AutoMediaPublisher.git

			@@ -1,7 +1,8 @@
			from docx import Document
			from docx.shared import Inches
			from docx.enum.text import WD_ALIGN_PARAGRAPH
			from docx.shared import Pt, Inches
			from docx.oxml.ns import qn
			from io import BytesIO
			from docx.enum.text import WD_ALIGN_PARAGRAPH
			import random
			import threading
			import re
			@@ -11,20 +12,23 @@
			from io import BytesIO
			from PIL import Image
			import google.generativeai as genai
			from auto_media_publisher.config.conf_base import GOOGLE_API_KEY,GOOGLE_MODEL,GOOGLE_CHAT_MIN_INTERVAL
			from auto_media_publisher.config.conf_base import GOOGLE_GEMINI_TEMPERATURE_MIN,GOOGLE_GEMINI_TEMPERATURE_MAX,GOOGLE_REWRITE_STYLE_MAP,GEMINI_PROMPT_INIT_DEFAULT
			from auto_media_publisher.config.conf_base import GOOGLE_API_KEY,GOOGLE_MODEL,GOOGLE_CHAT_MIN_INTERVAL,DEFAULT_PREFIX_MAP,DEFAULT_SURFIX_MAP
			from auto_media_publisher.config.conf_base import GOOGLE_GEMINI_TEMPERATURE_MIN,GOOGLE_GEMINI_TEMPERATURE_MAX,GOOGLE_GEMINI_TOPN_MIN,GOOGLE_GEMINI_TOPN_MAX,GOOGLE_REWRITE_STYLE_MAP,GEMINI_PROMPT_INIT_DEFAULT
			from auto_media_publisher.config.conf_base import GEMINI_PROMPT_TITLE,GEMINI_PROMPT_DESCRIPTION,GEMINI_PROMPT_TAG
			from auto_media_publisher.utils.logger import get_logger
			from auto_media_publisher.utils.utils_text2image_pollinations import Text2ImageGenerator

			class GeminiDocxRewriter:
			def __init__(self,rewrite_style="default" ):

			class DocxRewriterGemini:
			def __init__(self,rewrite_style="default" ,docx_author="我"):
			genai.configure(api_key=GOOGLE_API_KEY)
			self.temperature = round(random.uniform(GOOGLE_GEMINI_TEMPERATURE_MIN, GOOGLE_GEMINI_TEMPERATURE_MAX), 2)
			self.top_n = round(random.uniform(GOOGLE_GEMINI_TOPN_MIN, GOOGLE_GEMINI_TOPN_MAX), 2)
			self.model = genai.GenerativeModel(GOOGLE_MODEL,
			generation_config=genai.types.GenerationConfig(
			temperature=self.temperature,
			top_p=1.0
			top_p=self.top_n,
			top_k=random.choice([40, 50, 80, 100])
			)
			)
			if rewrite_style =="random":
			@@ -32,11 +36,16 @@
			else:
			self.rewrite_style = rewrite_style

			self.prefix_text = f"我是{docx_author}，{random.choice(DEFAULT_PREFIX_MAP)}"
			self.surfix_text = random.choice(DEFAULT_SURFIX_MAP)

			self.last_call_time = 0
			self.min_interval = GOOGLE_CHAT_MIN_INTERVAL # 限制最短请求间隔（秒）
			self.lock = threading.Lock() # 多线程并发时也安全
			self.logger = get_logger(self.__class__.__name__,"admin")
			self.text2imager = Text2ImageGenerator()
			self.text2imager = Text2ImageGenerator(width=900,height=383)

			self.summary_info = {}

			#@retry(stop=stop_after_attempt(3), wait=wait_fixed(2)) # 自动重试机制
			def safe_send_message(self, content: str):
			@@ -59,16 +68,11 @@
			rewritten = self._rewrite_blocks(blocks)
			self._save_to_new_docx(rewritten, output_path)

			summary_info = self.summarize_title_intro_keywords()
			print(summary_info)
			self.text2imager.generate_image(summary_info["summary"],output_path.replace(".docx",".jpg"))
			summary_info["cover_image"] = output_path.replace(".docx",".jpg")
			self.summary_info = self._summarize_title_intro_keywords()
			print(self.summary_info)
			self.text2imager.generate_image(self.summary_info["description"],output_path.replace(".docx",".png"))
			self.summary_info["path_cover"] = output_path.replace(".docx",".png")

			# 3. 写入磁盘
			with open(output_path.replace(".docx",".json"), "w", encoding="utf-8") as f:
			json.dump(summary_info, f, ensure_ascii=False, indent=4)

			return summary_info

			def extract_docx_blocks_by_order(self,docx_path):
			doc = Document(docx_path)
			@@ -146,28 +150,53 @@

			def _save_to_new_docx(self, blocks, output_path):
			doc = Document()
			# 设置默认样式
			style = doc.styles['Normal']
			font = style.font
			font.name = '楷体' # 西文字体设置为楷体
			font.size = Pt(10.5) # 小四字体
			style._element.rPr.rFonts.set(qn('w:eastAsia'), '楷体') # 中文字体设定

			# 前言
			para = doc.add_paragraph(self.prefix_text)
			para.paragraph_format.line_spacing = 1.5
			para.paragraph_format.space_before = Pt(0)
			para.paragraph_format.space_after = Pt(0)


			for block in blocks:
			if block["type"]=="text":
			if block["type"] == "text":
			for line in block["content"].split("\n"):
			doc.add_paragraph(line.strip())
			para = doc.add_paragraph(line.strip())
			para.paragraph_format.line_spacing = 1.5 # 行距1.5倍
			para.paragraph_format.space_before = Pt(0) # 段前0磅
			para.paragraph_format.space_after = Pt(0) # 段后0磅
			elif block["type"] == "image":
			image_data = block["content"]
			image_stream = BytesIO(image_data)
			paragraph = doc.add_paragraph()
			run = paragraph.add_run()
			run.add_picture(image_stream, width=Inches(5.5)) # 适当调整宽度
			run.add_picture(image_stream, width=Inches(5.5))
			paragraph.alignment = WD_ALIGN_PARAGRAPH.CENTER

			doc.save(output_path)

			# 结语
			para2 = doc.add_paragraph(self.surfix_text)
			para2.paragraph_format.line_spacing = 1.5
			para2.paragraph_format.space_before = Pt(12)
			para2.paragraph_format.space_after = Pt(0)

			def summarize_title_intro_keywords(self) -> dict:
			doc.save(output_path)


			def _summarize_title_intro_keywords(self) -> dict:
			title = self.safe_send_message(GEMINI_PROMPT_TITLE).text.strip()
			summary = self.safe_send_message(GEMINI_PROMPT_DESCRIPTION).text.strip()
			description = self.safe_send_message(GEMINI_PROMPT_DESCRIPTION).text.strip()
			keywords = self.safe_send_message(GEMINI_PROMPT_TAG).text.strip()

			return {
			"title": title,
			"summary": summary,
			"keywords": re.split(r'[、,\s]+', keywords.strip("。"))
			"title": title.splitlines()[0].strip()[:64],
			"description": description[:120],
			"tags": re.split(r'[、,\s]+', keywords.strip("。"))
			}