PDF Converter with Multiple Modes:

PDF Converter with Multiple Modes:#

This script demonstrates the use of the PDFConverter for generating PDFs in two different modes:

  • Template-Based PDF Generation: Utilize a YAML template to render dynamic content into a PDF.

  • Direct Prompt PDF Generation: Convert plain string prompts into PDFs without using a template.

The PromptSendingOrchestrator is used to handle the interaction with the PDFConverter and the mock TextTarget target system.

Key Features#

  1. Template-Based Generation:

    • Populate placeholders in a YAML-based template using dynamic data.

    • Convert the rendered content into a PDF using the PDFConverter.

  2. String-Based Generation:

    • Accept plain string prompts directly.

    • Generate PDFs from the provided string without using any template.

import pathlib

from pyrit.common import IN_MEMORY, initialize_pyrit
from pyrit.common.path import DATASETS_PATH
from pyrit.models import SeedPrompt
from pyrit.orchestrator import PromptSendingOrchestrator
from pyrit.prompt_converter import PDFConverter
from pyrit.prompt_normalizer import PromptConverterConfiguration
from pyrit.prompt_target import TextTarget

initialize_pyrit(memory_db_type=IN_MEMORY)

prompt_data = {
    "hiring_manager_name": "Jane Doe",
    "current_role": "AI Engineer",
    "company": "CyberDefense Inc.",
    "red_teaming_reason": "to creatively identify security vulnerabilities while enjoying free coffee",
    "applicant_name": "John Smith",
}

# Load the YAML template for the PDF generation
template_path = (
    pathlib.Path(DATASETS_PATH) / "prompt_converters" / "pdf_converters" / "red_teaming_application_template.yaml"
)
if not template_path.exists():
    raise FileNotFoundError(f"Template file not found: {template_path}")

# Load the SeedPrompt from the YAML file
prompt_template = SeedPrompt.from_yaml_file(template_path)

# Initialize the Azure OpenAI chat target (or mock target if not needed)
prompt_target = TextTarget()

# Initialize the PDFConverter
pdf_converter = PromptConverterConfiguration.from_converters(
    converters = [
        PDFConverter(
            prompt_template=prompt_template,
            font_type="Arial",
            font_size=12,
            page_width=210,
            page_height=297,
        )
    ]
)

# Define a single prompt for the orchestrator
prompts = str(prompt_data)

# Initialize the orchestrator
orchestrator = PromptSendingOrchestrator(
    objective_target=prompt_target,  # Target system (Azure OpenAI or other LLM target)
    request_converter_configurations=pdf_converter,  # Attach the PDFConverter
    verbose=False,  # Set to True for detailed logging
)

await orchestrator.run_attack_async(objective=prompts)  # type: ignore
memory = orchestrator.get_memory()

for entry in memory:
    print(entry)

{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: C:\git\PyRIT\dbdata\prompt-memory-entries\urls\1746483090633952.pdf
{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: C:\git\PyRIT\dbdata\prompt-memory-entries\urls\1746483090633952.pdf

Direct Prompt PDF Generation (No Template)#

# Define a simple string prompt (no templates)
prompt = "This is a simple test string for PDF generation. No templates here!"

# Initialize the TextTarget (mock target for testing)
prompt_target = TextTarget()

# Initialize the PDFConverter without a template
pdf_converter = PromptConverterConfiguration.from_converters(
    converters = [
        PDFConverter(
            prompt_template=None,  # No template provided
            font_type="Arial",
            font_size=12,
            page_width=210,
            page_height=297,
        )
    ]
)

# Initialize the orchestrator
orchestrator = PromptSendingOrchestrator(
    objective_target=prompt_target,
    request_converter_configurations=pdf_converter,
    verbose=False,
)

await orchestrator.run_attack_async(objective=prompt)  # type: ignore

memory = orchestrator.get_memory()

for entry in memory:
    print(entry)

{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: C:\git\PyRIT\dbdata\prompt-memory-entries\urls\1746483186494502.pdf
{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: C:\git\PyRIT\dbdata\prompt-memory-entries\urls\1746483186494502.pdf

Modify Existing PDF with Injection Items#

import tempfile
from pathlib import Path

import requests

from pyrit.common import IN_MEMORY, initialize_pyrit
from pyrit.orchestrator import PromptSendingOrchestrator
from pyrit.prompt_converter import PDFConverter
from pyrit.prompt_target import TextTarget

initialize_pyrit(memory_db_type=IN_MEMORY)

# This file isn't copied to our pipeline
url = "https://raw.githubusercontent.com/Azure/PyRIT/main/pyrit/datasets/prompt_converters/pdf_converters/fake_CV.pdf"

with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp_file:
    response = requests.get(url)
    tmp_file.write(response.content)

cv_pdf_path = Path(tmp_file.name)

# Define injection items
injection_items = [
    {"page": 0, "x": 50, "y": 700, "text": "Injected Text", "font_size": 12, "font": "Helvetica", "font_color": (255, 0, 0)},  # Red text
    {"page": 1, "x": 100, "y": 600, "text": "Confidential", "font_size": 10, "font": "Helvetica", "font_color": (0, 0, 255)}  # Blue text
]

# Define a simple string prompt (no templates)
prompt = "This is a simple test string for PDF generation. No templates here!"

# Initialize the TextTarget (mock target for testing)
prompt_target = TextTarget()

# Initialize the PDFConverter with the existing PDF and injection items
pdf_converter = PromptConverterConfiguration.from_converters(
    converters = [
        PDFConverter(
            prompt_template=None,  # No template provided
            font_type="Arial",
            font_size=12,
            page_width=210,
            page_height=297,
            existing_pdf=cv_pdf_path,  # Provide the existing PDF
            injection_items=injection_items,  # Provide the injection items
        )
    ]
)

orchestrator = PromptSendingOrchestrator(
    objective_target=prompt_target,
    request_converter_configurations=pdf_converter,
    verbose=False,
)

await orchestrator.run_attack_async(objective=prompt)  # type: ignore

memory = orchestrator.get_memory()

for entry in memory:
    print(entry)

[15:15:33][316][ai-red-team][INFO][Processing page 0 with 2 injection items.]
[15:15:33][326][ai-red-team][INFO][Processing page 1 with 2 injection items.]
[15:15:33][329][ai-red-team][INFO][Processing page 2 with 2 injection items.]
{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: C:\git\PyRIT\dbdata\prompt-memory-entries\urls\1746483333332260.pdf
{'__type__': 'TextTarget', '__module__': 'pyrit.prompt_target.text_target'}: user: C:\git\PyRIT\dbdata\prompt-memory-entries\urls\1746483333332260.pdf
Contents