Quickstart

The one function you need

Everything in tsauditor goes through a single entry point:

import tsauditor as tsa
 
report = tsa.scan(df, target="Direction", domain="finance")

scan() returns a GuardReport object. From there, you decide what to do with the findings.

---

Minimal example

import pandas as pd
import tsauditor as tsa
 
# Load your time-series DataFrame
df = pd.read_csv("data.csv", index_col="Date", parse_dates=True)
 
# Run the audit
report = tsa.scan(df, target="Direction", domain="finance")
 
# Print a formatted summary to the terminal
report.summary()

If your datetime column isn't already the index, pass it via time_col:

report = tsa.scan(df, target="Direction", time_col="Date", domain="finance")

---

Reading the report

Print to terminal

report.summary()

Prints a colour-coded table showing every issue found, grouped by severity, with suggested actions below.

Access issues programmatically

# Issues that block modeling — fix these first
report.critical
 
# Issues worth reviewing
report.warnings
 
# Informational findings
report.info
 
# All issues, sorted by severity
report.all_issues

Filter by code, module, or severity

# All leakage issues
report.filter(module="leakage")
 
# Specific issue code
report.filter(code="LEK001")
 
# All critical issues from the profiler
report.filter(module="profiler", severity="critical")

Get columns flagged for leakage

# Returns a list of column names to review or remove
report.leaky_columns()

Export to JSON

report.to_json("report.json")

The JSON file contains the full issue list with evidence and suggested actions, suitable for logging in automated pipelines.

---

Sensor data example

import pandas as pd
import tsauditor as tsa
 
df = pd.read_csv("sensor_log.csv", index_col="timestamp", parse_dates=True)
 
# No target needed for structural + anomaly checks only
report = tsa.scan(df, domain="sensor")
report.summary()

---

Selective module runs

Run only the checks you need:

# Only structural profiling, skip anomaly and leakage
report = tsa.scan(df, run_anomaly=False, run_leakage=False)
 
# Only leakage detection
report = tsa.scan(df, target="label", run_profiler=False, run_anomaly=False)

---

Next steps

• How It Works — understand what each module is checking
• Issue Code Reference — look up any specific issue code
• Domain Presets — understand finance vs sensor threshold differences