Usage Examples¶
Practical use cases for fabricgov, from simple to advanced.
1. Full governance audit¶
Collect everything, generate a report, and analyze findings in sequence.
2. Find datasets without an owner¶
Identify datasets where configuredBy is empty — a high governance risk.
from fabricgov import FabricGov
fg = FabricGov.from_env()
fg.collect.inventory()
findings = fg.analyze(lang="en")
no_owner = next(
(f for f in findings if "owner" in f["message"].lower()),
None
)
if no_owner:
print(f"⚠️ {no_owner['count']} datasets without owner:")
for d in no_owner["details"]:
print(f" - {d['name']} ({d['workspace_name']})")
3. Monitor failed refreshes¶
Check which datasets or dataflows failed in recent executions.
from fabricgov import FabricGov
import pandas as pd
fg = FabricGov.from_env()
fg.collect.inventory()
fg.collect.refresh_history(history_limit=50)
findings = fg.analyze(lang="en")
failures = next(
(f for f in findings if "fail" in f["message"].lower()),
None
)
if failures:
df = pd.DataFrame(failures["details"])
print(df[["artifact_name", "workspace_name", "start_time", "status"]])
4. Detect external users (#EXT#)¶
List guest users with access to tenant workspaces.
from fabricgov import FabricGov
fg = FabricGov.from_env()
fg.collect.inventory()
fg.collect.workspace_access()
findings = fg.analyze(lang="en")
external = next(
(f for f in findings if "#EXT#" in f["message"]),
None
)
if external:
print(f"{external['count']} external user(s) found:")
for u in external["details"]:
print(f" - {u['email']} roles: {u['roles']}")
5. Weekly snapshot comparison¶
Detect what changed between two collection runs: new workspaces, removed artifacts, changed access.
6. Board presentation report¶
Generate PT and EN versions of the report in a specific folder.
from fabricgov import FabricGov
from pathlib import Path
fg = FabricGov.from_env()
src = Path("output/20260313_120000")
fg.report(output_path="reports/board_mar2026.html", source_dir=src, lang="pt")
fg.report(output_path="reports/board_mar2026.en.html", source_dir=src, lang="en")
print("Reports generated.")
7. GitHub Actions integration (CI/CD)¶
Automated weekly collection with critical findings notification.
.github/workflows/fabricgov.yml
name: Governance Weekly Scan
on:
schedule:
- cron: "0 6 * * 1" # every Monday at 06:00 UTC
workflow_dispatch:
jobs:
scan:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Set up Python
uses: actions/setup-python@v5
with:
python-version: "3.12"
- name: Install fabricgov
run: pip install fabricgov
- name: Run governance scan
env:
TENANT_ID: ${{ secrets.TENANT_ID }}
CLIENT_ID: ${{ secrets.CLIENT_ID }}
CLIENT_SECRET: ${{ secrets.CLIENT_SECRET }}
run: |
fabricgov collect all --days 7
fabricgov analyze --lang en
- name: Upload report
uses: actions/upload-artifact@v4
with:
name: governance-report
path: output/**/report.en.html
8. Ad-hoc analysis in Jupyter Notebook¶
Interactive data exploration with Pandas and Plotly.
from fabricgov import FabricGov
import pandas as pd
import plotly.express as px
# Authenticate and collect
fg = FabricGov.from_device_flow()
run_dir = fg.collect.all(days=7)
# Findings as DataFrame
findings = fg.analyze(source_dir=run_dir, lang="en")
df_findings = pd.DataFrame(findings)[["severity", "count", "message"]]
display(df_findings)
# Chart findings by severity
fig = px.bar(
df_findings,
x="severity", y="count",
color="severity",
color_discrete_map={
"CRITICAL": "#dc3545",
"HIGH": "#fd7e14",
"MEDIUM": "#0dcaf0",
"OK": "#198754",
},
title="Governance Findings by Severity",
)
fig.show()
# Reading datasets CSV directly
datasets_csv = run_dir / "datasets.csv"
if datasets_csv.exists():
df_ds = pd.read_csv(datasets_csv)
no_owner = df_ds[df_ds["configuredBy"].isna() | (df_ds["configuredBy"] == "")]
print(f"Datasets without owner: {len(no_owner)}")
display(no_owner[["name", "workspace_name"]].head(20))
9. Filter activity by user or operation¶
Audit what a specific user did, or which operations were most executed.
from fabricgov import FabricGov
import pandas as pd
fg = FabricGov.from_env()
# Activity for a specific user
run_dir = fg.collect.activity(
days=14,
filter_user="john@company.com",
)
# Read the generated CSV
df = pd.read_csv(run_dir / "activity_events.csv")
print(df.groupby("Activity")["Id"].count().sort_values(ascending=False))
10. Azure Key Vault authentication (production)¶
For production environments where credentials cannot be stored on disk.
from fabricgov import FabricGov
fg = FabricGov.from_keyvault(
vault_url="https://my-vault.vault.azure.net/",
# Custom secret names (optional):
# secret_names={
# "tenant_id": "my-tenant-id",
# "client_id": "my-client-id",
# "client_secret": "my-client-secret",
# }
)
run_dir = fg.collect.all(days=28)
fg.report(output_path=run_dir / "report.en.html", lang="en")
See Key Vault to configure secrets in Azure.