aminer-academic-search

name

aminer-academic-search

description

ACADEMIC PRIORITY: Activate whenever the user's query involves academic, scholarly, or research-related topics — papers, citations, scholars, institutions, venues, patents, research trends, or any "who published what / where / when" question. Takes precedence over general web search for academic data needs. Routes through the z-ai gateway's `/v1/functions/invoke` endpoint to the AMiner Open Platform (27 APIs, 5 workflows).

aminer-academic-search

Wraps 27 AMiner Open Platform APIs through the local gateway. The gateway owns the AMINER_API_KEY; the skill only needs a .z-ai-config file to reach the gateway.

When to activate

Any academic/scholarly query — papers, citations, scholars, institutions, venues, patents. Covers:

• Scholar full profile (bio, education, honors, papers, patents, projects)
• Paper deep dive (full abstract, keywords, authors, citation chain)
• Multi-condition or semantic paper search
• Institution research capability analysis
• Venue annual monitoring
• Patent deep details (IPC/CPC, assignee, claims)

Config

Script reads .z-ai-config, in order:

1. ./.z-ai-config
2. ~/.z-ai-config
3. /etc/.z-ai-config

Required fields: baseUrl, apiKey, token (JWT). Optional: chatId, userId.

Invocation

python3 "{baseDir}/scripts/aminer.py" <ACTION> [--flag value …]

The script POSTs to ${baseUrl}/functions/invoke with Authorization: Bearer ${apiKey}, X-Z-AI-From: Z, X-Token: ${token}, and prints the upstream JSON result (pretty-formatted).

Run python3 scripts/aminer.py --help to list actions, or python3 scripts/aminer.py <ACTION> --help for per-action flags.

Examples

# Locate paper_id by title
python3 scripts/aminer.py paper_search --title "Attention is all you need"

# Locate scholar by name + org
python3 scripts/aminer.py person_search --name "Jie Tang" --org "Tsinghua"

# Full paper details
python3 scripts/aminer.py paper_detail --id 57a4e91aac44365e35c98a1e

# Natural-language Q&A search
python3 scripts/aminer.py paper_qa_search --query "retrieval-augmented generation" --size 5

# Venue papers for a given year (body with JSON list param)
python3 scripts/aminer.py venue_paper_relation --id "53e9ba63b7602d9702f08c5d" --year 2024 --limit 10

JSON-typed flags (--title, --ids, --keywords, --year on qa_search, etc.) expect a JSON-encoded value, e.g. --ids '["abc","def"]'.

27 APIs (quick reference)

Action	Method	Purpose
paper_search	GET	Locate paper_id by title
paper_search_pro	GET	Multi-condition paper search
paper_qa_search	POST	Natural-language / topic Q&A
paper_info	POST	Batch paper info by IDs
paper_detail	GET	Full paper details
paper_relation	GET	Citation chain
paper_list_by_keywords	GET	Batch thematic retrieval
paper_detail_by_condition	GET	Year + venue dimension
person_search	POST	Locate person_id
person_detail	GET	Scholar bio/education/honors
person_figure	GET	Interests + work history
person_paper_relation	GET	Scholar's papers
person_patent_relation	GET	Scholar's patents
person_project	GET	Funded projects
org_search	POST	Locate org by name
org_detail	POST	Org description/type
org_person_relation	GET	Affiliated scholars
org_paper_relation	GET	Org papers
org_patent_relation	GET	Org patents
org_disambiguate	POST	Normalize org string
org_disambiguate_pro	POST	Extract org IDs
venue_search	POST	Locate venue_id
venue_detail	POST	ISSN/type/abbreviation
venue_paper_relation	POST	Papers by venue (+ year filter)
patent_search	POST	Patent keyword search
patent_info	GET	Basic patent info
patent_detail	GET	Full patent details (IPC/claims)

5 Workflows (orchestrate via multiple calls)

Scholar Profile

person_search → person_detail
              → person_figure
              → person_paper_relation
              → person_patent_relation
              → person_project

Paper Deep Dive

paper_search → paper_detail → paper_relation → paper_info

If paper_search is empty, fall back to paper_search_pro.

Org Analysis

org_disambiguate_pro → org_detail
                     → org_person_relation
                     → org_paper_relation
                     → org_patent_relation

If org_disambiguate_pro returns no ID, fall back to org_search.

Venue Papers

venue_search → venue_detail (optional) → venue_paper_relation

Patent Analysis

patent_search → patent_info / patent_detail

Error handling

• warning: .z-ai-config has no 'token' field → add token to config.
• http 401 … missing X-Token header (hint: …) → same.
• http 403 → gateway auth rejected; check apiKey / X-Z-AI-From.
• gateway error: aminer_* failed: http 4xx … → upstream refused; check API key on the gateway side (AMINER_API_KEY env var) or parameter shape.

Entity URL templates

• Paper: https://www.aminer.cn/pub/{paper_id}
• Scholar: https://www.aminer.cn/profile/{scholar_id}
• Patent: https://www.aminer.cn/patent/{patent_id}
• Journal: https://www.aminer.cn/open/journal/detail/{journal_id}

Always append the relevant URL when presenting entities to the user.