agent-recipes 0.0.5__py3-none-any.whl

agent_recipes/templates/ai-sentiment-analyzer/TEMPLATE.yaml

@@ -0,0 +1,28 @@
+name: ai-sentiment-analyzer
+version: "1.0.0"
+description: Analyze sentiment in text data
+author: praison
+license: Apache-2.0
+tags:
+  - data
+  - sentiment
+  - nlp
+
+requires:
+  tools:
+    - llm_tool
+  packages:
+    - openai
+  env:
+    - OPENAI_API_KEY
+
+cli:
+  command: praison recipes run ai-sentiment-analyzer
+  examples:
+    - praison recipes run ai-sentiment-analyzer <input>
+    - praison recipes run ai-sentiment-analyzer <input> --output ./output/
+    - praison recipes run ai-sentiment-analyzer <input> --dry-run
+
+safety:
+  dry_run_default: false
+  overwrites_files: true

agent_recipes/templates/ai-seo-optimizer/README.md

@@ -0,0 +1,59 @@
+# Seo Optimizer
+
+Optimize content for SEO with keywords
+
+## Installation
+
+```bash
+pip install agent-recipes
+```
+
+## CLI Usage
+
+```bash
+# Basic usage
+praison recipes run ai-seo-optimizer <input>
+
+# With output directory
+praison recipes run ai-seo-optimizer <input> --output ./output/
+
+# Dry run (check dependencies only)
+praison recipes run ai-seo-optimizer <input> --dry-run
+
+# Show recipe info
+praison recipes info ai-seo-optimizer
+
+# Check dependencies
+praison recipes doctor ai-seo-optimizer
+```
+
+## Python Usage
+
+```python
+from praisonaiagents import Agent
+from agent_recipes import load_recipe
+
+# Load and run recipe
+recipe = load_recipe("ai-seo-optimizer")
+result = recipe.run(input="<your-input>")
+print(result)
+```
+
+## Requirements
+
+### Packages
+- openai
+
+### Environment Variables
+- OPENAI_API_KEY
+
+### External Tools
+- None
+
+## Tags
+
+`web`, `seo`, `optimization`
+
+## License
+
+Apache-2.0

agent_recipes/templates/ai-seo-optimizer/TEMPLATE.yaml

@@ -0,0 +1,28 @@
+name: ai-seo-optimizer
+version: "1.0.0"
+description: Optimize content for SEO with keywords
+author: praison
+license: Apache-2.0
+tags:
+  - web
+  - seo
+  - optimization
+
+requires:
+  tools:
+    - llm_tool
+  packages:
+    - openai
+  env:
+    - OPENAI_API_KEY
+
+cli:
+  command: praison recipes run ai-seo-optimizer
+  examples:
+    - praison recipes run ai-seo-optimizer <input>
+    - praison recipes run ai-seo-optimizer <input> --output ./output/
+    - praison recipes run ai-seo-optimizer <input> --dry-run
+
+safety:
+  dry_run_default: false
+  overwrites_files: true

agent_recipes/templates/ai-signal-ranker/recipe.yaml

@@ -0,0 +1,54 @@
+name: ai-signal-ranker
+version: 1.0.0
+description: Rank news articles by novelty, velocity, and relevance signals
+author: PraisonAI
+tags:
+  - news
+  - ranking
+  - signals
+  - relevance
+
+config:
+  ranking_criteria:
+    - novelty
+    - velocity
+    - relevance
+    - engagement
+  weights:
+    novelty: 0.3
+    velocity: 0.25
+    relevance: 0.25
+    engagement: 0.2
+  output_format: json
+
+input:
+  type: object
+  properties:
+    articles:
+      type: array
+    context:
+      type: string
+      description: Context for relevance scoring
+
+output:
+  type: object
+  properties:
+    ranked_articles:
+      type: array
+    signals:
+      type: object
+
+requires:
+  env:
+    - OPENAI_API_KEY
+
+workflow:
+  agents:
+    - name: ranker
+      role: Signal Analyst
+      goal: Rank articles by importance signals
+      tools:
+        - calculate_novelty
+        - calculate_velocity
+        - calculate_relevance
+        - rank_articles

agent_recipes/templates/ai-signal-ranker/tools.py

@@ -0,0 +1,256 @@
+"""
+AI Signal Ranker Tools
+
+Tools for ranking news articles by:
+- Novelty: How new/unique is this information
+- Velocity: How fast is this story spreading
+- Relevance: How relevant to AI/tech audience
+- Engagement: Social signals (comments, shares)
+"""
+
+import logging
+import os
+from datetime import datetime, timezone
+from typing import Any, Dict, List, Optional
+
+logger = logging.getLogger(__name__)
+
+
+def calculate_novelty(
+    article: Dict[str, Any],
+    existing_topics: Optional[List[str]] = None,
+) -> float:
+    """
+    Calculate novelty score for an article.
+    
+    Args:
+        article: Article dictionary
+        existing_topics: List of known topics to compare against
+        
+    Returns:
+        Novelty score between 0 and 1
+    """
+    score = 0.5  # Base score
+    
+    title = article.get("title", "").lower()
+    content = article.get("content", "").lower()
+    text = f"{title} {content}"
+    
+    # Novelty indicators
+    novelty_keywords = [
+        "breakthrough", "first", "new", "revolutionary", "unprecedented",
+        "announces", "launches", "releases", "introduces", "unveils",
+        "discovers", "achieves", "beats", "surpasses", "record"
+    ]
+    
+    for keyword in novelty_keywords:
+        if keyword in text:
+            score += 0.1
+    
+    # Check against existing topics
+    if existing_topics:
+        is_new_topic = not any(topic.lower() in text for topic in existing_topics)
+        if is_new_topic:
+            score += 0.2
+    
+    return min(1.0, score)
+
+
+def calculate_velocity(
+    article: Dict[str, Any],
+    related_articles: Optional[List[Dict[str, Any]]] = None,
+) -> float:
+    """
+    Calculate velocity score based on how fast the story is spreading.
+    
+    Args:
+        article: Article dictionary
+        related_articles: Other articles on the same topic
+        
+    Returns:
+        Velocity score between 0 and 1
+    """
+    score = 0.3  # Base score
+    
+    # Check engagement metrics
+    comments = article.get("comments", 0)
+    article_score = article.get("score", 0)
+    
+    # Normalize engagement
+    if comments > 100:
+        score += 0.3
+    elif comments > 50:
+        score += 0.2
+    elif comments > 10:
+        score += 0.1
+    
+    if article_score > 500:
+        score += 0.2
+    elif article_score > 100:
+        score += 0.1
+    
+    # Check recency
+    published = article.get("published", "")
+    if published:
+        try:
+            pub_time = datetime.fromisoformat(published.replace("Z", "+00:00"))
+            hours_old = (datetime.now(timezone.utc) - pub_time).total_seconds() / 3600
+            
+            if hours_old < 2:
+                score += 0.2
+            elif hours_old < 6:
+                score += 0.1
+        except Exception:
+            pass
+    
+    # Check if multiple sources are covering it
+    if related_articles and len(related_articles) > 2:
+        score += 0.2
+    
+    return min(1.0, score)
+
+
+def calculate_relevance(
+    article: Dict[str, Any],
+    context: str = "AI technology news",
+) -> float:
+    """
+    Calculate relevance score for target audience.
+    
+    Args:
+        article: Article dictionary
+        context: Context for relevance (e.g., "AI technology news")
+        
+    Returns:
+        Relevance score between 0 and 1
+    """
+    score = 0.3  # Base score
+    
+    title = article.get("title", "").lower()
+    content = article.get("content", "").lower()
+    text = f"{title} {content}"
+    
+    # High relevance keywords for AI audience
+    high_relevance = [
+        "gpt", "llm", "openai", "anthropic", "claude", "gemini", "google ai",
+        "machine learning", "deep learning", "neural network", "transformer",
+        "ai agent", "rag", "fine-tuning", "training", "inference",
+        "chatgpt", "copilot", "midjourney", "stable diffusion", "dall-e"
+    ]
+    
+    medium_relevance = [
+        "artificial intelligence", "automation", "robotics", "nlp",
+        "computer vision", "speech recognition", "embedding", "vector",
+        "model", "benchmark", "dataset", "research", "paper"
+    ]
+    
+    for keyword in high_relevance:
+        if keyword in text:
+            score += 0.15
+    
+    for keyword in medium_relevance:
+        if keyword in text:
+            score += 0.05
+    
+    # Source quality bonus
+    quality_sources = ["arxiv", "openai", "anthropic", "google", "meta", "microsoft"]
+    source = article.get("source", "").lower()
+    if any(qs in source for qs in quality_sources):
+        score += 0.1
+    
+    return min(1.0, score)
+
+
+def calculate_engagement(article: Dict[str, Any]) -> float:
+    """
+    Calculate engagement score from social signals.
+    
+    Args:
+        article: Article dictionary
+        
+    Returns:
+        Engagement score between 0 and 1
+    """
+    score = 0.2  # Base score
+    
+    comments = article.get("comments", 0)
+    article_score = article.get("score", 0)
+    
+    # Logarithmic scaling for engagement
+    import math
+    
+    if comments > 0:
+        score += min(0.3, math.log10(comments + 1) / 10)
+    
+    if article_score > 0:
+        score += min(0.3, math.log10(article_score + 1) / 10)
+    
+    # Check for author reputation (if available)
+    author = article.get("author", "")
+    if author and len(author) > 0:
+        score += 0.1
+    
+    return min(1.0, score)
+
+
+def rank_articles(
+    articles: List[Dict[str, Any]],
+    weights: Optional[Dict[str, float]] = None,
+    context: str = "AI technology news",
+    top_n: Optional[int] = None,
+) -> Dict[str, Any]:
+    """
+    Rank articles by combined signal scores.
+    
+    Args:
+        articles: List of article dictionaries
+        weights: Custom weights for each signal
+        context: Context for relevance scoring
+        top_n: Return only top N articles
+        
+    Returns:
+        Dictionary with ranked articles and signal breakdown
+    """
+    if not articles:
+        return {"ranked_articles": [], "signals": {}}
+    
+    weights = weights or {
+        "novelty": 0.3,
+        "velocity": 0.25,
+        "relevance": 0.25,
+        "engagement": 0.2,
+    }
+    
+    ranked = []
+    
+    for article in articles:
+        signals = {
+            "novelty": calculate_novelty(article),
+            "velocity": calculate_velocity(article),
+            "relevance": calculate_relevance(article, context),
+            "engagement": calculate_engagement(article),
+        }
+        
+        # Calculate weighted score
+        total_score = sum(signals[k] * weights.get(k, 0.25) for k in signals)
+        
+        ranked.append({
+            **article,
+            "signals": signals,
+            "total_score": round(total_score, 3),
+        })
+    
+    # Sort by total score
+    ranked.sort(key=lambda x: x["total_score"], reverse=True)
+    
+    if top_n:
+        ranked = ranked[:top_n]
+    
+    return {
+        "ranked_articles": ranked,
+        "signals": {
+            "weights_used": weights,
+            "total_articles": len(ranked),
+            "avg_score": sum(a["total_score"] for a in ranked) / len(ranked) if ranked else 0,
+        }
+    }

agent_recipes/templates/ai-sitemap-generator/README.md

@@ -0,0 +1,59 @@
+# Sitemap Generator
+
+Generate XML sitemaps from URLs
+
+## Installation
+
+```bash
+pip install agent-recipes
+```
+
+## CLI Usage
+
+```bash
+# Basic usage
+praison recipes run ai-sitemap-generator <input>
+
+# With output directory
+praison recipes run ai-sitemap-generator <input> --output ./output/
+
+# Dry run (check dependencies only)
+praison recipes run ai-sitemap-generator <input> --dry-run
+
+# Show recipe info
+praison recipes info ai-sitemap-generator
+
+# Check dependencies
+praison recipes doctor ai-sitemap-generator
+```
+
+## Python Usage
+
+```python
+from praisonaiagents import Agent
+from agent_recipes import load_recipe
+
+# Load and run recipe
+recipe = load_recipe("ai-sitemap-generator")
+result = recipe.run(input="<your-input>")
+print(result)
+```
+
+## Requirements
+
+### Packages
+- requests
+
+### Environment Variables
+- None
+
+### External Tools
+- None
+
+## Tags
+
+`web`, `sitemap`, `seo`
+
+## License
+
+Apache-2.0

agent_recipes/templates/ai-sitemap-generator/TEMPLATE.yaml

@@ -0,0 +1,26 @@
+name: ai-sitemap-generator
+version: "1.0.0"
+description: Generate XML sitemaps from URLs
+author: praison
+license: Apache-2.0
+tags:
+  - web
+  - sitemap
+  - seo
+
+requires:
+  tools:
+    - web_tool
+  packages:
+    - requests
+
+cli:
+  command: praison recipes run ai-sitemap-generator
+  examples:
+    - praison recipes run ai-sitemap-generator <input>
+    - praison recipes run ai-sitemap-generator <input> --output ./output/
+    - praison recipes run ai-sitemap-generator <input> --dry-run
+
+safety:
+  dry_run_default: false
+  overwrites_files: true

agent_recipes/templates/ai-sitemap-scraper/README.md

@@ -0,0 +1,13 @@
+# AI Sitemap Scraper
+
+Scrape content from sitemap URLs and save as structured data.
+
+## Usage
+```bash
+praison run ai-sitemap-scraper https://example.com/sitemap.xml
+praison run ai-sitemap-scraper https://example.com/sitemap.xml --limit 50
+```
+
+## Output
+- `pages.json` - URL list with metadata
+- `content/` - Scraped content files

agent_recipes/templates/ai-sitemap-scraper/TEMPLATE.yaml

@@ -0,0 +1,41 @@
+name: ai-sitemap-scraper
+version: "1.0.0"
+description: Scrape content from sitemap URLs and save as structured data.
+author: praisonai
+license: Apache-2.0
+tags: [web, sitemap, scraping, extraction]
+
+requires:
+  tools: [shell_tool]
+  packages: [praisonai-tools, requests, beautifulsoup4]
+
+config:
+  input:
+    type: string
+    required: true
+  output:
+    type: string
+  limit:
+    type: integer
+    default: 100
+
+cli:
+  command: ai-sitemap-scraper
+  args:
+    - name: input
+      positional: true
+      required: true
+    - name: --output
+      short: -o
+    - name: --limit
+      default: "100"
+  examples:
+    - "praison run ai-sitemap-scraper https://example.com/sitemap.xml"
+
+outputs:
+  - name: pages.json
+    type: file
+  - name: content/
+    type: directory
+  - name: run.json
+    type: file

agent_recipes/templates/ai-slide-generator/README.md

@@ -0,0 +1,60 @@
+# Slide Generator
+
+Generate presentation slides from text/outline
+
+## Installation
+
+```bash
+pip install agent-recipes
+```
+
+## CLI Usage
+
+```bash
+# Basic usage
+praison recipes run ai-slide-generator <input>
+
+# With output directory
+praison recipes run ai-slide-generator <input> --output ./output/
+
+# Dry run (check dependencies only)
+praison recipes run ai-slide-generator <input> --dry-run
+
+# Show recipe info
+praison recipes info ai-slide-generator
+
+# Check dependencies
+praison recipes doctor ai-slide-generator
+```
+
+## Python Usage
+
+```python
+from praisonaiagents import Agent
+from agent_recipes import load_recipe
+
+# Load and run recipe
+recipe = load_recipe("ai-slide-generator")
+result = recipe.run(input="<your-input>")
+print(result)
+```
+
+## Requirements
+
+### Packages
+- openai
+- python-pptx
+
+### Environment Variables
+- OPENAI_API_KEY
+
+### External Tools
+- None
+
+## Tags
+
+`document`, `presentation`, `slides`
+
+## License
+
+Apache-2.0

agent_recipes/templates/ai-slide-generator/TEMPLATE.yaml

@@ -0,0 +1,29 @@
+name: ai-slide-generator
+version: "1.0.0"
+description: Generate presentation slides from text/outline
+author: praison
+license: Apache-2.0
+tags:
+  - document
+  - presentation
+  - slides
+
+requires:
+  tools:
+    - llm_tool
+  packages:
+    - openai
+    - python-pptx
+  env:
+    - OPENAI_API_KEY
+
+cli:
+  command: praison recipes run ai-slide-generator
+  examples:
+    - praison recipes run ai-slide-generator <input>
+    - praison recipes run ai-slide-generator <input> --output ./output/
+    - praison recipes run ai-slide-generator <input> --dry-run
+
+safety:
+  dry_run_default: false
+  overwrites_files: true

agent_recipes/templates/ai-slide-to-notes/README.md

@@ -0,0 +1,12 @@
+# AI Slide to Notes
+
+Convert presentation slides to structured notes.
+
+## Usage
+```bash
+praison run ai-slide-to-notes presentation.pdf
+```
+
+## Output
+- `notes.md` - Structured notes
+- `outline.json` - Presentation outline

agent_recipes/templates/ai-slide-to-notes/TEMPLATE.yaml

@@ -0,0 +1,37 @@
+name: ai-slide-to-notes
+version: "1.0.0"
+description: Convert presentation slides to structured notes.
+author: praisonai
+license: Apache-2.0
+tags: [slides, notes, presentation, extraction]
+
+requires:
+  tools: [shell_tool]
+  packages: [praisonai-tools]
+  env: [OPENAI_API_KEY]
+
+config:
+  input:
+    type: string
+    required: true
+  output:
+    type: string
+
+cli:
+  command: ai-slide-to-notes
+  args:
+    - name: input
+      positional: true
+      required: true
+    - name: --output
+      short: -o
+  examples:
+    - "praison run ai-slide-to-notes presentation.pdf"
+
+outputs:
+  - name: notes.md
+    type: file
+  - name: outline.json
+    type: file
+  - name: run.json
+    type: file