Analyze a workbook to JSON
from excelminer import AnalysisOptions, analyze_to_dict
result = analyze_to_dict(
"workbook.xlsx",
options=AnalysisOptions(include_formulas=True),
)
print(result["graph"]["stats"])
Inspect the normalized graph
from excelminer import AnalysisOptions, analyze_workbook
graph, reports, ctx = analyze_workbook(
"workbook.xlsx",
options=AnalysisOptions(include_formulas=True),
)
print(graph.stats())
print(ctx.issues)
What excelminer extracts
The pipeline is designed for inventory, reproducible diffs, and analysis. Formulas are captured as text, macros are never executed.
Workbook structure
- Sheets, defined names, charts (when referenced)
- Connections + inferred sources
- Pivot tables + caches
Query & automation hints
- Power Query definitions + M scripts
- Binary mashup detection (best-effort extraction)
- VBA project module text
Optional enrichment
- Formula text & dependencies
- Used-range value blocks (calamine)
- COM automation (Windows + Excel)
Backend pipeline at a glance
The default backend order can be overridden, but this is the typical flow.
- OOXML zip parsing (sheets, defined names, charts, connections).
- VBA project module extraction for macro-enabled files.
- Power Query extraction + script/source inference.
- Pivot tables and caches.
- Calamine used-range scanning (optional).
- Openpyxl formula text (optional).
- Excel COM enrichment (Windows-only, opt-in).
Serving locally
Tip: Write the bundled site files to a folder and open index.html locally:
python -m excelminer.docs --write-site .\excelminer-site
Then open excelminer-site\index.html in your browser.
Core output shape
Every analysis returns a deterministic graph.
Node
{
"id": "sheet:Sheet1",
"kind": "sheet",
"key": "Sheet1",
"attrs": {"index": 0}
}Edge
{
"src": "sheet:Sheet1",
"dst": "formula_cell:Sheet1!A1",
"kind": "contains",
"attrs": {}
}Explore the details
Implementation guidance
Learn about architecture decisions, backend responsibilities, and data flow.
Safety first
Understand connection sanitization, Power Query privacy risks, and COM automation constraints.