key :param base_url: OpenAI Base URL :param model_name: LLM 模型名称 @@ -274,6 +403,8 @@ def generate_chapter_with_state( :param filepath: 文件存放的目录 :param word_number: 单章目标字数 :param lastchapter: 上一章内容（若为空字符串，表示无上一章） + :param user_guidance: 用户对当前章节的额外指导或想法 + :param temperature: 生成温度 :return: 本章生成的正文内容 """ # 确保文件夹存在 @@ -283,9 +414,15 @@ def generate_chapter_with_state( model=model_name, api_key=api_key, base_url=base_url, - temperature=0.9 + temperature=temperature ) + # 调试输出函数 + def debug_log(prompt: str, response_content: str): + """在控制台打印或记录下每次的 Prompt 与 Response，便于观察生成过程。""" + logging.info(f"\n[Prompt >>>]\n{prompt}\n") + logging.info(f"[Response <<<]\n{response_content}\n") + # --- 文件路径定义 --- chapters_dir = os.path.join(filepath, "chapters") os.makedirs(chapters_dir, exist_ok=True) @@ -308,6 +445,7 @@ def generate_chapter_with_state( if not response: logging.warning("update_global_summary: No response.") return old_summary + debug_log(prompt, response.content) return response.content.strip() if lastchapter.strip(): @@ -325,6 +463,7 @@ def generate_chapter_with_state( if not response: logging.warning("update_character_state: No response.") return old_state + debug_log(prompt, response.content) return response.content.strip() if lastchapter.strip(): @@ -334,53 +473,78 @@ def generate_chapter_with_state( # 3) 从向量库检索上下文 relevant_context = get_relevant_context_from_vector_store( - api_key, base_url, "回顾剧情", k=2 # <-- 多传一个 base_url + api_key, base_url, "回顾剧情", k=2 ) - # 4) 生成大纲 + # 4) 解析本章及前后章节目录信息 + this_and_related_chapters = parse_chapter_title_from_directory(novel_novel_directory, novel_number, range_size=1) + + # 5) 生成大纲 def outline_chapter( novel_setting: str, char_state: str, global_summary: str, chap_num: int, - extra_context: str + extra_context: str, + directory_hint: str, + user_guide: str ) -> str: - prompt = chapter_outline_prompt.format( + """ + 将目录提示以及用户额外指导内容一起放入 Prompt 中。 + """ + # 适度修改章节提纲提示词，以整合目录信息 & 用户指导 + outline_prompt = ( + chapter_outline_prompt + + "\n\n【目录参考】\n" + directory_hint + + "\n\n【用户指导】\n" + user_guide + ).format( novel_setting=novel_setting, character_state=char_state + "\n\n【历史上下文】\n" + extra_context, global_summary=global_summary, novel_number=chap_num ) - response = model.invoke(prompt) + + response = model.invoke(outline_prompt) if not response: logging.warning("outline_chapter: No response.") return "" + debug_log(outline_prompt, response.content) return response.content.strip() chap_outline = outline_chapter( - novel_settings, new_char_state, new_global_summary, novel_number, relevant_context + novel_settings, new_char_state, new_global_summary, novel_number, + relevant_context, this_and_related_chapters, user_guidance ) - # 5) 生成正文 + # 6) 生成正文 def write_chapter( novel_setting: str, char_state: str, global_summary: str, outline: str, wnum: int, - extra_context: str + extra_context: str, + directory_hint: str, + user_guide: str ) -> str: - prompt = chapter_write_prompt.format( + # 同理，整合目录信息和用户指导 + writing_prompt = ( + chapter_write_prompt + + "\n\n【目录参考】\n" + directory_hint + + "\n\n【用户指导】\n" + user_guide + ).format( novel_setting=novel_setting, character_state=char_state + "\n\n【历史上下文】\n" + extra_context, global_summary=global_summary, chapter_outline=outline, word_number=wnum ) - response = model.invoke(prompt) + + response = model.invoke(writing_prompt) if not response: logging.warning("write_chapter: No response.") return "" + debug_log(writing_prompt, response.content) return response.content.strip() chapter_content = write_chapter( @@ -389,7 +553,9 @@ def generate_chapter_with_state( new_global_summary, chap_outline, word_number, - relevant_context + relevant_context, + this_and_related_chapters, + user_guidance ) # 写入文件并更新记录 @@ -407,10 +573,117 @@ def generate_chapter_with_state( clear_file_content(global_summary_file) save_string_to_txt(new_global_summary, global_summary_file) - # 6) 更新向量检索库 + # 7) 更新向量检索库 update_vector_store(api_key, base_url, chapter_content) logging.info(f"Chapter {novel_number} generated successfully.") else: logging.warning(f"Chapter {novel_number} generation failed.") return chapter_content + +def import_knowledge_file(api_key: str, base_url: str, file_path: str) -> None: + """ + 将用户选定的文本文件导入到向量库，以便在写作时检索。 + 可以在UI中提供按钮来调用此函数。 + """ + + # 1. 检查文件路径是否有效 + if not os.path.exists(file_path): + logging.warning(f"知识库文件不存在: {file_path}") + return + + # 2. 读取文件内容 + content = read_file(file_path) + if not content.strip(): + logging.warning("知识库文件内容为空。") + return + + # 3. 对内容进行高级切分处理 + paragraphs = advanced_split_content(content) + + # 4. 加载或初始化向量存储 + store = load_vector_store(api_key, base_url) + if not store: + logging.info("Vector store does not exist. Initializing a new one for knowledge import...") + init_vector_store(api_key, base_url, paragraphs) + return + + # 5. 创建Document对象并更新到向量库 + docs = [Document(page_content=p) for p in paragraphs] + store.add_documents(docs) + store.persist() + logging.info("知识库文件已成功导入至向量库。") + + +def advanced_split_content(content: str, + similarity_threshold: float = 0.7, + max_length: int = 500) -> List[str]: + """ + 将文本先按句子切分，然后根据语义相似度进行合并，最后根据max_length进行二次切分。 + + :param content: 原始文本内容 + :param similarity_threshold: 相邻句子合并的语义相似度阈值，小于此值则会开启新的段落 + :param max_length: 每个段落的最大长度（按字符数计算，超过则进一步拆分） + :return: 切分好的段落列表 + """ + + # 1. 按句子切分 + nltk.download('punkt', quiet=True) # 确保 punkt 数据可用 + sentences = nltk.sent_tokenize(content) + + if not sentences: + return [] + + # 2. 加载 SentenceTransformer 模型，用于计算语义相似度 + model = SentenceTransformer('paraphrase-MiniLM-L6-v2') + embeddings = model.encode(sentences) + + # 3. 根据相邻句子的语义相似度合并段落 + merged_paragraphs = [] + current_sentences = [sentences[0]] + current_embedding = embeddings[0] + + for i in range(1, len(sentences)): + sim = cosine_similarity([current_embedding], [embeddings[i]])[0][0] + if sim >= similarity_threshold: + # 语义相似则并入当前段落 + current_sentences.append(sentences[i]) + # 更新current_embedding为合并后的平均值（可选，也可只采用最后一句做比较） + current_embedding = (current_embedding + embeddings[i]) / 2.0 + else: + # 语义相似度不足，另起一个新段落 + merged_paragraphs.append(" ".join(current_sentences)) + current_sentences = [sentences[i]] + current_embedding = embeddings[i] + + # 把最后一段加进去 + if current_sentences: + merged_paragraphs.append(" ".join(current_sentences)) + + # 4. 根据最大长度 max_length 做二次拆分，避免段落过长 + final_segments = [] + for para in merged_paragraphs: + # 如果段落长度超过max_length，进一步切分 + if len(para) > max_length: + sub_segments = split_by_length(para, max_length=max_length) + final_segments.extend(sub_segments) + else: + final_segments.append(para) + + # 返回最终段落列表 + return final_segments + + +def split_by_length(text: str, max_length: int = 500) -> List[str]: + """ + 将文本按照max_length进行拆分，以避免段落过长。 + 这里以字符数为单位进行简单的拆分，也可以改为按词数或token数等。 + """ + segments = [] + start_idx = 0 + while start_idx < len(text): + end_idx = min(start_idx + max_length, len(text)) + segment = text[start_idx:end_idx] + segments.append(segment.strip()) + start_idx = end_idx + return segments diff --git a/prompt_definitions.py b/prompt_definitions.py index b1d9af3..cd48375 100644 --- a/prompt_definitions.py +++ b/prompt_definitions.py @@ -1,3 +1,5 @@ +# prompt_definitions.py +# -*- coding: utf-8 -*- """ 集中存放所有提示词(Prompt)，便于统一管理和修改。 """ @@ -10,7 +12,7 @@ set_prompt = """\ 1. 小说名称、总字数走向（大致范围即可）。 2. 小说类型与基调（如：都市、穿越、战争等类型，以及轻松、爆笑、暗黑等基调）。 3. 写作风格（正式 / 轻松；细腻 / 简洁；抒情 / 客观；叙事视角等）。 -4. 整体世界观（时间背景、地理环境、社会结构、科技或魔法水平、重要历史传说或事件等）。 +4. 整体世界观（时间背景、地理环境、社会结构、科技或魔法水平、重要历史事件等）。 5. 核心内容梗概（可以使用常见叙事结构，如三幕结构、英雄之旅等）。 6. 初步的情节安排设想（主线、副线、交织等关键思路）。 7. 初步的人物关系与主要角色设定（角色定位、主要冲突或关系）。 @@ -58,7 +60,7 @@ novel_directory_prompt = """\ 根据以下最终《小说设定》： {final_novel_setting} 并按照下面的小说目录模板生成 {number_of_chapters} 章的目录，同时确保目录符合小说设定中的叙事结构、角色发展及暗线伏笔。 -目录模板： +目录模板（示例）： 第1章 ：< text > 第2章 ：< text > ... @@ -127,3 +129,4 @@ chapter_write_prompt = """\ 3. 可以着重描写人物心理、环境氛围等，以保证足够长度。 4. 在结尾部分保留一定悬念或剧情转折，为下一章做铺垫。 """ + diff --git a/requirements.txt b/requirements.txt index 693dcb1..d492909 100644 --- a/requirements.txt +++ b/requirements.txt @@ -4,3 +4,6 @@ langgraph openai chromadb langchain-community +sentence_transformers +scikit-learn +nltk \ No newline at end of file diff --git a/ui.py b/ui.py index 08b3768..4af9c40 100644 --- a/ui.py +++ b/ui.py @@ -1,3 +1,5 @@ +# ui.py +# -*- coding: utf-8 -*- import os import tkinter as tk from tkinter import ttk, filedialog, scrolledtext, messagebox @@ -7,7 +9,8 @@ from config_manager import load_config, save_config from utils import read_file from novel_generator import ( Novel_novel_directory_generate, - generate_chapter_with_state + generate_chapter_with_state, + import_knowledge_file ) from consistency_checker import check_consistency @@ -60,9 +63,9 @@ class NovelGeneratorGUI: self.chapter_result = scrolledtext.ScrolledText(chapter_frame, width=80, height=10, foreground="blue") self.chapter_result.grid(row=0, column=0, sticky="nsew") - def build_right_layout(self, ): + def build_right_layout(self): # 行列配置 - for i in range(12): + for i in range(15): self.right_frame.rowconfigure(i, weight=0) self.right_frame.columnconfigure(1, weight=1) @@ -81,57 +84,77 @@ class NovelGeneratorGUI: self.model_name_var = tk.StringVar(value=self.loaded_config.get("model_name", "gpt-4o-mini")) ttk.Entry(self.right_frame, textvariable=self.model_name_var, width=32).grid(row=2, column=1, padx=5, pady=5, sticky="w") - # 4. 主题(Topic) 多行输入 - ttk.Label(self.right_frame, text="主题(Topic):").grid(row=3, column=0, padx=5, pady=5, sticky="ne") + # 4. Temperature + ttk.Label(self.right_frame, text="Temperature:").grid(row=3, column=0, padx=5, pady=5, sticky="e") + self.temperature_var = tk.DoubleVar(value=self.loaded_config.get("temperature", 0.7)) + self.temp_value_label = ttk.Label(self.right_frame, text=f"{self.temperature_var.get():.2f}") + self.temp_value_label.grid(row=3, column=2, padx=5, pady=5, sticky="w") + + temp_scale = ttk.Scale(self.right_frame, from_=0.0, to=1.0, orient=tk.HORIZONTAL, variable=self.temperature_var) + temp_scale.grid(row=3, column=1, padx=5, pady=5, sticky="we") + def update_temp_label(*args): + self.temp_value_label.config(text=f"{self.temperature_var.get():.2f}") + self.temperature_var.trace("w", update_temp_label) + + # 5. 主题(Topic) 多行输入 + ttk.Label(self.right_frame, text="主题(Topic):").grid(row=4, column=0, padx=5, pady=5, sticky="ne") self.topic_text = scrolledtext.ScrolledText(self.right_frame, width=32, height=4) - self.topic_text.grid(row=3, column=1, padx=5, pady=5, sticky="w") + self.topic_text.grid(row=4, column=1, padx=5, pady=5, sticky="w") topic_default = self.loaded_config.get("topic", "") if topic_default: self.topic_text.insert(tk.END, topic_default) - # 5. 类型(Genre) - ttk.Label(self.right_frame, text="类型(Genre):").grid(row=4, column=0, padx=5, pady=5, sticky="e") + # 6. 类型(Genre) + ttk.Label(self.right_frame, text="类型(Genre):").grid(row=5, column=0, padx=5, pady=5, sticky="e") self.genre_var = tk.StringVar(value=self.loaded_config.get("genre", "玄幻")) - ttk.Entry(self.right_frame, textvariable=self.genre_var, width=32).grid(row=4, column=1, padx=5, pady=5, sticky="w") + ttk.Entry(self.right_frame, textvariable=self.genre_var, width=32).grid(row=5, column=1, padx=5, pady=5, sticky="w") - # 6. 章节数 - ttk.Label(self.right_frame, text="章节数:").grid(row=5, column=0, padx=5, pady=5, sticky="e") + # 7. 章节数 + ttk.Label(self.right_frame, text="章节数:").grid(row=6, column=0, padx=5, pady=5, sticky="e") self.num_chapters_var = tk.IntVar(value=self.loaded_config.get("num_chapters", 10)) - ttk.Entry(self.right_frame, textvariable=self.num_chapters_var, width=8).grid(row=5, column=1, padx=5, pady=5, sticky="w") + ttk.Entry(self.right_frame, textvariable=self.num_chapters_var, width=8).grid(row=6, column=1, padx=5, pady=5, sticky="w") - # 7. 每章字数 - ttk.Label(self.right_frame, text="每章字数:").grid(row=6, column=0, padx=5, pady=5, sticky="e") + # 8. 每章字数 + ttk.Label(self.right_frame, text="每章字数:").grid(row=7, column=0, padx=5, pady=5, sticky="e") self.word_number_var = tk.IntVar(value=self.loaded_config.get("word_number", 3000)) - ttk.Entry(self.right_frame, textvariable=self.word_number_var, width=8).grid(row=6, column=1, padx=5, pady=5, sticky="w") + ttk.Entry(self.right_frame, textvariable=self.word_number_var, width=8).grid(row=7, column=1, padx=5, pady=5, sticky="w") - # 8. 文件保存路径 - ttk.Label(self.right_frame, text="保存路径:").grid(row=7, column=0, padx=5, pady=5, sticky="e") + # 9. 文件保存路径 + ttk.Label(self.right_frame, text="保存路径:").grid(row=8, column=0, padx=5, pady=5, sticky="e") self.filepath_var = tk.StringVar(value=self.loaded_config.get("filepath", "")) - ttk.Entry(self.right_frame, textvariable=self.filepath_var, width=32).grid(row=7, column=1, padx=5, pady=5, sticky="w") - ttk.Button(self.right_frame, text="浏览...", command=self.browse_folder).grid(row=7, column=2, padx=5, pady=5, sticky="w") + ttk.Entry(self.right_frame, textvariable=self.filepath_var, width=32).grid(row=8, column=1, padx=5, pady=5, sticky="w") + ttk.Button(self.right_frame, text="浏览...", command=self.browse_folder).grid(row=8, column=2, padx=5, pady=5, sticky="w") # 保存/加载配置按钮 config_frame = ttk.Frame(self.right_frame) - config_frame.grid(row=8, column=1, sticky="w") + config_frame.grid(row=9, column=1, sticky="w") ttk.Button(config_frame, text="保存配置", command=self.save_config_btn).grid(row=0, column=0, padx=5) ttk.Button(config_frame, text="加载配置", command=self.load_config_btn).grid(row=0, column=1, padx=5) - # 按钮区域 - row_base = 9 - ttk.Label(self.right_frame, text="章节号:").grid(row=row_base, column=0, sticky="e") + # 10. 章节号 + ttk.Label(self.right_frame, text="章节号:").grid(row=10, column=0, sticky="e") self.chapter_num_var = tk.IntVar(value=1) - ttk.Entry(self.right_frame, textvariable=self.chapter_num_var, width=6).grid(row=row_base, column=1, padx=5, pady=5, sticky="w") + ttk.Entry(self.right_frame, textvariable=self.chapter_num_var, width=6).grid(row=10, column=1, padx=5, pady=5, sticky="w") + # 11. “用户指导” 多行输入 + ttk.Label(self.right_frame, text="本章指导:").grid(row=11, column=0, padx=5, pady=5, sticky="ne") + self.user_guide_text = scrolledtext.ScrolledText(self.right_frame, width=32, height=4) + self.user_guide_text.grid(row=11, column=1, padx=5, pady=5, sticky="w") + + # 按钮区域 + row_base = 12 self.btn_generate_full = ttk.Button(self.right_frame, text="1. 生成设定 & 目录", command=self.generate_full_novel) - self.btn_generate_full.grid(row=row_base+1, column=0, columnspan=2, padx=5, pady=5, sticky="ew") + self.btn_generate_full.grid(row=row_base, column=0, columnspan=2, padx=5, pady=5, sticky="ew") self.btn_generate_chapter = ttk.Button(self.right_frame, text="2. 生成单章(含角色状态)", command=self.generate_chapter_text) - self.btn_generate_chapter.grid(row=row_base+2, column=0, columnspan=2, padx=5, pady=5, sticky="ew") + self.btn_generate_chapter.grid(row=row_base+1, column=0, columnspan=2, padx=5, pady=5, sticky="ew") - # 可选：添加一个“一致性审校”按钮 self.btn_check_consistency = ttk.Button(self.right_frame, text="3. 一致性审校", command=self.do_consistency_check) - self.btn_check_consistency.grid(row=row_base+3, column=0, columnspan=2, padx=5, pady=5, sticky="ew") + self.btn_check_consistency.grid(row=row_base+2, column=0, columnspan=2, padx=5, pady=5, sticky="ew") + # 增加一个按钮来导入自定义知识库文件 + self.btn_import_knowledge = ttk.Button(self.right_frame, text="导入知识库", command=self.import_knowledge_handler) + self.btn_import_knowledge.grid(row=row_base+3, column=0, columnspan=2, padx=5, pady=5, sticky="ew") # -------------- 配置管理 -------------- def load_config_btn(self): @@ -140,6 +163,7 @@ class NovelGeneratorGUI: self.api_key_var.set(cfg.get("api_key", "")) self.base_url_var.set(cfg.get("base_url", "")) self.model_name_var.set(cfg.get("model_name", "")) + self.temperature_var.set(cfg.get("temperature", 0.7)) self.genre_var.set(cfg.get("genre", "")) self.num_chapters_var.set(cfg.get("num_chapters", 10)) self.word_number_var.set(cfg.get("word_number", 3000)) @@ -158,6 +182,7 @@ class NovelGeneratorGUI: "api_key": self.api_key_var.get(), "base_url": self.base_url_var.get(), "model_name": self.model_name_var.get(), + "temperature": self.temperature_var.get(), "topic": self.topic_text.get("1.0", tk.END).strip(), "genre": self.genre_var.get(), "num_chapters": self.num_chapters_var.get(), @@ -200,6 +225,7 @@ class NovelGeneratorGUI: num_chapters = self.num_chapters_var.get() word_number = self.word_number_var.get() filepath = self.filepath_var.get().strip() + temperature = self.temperature_var.get() if not filepath: messagebox.showwarning("警告", "请先选择保存文件路径") @@ -214,7 +240,8 @@ class NovelGeneratorGUI: genre=genre, number_of_chapters=num_chapters, word_number=word_number, - filepath=filepath + filepath=filepath, + temperature=temperature ) self.log("✅ 小说设定和目录生成完成。查看 Novel_setting.txt 和 Novel_directory.txt。") except Exception as e: @@ -226,7 +253,7 @@ class NovelGeneratorGUI: thread.start() def generate_chapter_text(self): - """多步生成章节：维护全局摘要+角色状态文档，向量检索辅助""" + """多步生成章节：维护全局摘要+角色状态文档，向量检索辅助，并结合目录信息和用户指导。""" def task(): self.disable_button(self.btn_generate_chapter) try: @@ -236,6 +263,7 @@ class NovelGeneratorGUI: novel_number = self.chapter_num_var.get() filepath = self.filepath_var.get().strip() word_number = self.word_number_var.get() + temperature = self.temperature_var.get() # 读取设定 & 目录 novel_settings_file = os.path.join(filepath, "Novel_setting.txt") @@ -253,6 +281,9 @@ class NovelGeneratorGUI: self.log("⚠️ 未找到 Novel_directory.txt，请先生成目录。") return + # 用户对当前章节的指导 + user_guidance = self.user_guide_text.get("1.0", tk.END).strip() + self.log(f"开始生成第{novel_number}章内容(含角色状态文档更新)...") chapter_text = generate_chapter_with_state( novel_settings=novel_settings, @@ -263,11 +294,13 @@ class NovelGeneratorGUI: novel_number=novel_number, filepath=filepath, word_number=word_number, - lastchapter=lastchapter + lastchapter=lastchapter, + user_guidance=user_guidance, + temperature=temperature ) if chapter_text: - self.log(f"✅ 第{novel_number}章内容生成完成。chapter.txt 已更新。") + self.log(f"✅ 第{novel_number}章内容生成完成。chapter_{novel_number}.txt 已更新。") self.chapter_result.delete("1.0", tk.END) self.chapter_result.insert(tk.END, chapter_text) self.chapter_result.see(tk.END) @@ -291,6 +324,7 @@ class NovelGeneratorGUI: base_url = self.base_url_var.get().strip() model_name = self.model_name_var.get().strip() filepath = self.filepath_var.get().strip() + temperature = self.temperature_var.get() # 读取关键文件 novel_settings_file = os.path.join(filepath, "Novel_setting.txt") @@ -315,7 +349,8 @@ class NovelGeneratorGUI: chapter_text=last_chapter_text, api_key=api_key, base_url=base_url, - model_name=model_name + model_name=model_name, + temperature=temperature ) self.log("审校结果：") self.log(result) @@ -327,3 +362,29 @@ class NovelGeneratorGUI: thread = threading.Thread(target=task) thread.start() + + def import_knowledge_handler(self): + """处理导入知识库文件。""" + selected_file = filedialog.askopenfilename( + title="选择要导入的知识库文件", + filetypes=[("Text Files", "*.txt"), ("All Files", "*.*")] + ) + if selected_file: + def task(): + self.disable_button(self.btn_import_knowledge) + try: + self.log(f"开始导入知识库文件: {selected_file}") + import_knowledge_file( + api_key=self.api_key_var.get().strip(), + base_url=self.base_url_var.get().strip(), + file_path=selected_file + ) + self.log("✅ 知识库文件导入完成。") + except Exception as e: + self.log(f"❌ 导入知识库时出错: {e}") + finally: + self.enable_button(self.btn_import_knowledge) + + thread = threading.Thread(target=task) + thread.start() + diff --git a/utils.py b/utils.py index 2c21cfe..96b2e0b 100644 --- a/utils.py +++ b/utils.py @@ -1,3 +1,5 @@ +# utils.py +# -*- coding: utf-8 -*- import os import json