Medical-research-skills uspto-database

Source: https://github.com/aipoch/medical-research-skills

When to Use

• Find relevant US patents by keyword, inventor, assignee, CPC/IPC class, or date range for prior-art screening.
• Retrieve patent application/prosecution history (PEDS) to understand office actions, events, and timeline status.
• Check trademark status and documents via TSDR for clearance, monitoring, or portfolio review.
• Validate patent lifecycle signals such as maintenance-fee-related status or expiration indicators as part of diligence.
• Pull assignment/ownership-related information to support chain-of-title checks.

Key Features

• Patent search client for querying USPTO/PatentsView-style endpoints and returning matching patent identifiers/records.
• PEDS client for fetching examination history and event data for applications.
• TSDR access for trademark status and document retrieval.
• Script-oriented workflow designed for automation (batch queries, pipelines, and integration into internal tools).

Dependencies

• Python

  >=3.9

• requests >=2.31.0

• uspto-opendata-python >=0.3.0

Example Usage

# Example: search patents by abstract text and print results
# Prerequisite: set required API keys in environment variables as expected by your scripts/clients.
# For example (names may vary by implementation):
#   export USPTO_API_KEY="..."
#   export PATENTSVIEW_API_KEY="..."

from scripts.patent_search import PatentSearchClient

def main():
    client = PatentSearchClient()

    # Query example: search for patents whose abstract contains "AI"
    query = {"_text_all": {"patent_abstract": "AI"}}

    results = client.search_patents(query)
    print(results)

if __name__ == "__main__":
    main()

Implementation Details

• Workflow

  1. Search: Use

     scripts/patent_search.py

     to locate relevant patent numbers/records from a text or fielded query.
  2. History: Use

     scripts/peds_client.py

     to retrieve prosecution/examination history (events, dates, and status signals).
  3. Status checks: Use retrieved metadata to assess maintenance-fee/expiration-related status where available.

• Query model

  • Patent search is driven by a JSON query object (e.g., full-text constraints such as

    "_text_all"

    over fields like

    patent_abstract

    ).
  • Returned results are intended to be fed into downstream steps (e.g., PEDS lookups by application/patent identifiers).

• Authentication

  • API keys for USPTO and/or PatentsView are required and should be provided via environment variables or configuration used by the clients.

• Networking

  • HTTP requests are performed via

    requests

    ; ensure timeouts/retries are configured in the underlying clients if running large batches.