🚀 OneClick Keyword / MeSH / Co-word Semantic Tool

Research utility

Step A: extract OpenAlex / PubMed keywords from your CSV.
Step C: convert abstract / co-word data into GPT-based semantic phrase sets (per document).
Both steps require a valid Club Membership Code (CMC); GPT usage is billed directly by OpenAI.

A. Keyword / MeSH Retriever (CMC required)

Upload CSV with DOI or PMID

For this standalone script, Step A currently supports:

DOI → OpenAlex concepts (column openalex_concepts)
PMID → PubMed MeSH terms (column mesh_terms)

Access control: a valid 10-digit Club Membership Code (CMC) is required. CMCs are checked annually by a backend rule.

Authors personally cover Google App Engine usage fees to keep this app online. If this tool supports your bibliometrics research, a voluntary donation is appreciated to help fund maintenance and development.

🚀 Donate

C. Abstract / Co-word → Semantic Phrases (GPT)

Upload data and get semantic phrase sets per record

This step starts from abstract or co-word data. You can upload:

Single-column abstracts (>20 words without /)
Multi-column co-word data (one row = one document; non-numeric cells merged)

Access control: the same 10-digit CMC used in Step A is also required for Step C. In this step GPT is always applied; a valid OpenAI API key is mandatory and all GPT charges are billed by OpenAI to your own account.

🔑 Club Membership Code (CMC, 10 digits) Same CMC rule as Step A (checked by current year).

Upload abstract / co-word data (.csv) Only meaningful terms in co-words are required, excluding those like country and department names

Output file contains semantic phrase sets per record (.csv) each row = document, columns = Item_1..Item_k

Number of rows/documents to analyze by GPT Default = 15. Increase this value to analyze more rows; for example, 100 or 9999.

Maximum semantic phrases per document Default = 10. This controls the output columns Item_1 ... Item_k.

Number of nodes shown in network plot Default = 30. Increase/decrease this scale to control visible nodes in the network.

Network node label font size Default = 26. Increase this value to make node labels larger and bolder.

OpenAI API Key (optional for one-column abstracts) If provided, Step C uses GPT. If blank, one-column abstract rows use the rule-based fallback extractor. Multi-column co-word files still require an API key for GPT semantic grouping.

Demo data

Run built-in demo files from the app root folder

These demos use local files placed in the same folder as this Python file: doi.csv, pmid.csv, and abstract.csv.

Download demo input CSV formats

doi.csv — DOI input for Mode A
pmid.csv — PMID input for Mode A
abstract.csv — one-column abstract input for Mode C

All demos return a network dashboard and ZIP download. DOI/PMID demos first retrieve keywords, then build a co-word network from the extracted concepts/MeSH terms.

D. Overview

What happens in each step?

Inputs & detection (Step A)

Upload one CSV per run.
Column 'DI' or first column with DOI → OpenAlex.
Column 'PMID' or numeric-only first column → PubMed MeSH.
CMC is validated using a backend rule and the current year.

Inputs & detection (Step C)

1 column & abstract-like → treat as abstract text.
>1 column → merge non-numeric cells as co-word text.
Meaningless rows (only digits or very short tokens) are skipped.

Outputs

Step A: *_openalex_pubmed.csv (columns openalex_concepts or mesh_terms)
Step C: *_gpt_semantic_phrases.csv — wide table:
- Doc_index, Item_1, ..., Item_k
- Up to max_terms semantic phrases per record.

Use Step C output as a co-word set per document for downstream FLCA/TAAA analysis.