Skip to content
Fergus-MW
PR-Report
Public

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository
BranchesTags

Repository files navigation

PR Coverage Report Generator

A FastAPI service that generates AI-powered PR coverage reports for any topic using Google News RSS feeds.

Components

  1. Main API (main.py) - Single endpoint API server
  2. Report Generator (report_generator.py) - Core logic for structured PDF reports with links
  3. RSS Agent (rss_agent.py) - LangChain-powered agent with Google Gemini for RSS processing and AI categorization

Installation

  1. Install the required dependencies:
pip install -r requirements.txt

Cloud Deployment

Deploy to Google Cloud Run using Cloud Build:

# Quick deployment
./deploy.sh

# Or manually
gcloud builds submit --config cloudbuild.yaml

See DEPLOYMENT.md for detailed instructions.

  1. Set up your Google API key (required for RSS agent):
export GOOGLE_API_KEY="your-google-api-key-here"

Get your API key from: https://aistudio.google.com/app/apikey Or pass it directly in API requests.

Usage

  1. Start the API server:
python main.py

Or use uvicorn directly:

uvicorn main:app --reload --host 0.0.0.0 --port 8000
  1. The API will be available at http://localhost:8000

API Endpoints

POST /generate-report

Generates an AI-powered PR coverage report for any subject by automatically fetching and categorizing news articles.

Request Body:

{
    "subject": "Harry Styles",
    "max_articles": 20,
    "filename": "harry-styles-coverage.pdf",
    "language": "en-US",
    "country": "US"
}

Response: PDF file (application/pdf) with categorized news coverage

Example using curl:

# Generate Harry Styles coverage report
curl -X POST "http://localhost:8000/generate-report" \
     -H "Content-Type: application/json" \
     -d '{"subject": "Harry Styles", "max_articles": 15}' \
     --output harry-styles-coverage.pdf

# Generate AI technology coverage report
curl -X POST "http://localhost:8000/generate-report" \
     -H "Content-Type: application/json" \
     -d '{"subject": "artificial intelligence", "max_articles": 25}' \
     --output ai-coverage.pdf

# Generate Apple Inc coverage report for UK market
curl -X POST "http://localhost:8000/generate-report" \
     -H "Content-Type: application/json" \
     -d '{"subject": "Apple Inc", "language": "en-GB", "country": "GB"}' \
     --output apple-uk-coverage.pdf

Example using Python requests:

import requests

# Generate Harry Styles coverage report
response = requests.post(
    "http://localhost:8000/generate-report",
    json={
        "subject": "Harry Styles",
        "max_articles": 20,
        "filename": "harry-styles-coverage.pdf"
    }
)

if response.status_code == 200:
    with open("harry-styles-coverage.pdf", "wb") as f:
        f.write(response.content)
    print("Coverage report generated successfully!")

# Generate tech company coverage report
response = requests.post(
    "http://localhost:8000/generate-report",
    json={
        "subject": "Tesla",
        "max_articles": 15,
        "language": "en-US",
        "country": "US"
    }
)

if response.status_code == 200:
    with open("tesla-coverage.pdf", "wb") as f:
        f.write(response.content)
    print("Tesla coverage report generated!")

# Generate international coverage report
response = requests.post(
    "http://localhost:8000/generate-report",
    json={
        "subject": "climate change",
        "max_articles": 30,
        "language": "en-GB",
        "country": "GB"
    }
)

if response.status_code == 200:
    with open("climate-coverage-uk.pdf", "wb") as f:
        f.write(response.content)
    print("Climate change coverage report generated!")

Using Direct Functions

Generate coverage reports:

from rss_agent import create_report_from_topic

# Generate Harry Styles coverage report
json_path, pdf_path = create_report_from_topic(
    topic="Harry Styles",
    max_articles=15,
    subject_override="Harry Styles Coverage Report"
)
print(f"Generated: {pdf_path}")

# Generate technology coverage report
json_path, pdf_path = create_report_from_topic(
    topic="artificial intelligence",
    max_articles=25,
    subject_override="AI Technology Coverage"
)
print(f"Generated: {pdf_path}")

Process topics to JSON only:

from rss_agent import ArticleCategorizer

# Process any topic to categorized JSON
agent = ArticleCategorizer()
data = agent.process_topic_to_json(
    topic="space exploration",
    max_articles=10
)
print(f"Categorized into {len(data['sections'])} sections")

POST /analytics

Generates structured JSON analytics data with detailed metrics for any subject. Returns comprehensive analytics including tier counts, sentiment analysis, and article metadata instead of a PDF report.

Request Body:

{
    "clientId": "harry-styles",
    "includeInternational": false,
    "date": "2025-01-22"
}

Response Body:

{
    "report": {
        "clientId": "harry-styles",
        "clientName": "Harry Styles",
        "date": "2025-01-22",
        "generatedAt": "2025-01-22T15:00:00Z",
        "summary": {
            "topTierCount": 3,
            "midTierCount": 8,
            "blogCount": 12,
            "totalMentions": 23,
            "sentimentBreakdown": {
                "positive": 15,
                "neutral": 6,
                "negative": 2
            }
        },
        "articles": [
            {
                "id": 1,
                "title": "Harry Styles Announces New Tour Dates",
                "url": "https://people.com/harry-styles-tour-2025",
                "outlet": "People",
                "tier": "Mid",
                "focusType": "Headline",
                "estViews": 100000,
                "publishedAt": "2025-01-22T14:30:00Z",
                "sentiment": "positive",
                "summary": "Pop star announces highly anticipated world tour...",
                "includedInReport": true
            }
        ]
    }
}

Parameters:

  • clientId: Subject identifier (kebab-case format like "harry-styles")
  • includeInternational: Boolean flag for international vs US-only coverage
  • date: Report date in YYYY-MM-DD format

Analytics Features:

  • Tier Classification: Automatically categorizes outlets as Top/Mid/Blog tier
  • Sentiment Analysis: Keyword-based sentiment detection (positive/neutral/negative)
  • Focus Analysis: Determines if coverage is "Headline" focus vs "Mention"
  • View Estimation: Estimates reach based on outlet tier (500K/100K/25K)
  • Auto-calculated Summaries: Tier counts, total mentions, sentiment breakdown

Example using curl:

# Generate analytics for Harry Styles (US-only)
curl -X POST "http://localhost:8000/analytics" \
     -H "Content-Type: application/json" \
     -d '{"clientId": "harry-styles", "includeInternational": false, "date": "2025-01-22"}' \
     | jq '.'

# Generate international analytics for Apple Inc
curl -X POST "http://localhost:8000/analytics" \
     -H "Content-Type: application/json" \
     -d '{"clientId": "apple-inc", "includeInternational": true, "date": "2025-01-22"}' \
     | jq '.'

# Generate analytics for AI technology coverage
curl -X POST "http://localhost:8000/analytics" \
     -H "Content-Type: application/json" \
     -d '{"clientId": "artificial-intelligence", "includeInternational": false, "date": "2025-01-22"}' \
     | jq '.report.summary'

Example using Python requests:

import requests
import json

# Generate analytics for Tesla
response = requests.post(
    "http://localhost:8000/analytics",
    json={
        "clientId": "tesla",
        "includeInternational": false,
        "date": "2025-01-22"
    }
)

if response.status_code == 200:
    analytics = response.json()
    summary = analytics["report"]["summary"]
    
    print(f"Total mentions: {summary['totalMentions']}")
    print(f"Top tier: {summary['topTierCount']}")
    print(f"Mid tier: {summary['midTierCount']}")
    print(f"Blog tier: {summary['blogCount']}")
    print(f"Sentiment: {summary['sentimentBreakdown']}")
    
    # Process individual articles
    for article in analytics["report"]["articles"][:5]:
        print(f"- {article['title']} ({article['outlet']}, {article['tier']} tier)")

# Generate international coverage analytics
response = requests.post(
    "http://localhost:8000/analytics",
    json={
        "clientId": "climate-change",
        "includeInternational": true,
        "date": "2025-01-22"
    }
)

if response.status_code == 200:
    data = response.json()
    print(f"International coverage: {data['report']['summary']['totalMentions']} articles")

Additional Endpoints

GET /

Returns API information and supported topics.

GET /health

Health check endpoint showing service status.

Interactive Documentation

When the server is running, you can access the interactive API documentation at:

  • Swagger UI: http://localhost:8000/docs
  • ReDoc: http://localhost:8000/redoc

Features

AI-Powered Coverage Analysis

  • Automatic topic processing - Enter any subject and get comprehensive coverage
  • Google News integration - Fetches latest articles from Google News RSS
  • Smart categorization using Google Gemini 2.0 Flash AI
  • Media tier classification - Distinguishes top-tier, mid-tier, and low-tier sources
  • Coverage differentiation - Identifies headline coverage vs mentions
  • Multi-language support - Generate reports in different languages/regions

Professional PDF Output

  • Structured sections with clear headings and subheadings
  • Clickable links to original articles (blue links in PDF)
  • Source attribution with media tier and coverage type metadata
  • Professional formatting with proper spacing and typography
  • Custom filenames and automatic timestamping

System Architecture

  • Single endpoint design - One simple route for all coverage reports
  • Dynamic RSS URL generation - Builds Google News RSS URLs for any topic
  • AI-powered processing using LangChain + Google Gemini 2.0 Flash
  • Modular components - Separate RSS agent and report generator modules
  • Cloud-ready with Docker and Google Cloud Run deployment
  • Environment variable configuration for API keys and settings

Technical Features

  • FastAPI framework with automatic validation and documentation
  • Comprehensive error handling with detailed error messages
  • Automatic cleanup of temporary files
  • Health check endpoint for monitoring
  • Interactive API documentation at /docs
  • Multi-region support with language and country parameters

Example Usage

Quick Start with RSS

  1. Set your Google API key:

    export GOOGLE_API_KEY="your-key-here"

  2. Run the example:

    python example_rss_usage.py

  3. Or use the API:

    curl -X POST "http://localhost:8000/generate-report" \
     -H "Content-Type: application/json" \
     -d '{"subject": "Harry Styles", "max_articles": 15}' \
     --output harry-styles-coverage.pdf

Supported Topics

The service can generate coverage reports for any topic that appears in Google News:

People & Celebrities:

  • Harry Styles, Taylor Swift, Elon Musk, Joe Biden

Companies & Brands:

  • Apple, Tesla, Microsoft, Netflix, Google

Technology:

  • artificial intelligence, blockchain, quantum computing, ChatGPT

Events & Trends:

  • COP28, Olympics 2024, World Cup, climate change

Geographic Topics:

  • London news, California wildfires, Japan earthquake

Industries:

  • renewable energy, electric vehicles, space exploration

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages