SuperAPI AI Directory

ecommerce.patent-abstract-image-data

Design & creation

View repository

Updated 10-16 01:36

About this skill

Nexscope Patent Abstract Image

This skill guides you on how to retrieve abstract images (drawings) from the Nexscope patent database, helping users quickly obtain the illustrative figures associated with specific patents.

Core Concepts

Abstract images (abstract drawings) are the representative figures attached to a patent document's abstract section. They provide a quick visual overview of the invention. This tool queries the Nexscope patent database and returns download paths for these images.

Lookup logic: Provide at least one of a patent ID or a publication number. Each field accepts 1 to 100 values separated by English commas, up to 60,000 characters. If both fields are provided, patentId takes priority.

Parameter Guide

Parameter API Name Required Description Example
Patent ID patentId Conditionally (one of the two must be provided) Internal patent IDs; up to 100 English-comma-separated values, maximum 60,000 characters 5e6f7a8b9c,6f7a8b9c0d
Publication Number patentNumber Conditionally (one of the two must be provided) Patent publication/announcement numbers; up to 100 English-comma-separated values, maximum 60,000 characters CN115059423A,CN115059424A
  • At least one of patentId or patentNumber must be supplied.
  • If both are supplied, patentId takes precedence.
  • Pass at most 100 patent identifiers in one request. Use English commas without empty entries.

Credit use: This endpoint consumes credits. Keep the requested batch within the user's stated scope, and explain any additional calls before expanding that scope.

Response Fields

Field API Name Description
Patent ID patentId The internal patent identifier
Publication Number pn The publication/announcement number
Abstract Drawing Path abstractDrawingPath URL path to the abstract image file
Total total Total number of records returned
Cost Token costToken Tokens consumed by the query

Calling the Tool

  • API Endpoint: /zhihuiya/abstractImage (full parameters/responses/error codes in references/api.md)
  • Python Script: python scripts/patent_abstract_image_data.py '<JSON params>' [--inline]
  • Cost constraint: This tool consumes credits. Within the same session and same parameter combination, it defaults to a single call with a 24-hour local cache. Do not automatically retry with different keywords, pagination, or parameters on failure/empty results. Inform the user of additional credit consumption before continuing retrieval.

Output strategy (script default behavior):

  • Always write the full response to <cwd>/nexscope/<YYYY-MM-DD>/<session>/data/ecommerce.patent-abstract-image-data-<timestamp>.json (<cwd> is the working directory when the script executes, which in Claude Code is the current project directory; <session> is taken from the SESSION_ID environment variable, automatically grouped by user task; do not write to /tmp; error if the current directory is not writable)
  • Response body <= 8 KB: write to disk then print full JSON to stdout
  • Response body > 8 KB: write to disk then print only a summary to stdout (top-level fields, common counts like total/costToken, length of the largest list field + first 3 samples)
  • Add --inline to force full output to stdout (still writes to disk)

Reading data: Check the summary first to determine if it is sufficient. When specific fields are needed, use jq or ConvertFrom-Json to extract from the saved JSON file as needed, avoiding loading the entire JSON into context.

Authentication & Credits

If you encounter authentication or credit issues:

Error conditions

  • API Key not configured: The NEXSCOPE_API_KEY environment variable is not set.
  • HTTP 401 or 402 status code
  • Insufficient credits/balance: Response message indicates credit balance exhausted, quota exceeded, subscription expired, or recharge required.

Resolution steps

  1. Set the NEXSCOPE_API_KEY environment variable with a valid API key.
  2. If credits are insufficient, visit https://www.nexscope.ai/pricing?co-from=skillNS to top up your balance.
  3. For onboarding and setup guidance, refer to https://www.nexscope.ai/help/skills-external-access?co-from=skillNS

Usage Examples

1. Single patent lookup by publication number

Retrieve the abstract image for patent CN115059423A.

2. Batch lookup by publication number

Retrieve the abstract images for patents CN115059423A and CN115059424A.

Pass {"patentNumber":"CN115059423A,CN115059424A"} to the script for this batch.

3. Lookup by patent ID

Fetch the abstract image for patent ID 5e6f7a8b9c.

Display Rules

  1. Show the image: When the response includes an abstractDrawingPath, display the image directly using Markdown image syntax so the user can see the drawing inline.
  2. Patent identification: Always show the publication number (pn) alongside each image so the user knows which patent each drawing belongs to.
  3. Missing images: If a patent has no abstract drawing (empty abstractDrawingPath), explicitly inform the user that no abstract image is available for that patent.
  4. Batch results: For each returned patent, pair its image and metadata with its own publication number (pn).
  5. Error handling: When a query fails, explain the reason based on the response and suggest the user verify their patent IDs or publication numbers.
  6. No subjective analysis: Present the retrieved images and metadata without adding subjective patent analysis or legal interpretations.

User Expression & Scenario Quick Reference

Applicable -- Patent abstract image retrieval:

User Says Scenario
"Show me the abstract image for patent XX" Single patent image lookup
"What does the patent figure look like" Abstract drawing retrieval
"Retrieve patent illustrations for XX" Image download path retrieval
"I need the abstract drawing for publication number XX" Lookup by publication number

Not applicable -- Needs beyond abstract image retrieval:

  • Full patent text or claims analysis
  • Patent search by keyword or classification
  • Patent legal status or family information
  • Patent citation or prior art analysis
  • Patent valuation or infringement analysis