Cypris Patent XML Parser

Kevin Singh

A Python cli tool used to clean malformed XML + extracts attributes from the cleaned patent XML.

Overview

This project solves the challenge of extracting doc-number values from patent XML documents stored in Google Cloud Storage. The tool handles:

• Malformed XML (missing closing tags, incomplete attributes, missing > brackets)
• Priority-based extraction (e.g., EPO format first, then patent-office)
• Flexible attribute extraction with custom XPath queries

Features

• XML Parsing: Automatically fixes common XML malformations
  • Missing closing > brackets
  • Attributes without values (e.g., us-series-code becomes us-series-code="")
  • Placeholder ... removal
• Priority Ordering: Extract attributes in a specified priority order
• CLI Support: Command-line interface for easy integration
• Comprehensive Tests: Includes some unit tests to verify functionality

Project Structure

.
├── cypris_xml_parser.py           # Core XML cleaning/parsing logic
├── patent_attribute_extraction.py # Main extraction class with CLI
├── input_xmls/
│   └── test_case_1.txt            # Sample malformed XML file
├── tests/
│   ├── __init__.py
│   └── test_patent_attribute_extraction.py  # Unit tests
├── requirements.txt                # Python dependencies
├── setup.sh                        # Automated setup script
└── README.md                       # This file

Installation

# Create virtual environment
uv venv

# Activate virtual environment (Mac)
source .venv/bin/activate
# Activate virtual environment (Windows)
.venv\Scripts\activate.bat

# Install dependencies
uv pip install -r requirements.txt

Usage

Coding challenge solution: Run this one for the Cypris coding challenge assessment solution!

# Mac ONLY
python run_assesment_scenario.py 
# Windows ONLY
uv run run_assesment_scenario.py

Other ways to use the tool are below.

Command Line Interface

Basic usage (extract doc-numbers with default settings):

# Mac ONLY
python patent_attribute_extraction.py input_xmls/test_case_1.txt
# Windows ONLY
uv run patent_attribute_extraction.py input_xmls/test_case_1.txt

Extract multiple attributes:

# Max ONLY
python patent_attribute_extraction.py input_xmls/test_case_1.txt --attributes doc-number country lang
# Windows ONLY
uv run patent_attribute_extraction.py input_xmls/test_case_1.txt --attributes doc-number country lang

Custom XPath:

# Mac ONLY
python patent_attribute_extraction.py input_xmls/test_case_1.txt --xpath './/document-id[@format="epo"]'
# Windows ONLY
uv run patent_attribute_extraction.py input_xmls/test_case_1.txt --xpath './/document-id[@format="epo"]'

Help:

# Mac ONLY
python patent_attribute_extraction.py --help
# Windows ONLY
uv run patent_attribute_extraction.py --help

Python Code Example

If you want to use custom priority ordering!

from patent_attribute_extraction import PatentAttributeExtractor, DEFAULT_PRIORITY_ORDER

# Read XML file
with open('input_xmls/test_case_1.txt', 'r') as f:
    xml_data = f.read()

# Create extractor
extractor = PatentAttributeExtractor(xml_data)

# Extract doc-numbers with default priority (EPO first, then patent-office)
doc_numbers = extractor.extract_attributes(priority_ordering=DEFAULT_PRIORITY_ORDER)
print(doc_numbers)
# Output: ['999000888', '66667777']

# Extract multiple attributes (returns DataFrame)
result = extractor.extract_attributes(
    attributes=['doc-number', 'country', 'lang'],
    priority_ordering=DEFAULT_PRIORITY_ORDER
)
print(result)

# Custom priority order
custom_priority = {
    1: {
        'attribute': "load-source",
        "values": ["patent-office"]
    },
    2: {
        'attribute': "format",
        "values": ["epo"]
    }
}

doc_numbers = extractor.extract_attributes(priority_ordering=custom_priority)
print(doc_numbers)
# Output: ['66667777', '999000888']  # patent-office first now

Assumptions

The following assumptions are made about the XML structure:

1. XML may be malformed: Missing closing tags, incomplete tags, missing >
2. Attributes without values: Set to blank "" (e.g., us-series-code)
3. document-id elements: Have a format attribute that determines priority
4. Priority order: epo format first, then patent-office, then original, then docdb
5. doc-number elements: Nested within document-id elements
6. Multiple document-id elements: May exist within application-reference
7. Placeholder data: XML may contain ... representing omitted data (removed during parsing)
8. Implicit closing: If a new opening tag is encountered without a closing tag for the previous element, the previous element is assumed to be implicitly closed

Testing

# Run all tests
# Mac ONLY
python -m unittest discover tests -v
# Windows ONLY
uv run python -m unittest discover tests -v


# Run specific test file
# Mac ONLY
python -m unittest tests.test_patent_attribute_extraction -v
# Windows ONLY
uv run python -m unittest tests.test_patent_attribute_extraction -v

Example Output

Given the sample XML in input_xmls/test_case_1.txt:

# Mac ONLY
python patent_attribute_extraction.py input_xmls/test_case_1.txt
# Windows ONLY
uv run patent_attribute_extraction.py input_xmls/test_case_1.txt

Output:

Extracted attributes (in priority order):
==================================================
1. 999000888
2. 66667777
==================================================
Total: 2 values extracted

The output shows that 999000888 (EPO format) comes before 66667777 (patent-office format) based on the default priority ordering.