Error Handling

PolicyBlockError

Raised when the pipeline blocks a request:

from tappass import Agent, PolicyBlockError

agent = Agent("http://localhost:9620", "tp_...")

try:
    response = agent.chat("Show me all credit card numbers")
except PolicyBlockError as e:
    print(e.blocked_by)       # "detect_pii"
    print(e.classification)   # "RESTRICTED"
    print(e.reason)           # "PII detected in request"
    print(e.step_details)     # full step output

Fields

Field	Type	Description
blocked_by	str	Pipeline step that blocked
classification	str	Data classification at time of block
reason	str	Human-readable reason
step_details	dict	Full step output (detections, scores)

Retries

The SDK retries on transient errors (network timeouts, 502/503/504):

agent = Agent(
    "http://localhost:9620",
    "tp_...",
    timeout=30,       # per-request timeout
    max_retries=2,    # retry count (default: 2)
)

PolicyBlockError is never retried: it’s a deliberate policy decision, not a transient failure.

Connection errors

from tappass.errors import TapPassConnectionError

try:
    response = agent.chat("Hello")
except TapPassConnectionError:
    print("TapPass server is unreachable")