AI Outline Writer - Added image generation and display for sections
This commit is contained in:
ajaysi
@@ -10,6 +10,7 @@ from typing import Dict, List, Optional
|
||||
from enum import Enum
|
||||
from dataclasses import dataclass
|
||||
from loguru import logger
|
||||
import json
|
||||
|
||||
from lib.gpt_providers.text_generation.main_text_generation import llm_text_gen
|
||||
from lib.gpt_providers.text_to_image_generation.main_generate_image_from_prompt import generate_image
|
||||
@@ -84,216 +85,196 @@ class BlogOutlineGenerator:
|
||||
self.outline = {}
|
||||
self.section_contents = {}
|
||||
|
||||
async def generate_outline(self, topic: str) -> Dict:
|
||||
"""Generate a comprehensive outline based on the topic and configuration."""
|
||||
def generate_outline(self, topic: str) -> Dict[str, List[str]]:
|
||||
"""Generate a blog outline based on the topic and configuration."""
|
||||
try:
|
||||
# Step 1: Generate main sections
|
||||
main_sections = await self._generate_main_sections(topic)
|
||||
# Create a focused prompt for outline generation
|
||||
prompt = f"""Generate a blog outline for topic: {topic}
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Target Audience: {self.config.target_audience}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
Word Count Target: {self.config.target_word_count}
|
||||
Main Sections: {self.config.num_main_sections}
|
||||
Subsections per Section: {self.config.num_subsections_per_section}
|
||||
|
||||
Requirements:
|
||||
- Create exactly {self.config.num_main_sections} main sections
|
||||
- Each section should have exactly {self.config.num_subsections_per_section} subsections
|
||||
- Focus on {self.config.content_type.value} content style
|
||||
- Target {self.config.target_audience} audience
|
||||
- Maintain {self.config.content_depth.value} depth
|
||||
- Follow {self.config.outline_style.value} style
|
||||
- Optimize for {self.config.target_word_count} words total
|
||||
|
||||
IMPORTANT: You must return a valid JSON object with main sections as keys and lists of subsections as values.
|
||||
Example format: {{"Section 1": ["Subsection 1.1", "Subsection 1.2"], "Section 2": ["Subsection 2.1", "Subsection 2.2"]}}
|
||||
Do not include any additional text or explanations, only the JSON object."""
|
||||
|
||||
# Get outline from LLM
|
||||
outline_json = llm_text_gen(prompt)
|
||||
|
||||
# Step 2: Generate subsections for each main section
|
||||
detailed_sections = await self._generate_subsections(main_sections)
|
||||
# Clean the response to ensure it's valid JSON
|
||||
outline_json = outline_json.strip()
|
||||
if not outline_json.startswith('{'):
|
||||
outline_json = outline_json[outline_json.find('{'):]
|
||||
if not outline_json.endswith('}'):
|
||||
outline_json = outline_json[:outline_json.rfind('}')+1]
|
||||
|
||||
# Step 3: Add introduction and conclusion if requested
|
||||
# Parse the outline
|
||||
try:
|
||||
outline = json.loads(outline_json)
|
||||
except json.JSONDecodeError as e:
|
||||
logger.error(f"JSON parsing error: {str(e)}")
|
||||
logger.error(f"Raw response: {outline_json}")
|
||||
# Fallback to a basic outline structure
|
||||
outline = {
|
||||
f"Section {i+1}": [f"Subsection {i+1}.{j+1}" for j in range(self.config.num_subsections_per_section)]
|
||||
for i in range(self.config.num_main_sections)
|
||||
}
|
||||
|
||||
# Add introduction and conclusion if configured
|
||||
if self.config.include_introduction:
|
||||
detailed_sections["Introduction"] = await self._generate_introduction(topic)
|
||||
outline = {"Introduction": ["Overview", "Importance", "What to Expect"]} | outline
|
||||
|
||||
if self.config.include_conclusion:
|
||||
detailed_sections["Conclusion"] = await self._generate_conclusion(topic)
|
||||
outline["Conclusion"] = ["Summary", "Key Takeaways", "Next Steps"]
|
||||
|
||||
# Step 4: Add FAQs if requested
|
||||
# Add FAQs if configured
|
||||
if self.config.include_faqs:
|
||||
detailed_sections["FAQs"] = await self._generate_faqs(topic)
|
||||
|
||||
# Step 5: Add resources if requested
|
||||
if self.config.include_resources:
|
||||
detailed_sections["Additional Resources"] = await self._generate_resources(topic)
|
||||
|
||||
self.outline = detailed_sections
|
||||
|
||||
# Step 6: Generate content for each section
|
||||
await self._generate_section_contents(topic)
|
||||
|
||||
return self.outline
|
||||
|
||||
except Exception as err:
|
||||
logger.error(f"Failed to generate outline: {err}")
|
||||
raise
|
||||
|
||||
async def _generate_main_sections(self, topic: str) -> List[str]:
|
||||
"""Generate main sections for the outline."""
|
||||
prompt = f"""Generate {self.config.num_main_sections} main sections for a {self.config.content_type.value}
|
||||
article about {topic} with the following characteristics:
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Target Word Count: {self.config.target_word_count}
|
||||
Target Audience: {self.config.target_audience}
|
||||
Style: {self.config.outline_style.value}
|
||||
|
||||
Additional Requirements:
|
||||
- Each section should contribute to the overall word count goal
|
||||
- Sections should flow logically
|
||||
- Include key concepts and important points
|
||||
- Consider SEO optimization
|
||||
- Keywords to include: {', '.join(self.config.keywords or [])}
|
||||
- Topics to exclude: {', '.join(self.config.exclude_topics or [])}
|
||||
|
||||
Please provide only the section titles, one per line."""
|
||||
|
||||
response = await llm_text_gen(prompt)
|
||||
return [section.strip() for section in response.split('\n') if section.strip()]
|
||||
|
||||
async def _generate_subsections(self, main_sections: List[str]) -> Dict[str, List[str]]:
|
||||
"""Generate subsections for each main section."""
|
||||
detailed_sections = {}
|
||||
|
||||
for section in main_sections:
|
||||
prompt = f"""Generate {self.config.num_subsections_per_section} subsections for the following section:
|
||||
{section}
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
|
||||
Each subsection should:
|
||||
- Be specific and focused
|
||||
- Support the main section's topic
|
||||
- Include key points to cover
|
||||
- Consider SEO optimization
|
||||
|
||||
Please provide only the subsection titles, one per line."""
|
||||
|
||||
response = await llm_text_gen(prompt)
|
||||
detailed_sections[section] = [sub.strip() for sub in response.split('\n') if sub.strip()]
|
||||
|
||||
return detailed_sections
|
||||
|
||||
async def _generate_introduction(self, topic: str) -> List[str]:
|
||||
"""Generate introduction subsections."""
|
||||
prompt = f"""Generate introduction subsections for an article about {topic}.
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
|
||||
The introduction should:
|
||||
- Hook the reader
|
||||
- Present the main topic
|
||||
- Outline what's to come
|
||||
- Set the tone for the article
|
||||
|
||||
Please provide only the subsection titles, one per line."""
|
||||
|
||||
response = await llm_text_gen(prompt)
|
||||
return [sub.strip() for sub in response.split('\n') if sub.strip()]
|
||||
|
||||
async def _generate_conclusion(self, topic: str) -> List[str]:
|
||||
"""Generate conclusion subsections."""
|
||||
prompt = f"""Generate conclusion subsections for an article about {topic}.
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
|
||||
The conclusion should:
|
||||
- Summarize key points
|
||||
- Provide final thoughts
|
||||
- Include a call to action
|
||||
- Leave a lasting impression
|
||||
|
||||
Please provide only the subsection titles, one per line."""
|
||||
|
||||
response = await llm_text_gen(prompt)
|
||||
return [sub.strip() for sub in response.split('\n') if sub.strip()]
|
||||
|
||||
async def _generate_faqs(self, topic: str) -> List[str]:
|
||||
"""Generate FAQ subsections."""
|
||||
prompt = f"""Generate FAQ subsections for an article about {topic}.
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
|
||||
The FAQs should:
|
||||
- Address common questions
|
||||
- Cover important aspects
|
||||
- Be relevant to the target audience
|
||||
- Include both basic and advanced questions
|
||||
|
||||
Please provide only the FAQ questions, one per line."""
|
||||
|
||||
response = await llm_text_gen(prompt)
|
||||
return [sub.strip() for sub in response.split('\n') if sub.strip()]
|
||||
|
||||
async def _generate_resources(self, topic: str) -> List[str]:
|
||||
"""Generate resource subsections."""
|
||||
prompt = f"""Generate resource subsections for an article about {topic}.
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
|
||||
The resources should:
|
||||
- Include relevant links
|
||||
- Suggest further reading
|
||||
- Provide tools or references
|
||||
- Include related materials
|
||||
|
||||
Please provide only the resource categories, one per line."""
|
||||
|
||||
response = await llm_text_gen(prompt)
|
||||
return [sub.strip() for sub in response.split('\n') if sub.strip()]
|
||||
|
||||
async def _generate_section_contents(self, topic: str):
|
||||
"""Generate content and images for each section."""
|
||||
for section, subsections in self.outline.items():
|
||||
if section not in ["Introduction", "Conclusion", "FAQs", "Additional Resources"]:
|
||||
# Generate content for the main section
|
||||
content_prompt = f"""Write a detailed section for a blog post about {topic}.
|
||||
Section Title: {section}
|
||||
Content Type: {self.config.content_type.value}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
Target Word Count: {self.config.target_word_count // self.config.num_main_sections}
|
||||
|
||||
Include:
|
||||
- Clear explanation of the main points
|
||||
- Examples and illustrations
|
||||
- Key takeaways
|
||||
- Relevant data or statistics
|
||||
"""
|
||||
|
||||
content = await llm_text_gen(content_prompt)
|
||||
|
||||
# Generate image prompt if images are enabled
|
||||
image_prompt = None
|
||||
image_path = None
|
||||
|
||||
if self.config.include_images:
|
||||
image_prompt = f"""Create a detailed image prompt for a blog section about {topic}.
|
||||
Section: {section}
|
||||
Content: {content[:200]}...
|
||||
Style: {self.config.image_style}
|
||||
"""
|
||||
# Generate topic-specific FAQs
|
||||
faq_prompt = f"""Generate 3 specific and relevant FAQ questions for a blog post about: {topic}
|
||||
|
||||
Content Type: {self.config.content_type.value}
|
||||
Target Audience: {self.config.target_audience}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
|
||||
Requirements:
|
||||
- Questions should be specific to the topic
|
||||
- Cover common concerns and important aspects
|
||||
- Be relevant to the target audience
|
||||
- Include both basic and advanced questions
|
||||
|
||||
Format: Return only a JSON array of 3 questions.
|
||||
Example format: ["Question 1?", "Question 2?", "Question 3?"]"""
|
||||
|
||||
try:
|
||||
faq_json = llm_text_gen(faq_prompt)
|
||||
faq_json = faq_json.strip()
|
||||
if not faq_json.startswith('['):
|
||||
faq_json = faq_json[faq_json.find('['):]
|
||||
if not faq_json.endswith(']'):
|
||||
faq_json = faq_json[:faq_json.rfind(']')+1]
|
||||
|
||||
image_prompt = await llm_text_gen(image_prompt)
|
||||
try:
|
||||
image_path = generate_image(
|
||||
image_prompt,
|
||||
title=section,
|
||||
description=content[:100],
|
||||
tags=self.config.keywords
|
||||
)
|
||||
except Exception as err:
|
||||
logger.warning(f"Failed to generate image for section {section}: {err}")
|
||||
|
||||
self.section_contents[section] = SectionContent(
|
||||
title=section,
|
||||
content=content,
|
||||
image_prompt=image_prompt,
|
||||
image_path=image_path
|
||||
)
|
||||
|
||||
faqs = json.loads(faq_json)
|
||||
outline["Frequently Asked Questions"] = faqs
|
||||
except Exception as e:
|
||||
logger.error(f"Error generating FAQs: {str(e)}")
|
||||
outline["Frequently Asked Questions"] = [
|
||||
f"Common Question about {topic} 1",
|
||||
f"Common Question about {topic} 2",
|
||||
f"Common Question about {topic} 3"
|
||||
]
|
||||
|
||||
# Add resources if configured
|
||||
if self.config.include_resources:
|
||||
outline["Additional Resources"] = [
|
||||
"Further Reading",
|
||||
"Tools and References",
|
||||
"Related Topics"
|
||||
]
|
||||
|
||||
return outline
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error generating outline: {str(e)}")
|
||||
return {}
|
||||
|
||||
def generate_section_content(self, section: str, subsections: List[str]) -> Optional[SectionContent]:
|
||||
"""Generate content for a section."""
|
||||
try:
|
||||
# Create a focused prompt for content generation
|
||||
prompt = f"""Generate content for section: {section}
|
||||
|
||||
Subsections: {', '.join(subsections)}
|
||||
Content Type: {self.config.content_type.value}
|
||||
Target Audience: {self.config.target_audience}
|
||||
Content Depth: {self.config.content_depth.value}
|
||||
Style: {self.config.outline_style.value}
|
||||
Word Count Target: {self.config.target_word_count // self.config.num_main_sections}
|
||||
|
||||
Requirements:
|
||||
- Write content for each subsection
|
||||
- Maintain {self.config.content_depth.value} depth
|
||||
- Target {self.config.target_audience} audience
|
||||
- Follow {self.config.outline_style.value} style
|
||||
- Optimize for {self.config.target_word_count // self.config.num_main_sections} words
|
||||
- Include relevant examples and data points
|
||||
- Use clear, engaging language
|
||||
|
||||
Format: Return only a JSON object with 'content' and 'image_prompt' fields.
|
||||
Example format: {{"content": "Section content here...", "image_prompt": "Image description here..."}}"""
|
||||
|
||||
# Get content from LLM
|
||||
content_json = llm_text_gen(prompt)
|
||||
content_data = json.loads(content_json)
|
||||
|
||||
# Generate image if configured
|
||||
image_path = None
|
||||
if self.config.include_images:
|
||||
image_path = self.generate_section_image(section)
|
||||
|
||||
return SectionContent(
|
||||
title=section,
|
||||
content=content_data["content"],
|
||||
image_prompt=content_data.get("image_prompt"),
|
||||
image_path=image_path
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error generating content for section {section}: {str(e)}")
|
||||
return None
|
||||
|
||||
def generate_section_image(self, section: str) -> Optional[str]:
|
||||
"""Generate an image for a section."""
|
||||
try:
|
||||
# Create a focused prompt for image generation
|
||||
prompt = f"""Generate an image prompt for section: {section}
|
||||
|
||||
Style: {self.config.image_style}
|
||||
Engine: {self.config.image_engine}
|
||||
Content Type: {self.config.content_type.value}
|
||||
Target Audience: {self.config.target_audience}
|
||||
|
||||
Requirements:
|
||||
- Create a {self.config.image_style} style image
|
||||
- Optimize for {self.config.image_engine} engine
|
||||
- Match {self.config.content_type.value} content type
|
||||
- Appeal to {self.config.target_audience} audience
|
||||
- Be visually engaging and relevant
|
||||
|
||||
Format: Return only a JSON object with an 'image_prompt' field.
|
||||
Example format: {{"image_prompt": "Detailed image description here..."}}"""
|
||||
|
||||
# Get image prompt from LLM
|
||||
prompt_json = llm_text_gen(prompt)
|
||||
prompt_data = json.loads(prompt_json)
|
||||
|
||||
# Generate image using the specified engine
|
||||
if self.config.image_engine == "Gemini-AI":
|
||||
image_path = generate_gemini_image(prompt_data["image_prompt"])
|
||||
elif self.config.image_engine == "Dalle3":
|
||||
image_path = generate_dalle_image(prompt_data["image_prompt"])
|
||||
else: # Stability-AI
|
||||
image_path = generate_stability_image(prompt_data["image_prompt"])
|
||||
|
||||
return image_path
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error generating image for section {section}: {str(e)}")
|
||||
return None
|
||||
|
||||
def to_markdown(self) -> str:
|
||||
"""Convert outline to markdown format with content and images."""
|
||||
markdown = f"# {self.outline.get('Introduction', [''])[0]}\n\n"
|
||||
|
||||
Reference in New Issue
Block a user