Evaluating Flags

`get_flag(key, context)`

Returns the data for a specific flag, or None if the flag does not exist.


from canarygate import FlagEvaluationContext
 
flag = client.get_flag("new-checkout", FlagEvaluationContext(user_id="user-42"))
# FlagData or None

The user_id in the context is used to evaluate rollout flags consistently: the same user always gets the same result. When the context or user_id is empty, a per-instance anonymous ID is used.

Safe check


if flag and flag.enabled:
    # Flag is enabled for this user
    pass

`get_flags(context)`

Returns all flags for the environment as a list.


flags = client.get_flags(FlagEvaluationContext(user_id="user-42"))
 
for flag in flags:
    print(flag.key, flag.enabled)

Examples by flag type

Boolean flag


maintenance = client.get_flag("maintenance-mode", context)
 
if maintenance and maintenance.enabled:
    return maintenance_page()

Rollout flag


new_dashboard = client.get_flag("new-dashboard", context)
 
# flag.enabled is True or False based on user_id and percentage
if new_dashboard and new_dashboard.enabled:
    return new_dashboard_page()
 
# To see the configured percentage:
if new_dashboard and new_dashboard.type == "rollout":
    print(f"{new_dashboard.percent}% of users will see this feature")

Before `init()`

If you call get_flag() before init(), the flag will return None (the local cache is still empty).


client = CanaryGate("cg_key")
# Do not do this:
client.get_flag("feature")  # None
 
# Do this instead:
client.init()
client.get_flag("feature")  # FlagData or None

Flags that do not exist

If the flag does not exist in the environment, get_flag() returns None. Always guard with a truthiness check to avoid breaking the application flow:


# If 'new-feature' does not exist in the environment, enabled will be False
flag = client.get_flag("new-feature")
enabled = flag.enabled if flag else False

get_flag(key, context)