| name | literature-survey-csv |
| description | Export literature search results to a structured CSV spreadsheet. Use when the user wants to view, share, or analyze search results in tabular form (Excel/Sheets-compatible). Triggers: CSV, spreadsheet, Excel, table, export, 导出, 表格, survey. |
| always | false |
Literature Survey CSV Export
Export deduplicated paper results to a CSV file for analysis in Excel, Google Sheets, or Python/pandas.
When to Use
After Stage 1 (Literature Search → Dedup → Citation Analysis), when you have a JSON results set from academic_dedup or academic_citation_network.
How to Generate
Option A: Run the script (deterministic, preferred)
python scripts/generate_csv.py <input.json> -o papers/literature_survey.csv
The script is located in the packaged skill directory:
acadclaw/skills/literature-survey-csv/scripts/generate_csv.py.
Copy it to the workspace or run it directly from the installed package path.
Input: JSON file with {"papers": [...]} or a plain JSON array of paper objects.
Output: UTF-8 CSV with BOM (for Excel compatibility).
Option B: Manual via write_file
If the script is unavailable, construct CSV manually:
- Header row:
Title,Authors,Year,Venue,Citations,DOI,URL,Source,Type,Topic
- One row per paper, fields comma-separated, text fields quoted
- Save to
papers/literature_survey.csv
CSV Columns
| Column | Source Field | Notes |
|---|
| Title | title | Always present |
| Authors | authors | Semicolon-separated |
| Year | year | Integer or empty |
| Venue | venue | Journal/conference name |
| Citations | citation_count | Integer, 0 if unknown |
| DOI | doi | Full DOI URL |
| URL | url | Landing page or PDF |
| Source | source | arXiv / Semantic Scholar / OpenAlex |
| Type | type | article, preprint, book, etc. |
| Topic | topic | Primary topic (OpenAlex) |
| Seminal | _seminal_rank | From citation analysis (1=top) |
| Cluster | _topic_cluster | From citation analysis |
Integration with Pipeline
In the academic-research pipeline, CSV export fits after Stage 1:
Search → Dedup → Citation Analysis → **CSV Export** → RAG Indexing → ...
Tips
- For large result sets (>100 papers), the CSV is more useful than scrolling chat
- Users often want to filter/sort by citation count or year in a spreadsheet
- After exporting, you can still use
academic_bibtex for citation management
- The CSV is complementary to BibTeX — CSV for analysis, BibTeX for writing
