Legacy Python Server SDK
Statsig's Legacy Server SDK for Python applications
Setup the SDK
Install the SDK
Install the sdk using pip3:The Statsig SDK isn't compatible with python 2. You must be on python 3.7+ to use the Statsig SDK.
bashpip3 install statsigInitialize the SDK
After installation, initialize the SDK using a Server Secret Key from the Statsig console.There is also an optional parameter namedDo NOT embed your Server Secret Key in client-side applications, or expose it in any external-facing documents. However, if you accidentally expose it, you can create a new one in the Statsig console.
optionsthat allows you to pass in a StatsigOptions to customize the SDK.pythonfrom statsig import statsig statsig.initialize("server-secret-key") # or with StatsigOptions options = StatsigOptions(tier=StatsigEnvironmentTier.development) statsig.initialize("server-secret-key", options) # check if sdk is initialized initialized = statsig.is_initialized()initializeperforms a network request. Afterinitializecompletes, virtually all SDK operations are synchronous (refer to Evaluating Feature Gates in the Statsig SDK). The SDK fetches updates from Statsig in the background, independently of API calls.
Working with the SDK
Checking a Feature Flag/Gate
After the SDK is initialized, you can check a Feature Gate. Feature Gates create logic branches in code that you can roll out to different users from the Statsig Console. Gates are always CLOSED or OFF (return false;) by default.All APIs require you to specify the user (refer to Statsig user) associated with the request. For example, to check a gate for a user:from statsig.statsig_user import StatsigUser
...
statsig.check_gate(StatsigUser("user-id"), "gate-name")
Reading a Dynamic Config
Feature Gates work well for simple on/off switches with optional user targeting. To send a different set of values (strings, numbers, and so on) to clients based on specific user attributes such as country, use Dynamic Configs. The Dynamic Config API is similar to Feature Gates, but returns a full JSON object configured on the server, from which you can fetch typed parameters.config = statsig.get_config(StatsigUser("user-id"), "config-name")
config_json = config.get_value()
Getting a Layer/Experiment
Use Layers/Experiments to run A/B/n experiments. Two APIs are available, but Statsig recommends layers for faster iterations with parameter reuse.# Values via getLayer
layer = statsig.get_layer(user, "user_promo_experiments")
title = layer.get("title", "Welcome to Statsig!")
discount = layer.get("discount", 0.1)
# or, via getExperiment
title_exp = statsig.get_experiment(user, "new_user_promo_title")
price_exp = statsig.get_experiment(user, "new_user_promo_price")
title = title_exp.get("title", "Welcome to Statsig!")
discount = price_exp.get("discount", 0.1)
...
price = msrp * (1 - discount)
Retrieving Feature Gate Metadata
When you need more than a boolean value from a gate evaluation, use the Get Feature Gate API, which returns a FeatureGate object with additional evaluation metadata:
gate = statsig.get_feature_gate(StatsigUser("user-id"), "gate-name")
print(gate.name) # 'gate-name'
print(gate.value) # True or False
print(gate.rule_id) # rule ID that was evaluated
print(gate.evaluation_details) # evaluation metadata
Logging an Event
To track custom events and measure how features or experiment groups affect those events, call the Log Event API. Specify the user and event name to log, and optionally provide a value and metadata object:
from statsig.statsig_user import StatsigUser
from statsig.statsig_event import StatsigEvent
statsig.log_event(StatsigEvent(StatsigUser("user-id"), "event-name"))
Python supports retry_queue_size, which allows you to adjust the memory allocated for handling retries. While service outages are rare, increasing the retry_queue_size can help minimize event loss by providing additional memory to buffer events during such occurrences. This option is generally not needed for typical use but offers added flexibility in exceptional situations.
Statsig User
When calling APIs that require a user, pass as much information as possible to take advantage of advanced gate and config conditions (like country or OS/browser level checks), and to correctly measure the impact of your experiments on your metrics/events. At least one identifier (userID or customID) is required to provide a consistent experience for a given user. Refer to userID requirements for more detail.In addition to userID, email, ip, userAgent, country, locale, and appVersion are available as top-level fields on StatsigUser. You can also pass any key-value pairs in an object/dictionary to the custom field to create targeting based on them.
Typing on the StatsigUser object is lenient: you can pass numbers, strings, arrays, objects, and even enums or classes. However, evaluation operators only work on primitive types, mostly strings and numbers. The SDK attempts to cast custom field types to match the operator, but evaluation results for other types are not guaranteed. For example, an array set as a custom field is only compared as a string: there is no operator to match a value within that array.
Private Attributes
To keep sensitive user PII data out of logs, use the privateAttributes field on the StatsigUser object. This field accepts an object/dictionary of private user attributes. Any attribute set in privateAttributes is used only for evaluation/targeting and is removed from all logs before Statsig sends them to its servers.
For example, if a feature gate should only pass for users with emails ending in "@statsig.com", but you don't want to log email addresses to Statsig, add the key-value pair { email: "my_user@statsig.com" } to privateAttributes on the user.
Statsig Options
initialize() takes an optional options parameter in addition to the secret key to customize the Statsig client. Create a StatsigOptions class with the following available parameters:
(unit of measure for time related options is seconds)
tierStatsigEnvironmentTier | strSets the environment tier (for gates to evaluate differently in development and production)
Set the environment tier using the StatsigEnvironmentTier enum or as a str.
timeoutintEnforces a minimum timeout on network requests from the SDK
init_timeoutintSets the maximum timeout on download config specs and id lists network requests for initialization
rulesets_sync_intervalintHow often the SDK updates rulesets from Statsig servers
idlists_sync_intervalintHow often the SDK updates idlists from Statsig servers
local_modeboolDisables all network requests. SDK returns default values and won't log events. Useful in combination with overrides to mock behavior for tests.
bootstrap_valuesstra string that represents all rules for all feature gates, dynamic configs and experiments. It can be provided to bootstrap the Statsig server SDK at initialization in case your server runs into network issue or Statsig server is down temporarily.
rules_updated_callbacktyping.Callablea callback function that's called whenever the rules update; it's called with a logical timestamp and a JSON string (used as is for bootstrapValues mentioned above). As of v0.6.0, the SDK calls this from a background thread that it uses to update config values.
event_queue_sizeintThe number of events to batch before flushing the queue to the network. Default 500.
Events are also batched every minute by a background thread
data_storeIDataStoreA data store with custom storage behavior for config specs. Can be used to bootstrap Statsig server (takes priority over bootstrap_values).
proxy_configsOptional[Dict[NetworkEndpoint, ProxyConfig]]Configuration network for each endpoint, for example, download_config_spec, get_id_lists
fallback_to_statsig_apiOptional[bool]Fallback to Statsig CDN for download config specs and get id lists if the overridden api failed.
initialize_sourcesOptional[List[DataSource]]List of sources SDK tries to get download_config_specs from when initialize. The list is ordered, SDK tries to get source from first element, and stops when getting dcs successfully
config_sync_sourcesOptional[List[DataSource]]List of sources SDK tries to get download_config_specs from when downloading. The list is ordered, SDK tries to get source from first element, and stops when getting dcs successfully
Example:
from statsig import statsig, StatsigEnvironmentTier, StatsigOptions
options = StatsigOptions(None, StatsigEnvironmentTier.development)
statsig.initialize("secret-key", options).wait()
You can also use the set_environment_parameter function, but that takes in string values only:
from statsig import statsig, StatsigEnvironmentTier, StatsigOptions
options = StatsigOptions()
options.set_environment_parameter("tier", StatsigEnvironmentTier.development.value)
statsig.initialize("secret-key", options).wait()
Shutdown
To gracefully shutdown the SDK and ensure all events are flushed:
statsig.shutdown()
Client SDK bootstrapping
The Statsig server SDK can generate the initialization values for a client SDK. This is useful for server-side rendering (SSR) or when you want to pre-fetch values for a client.
values = statsig.get_client_initialize_response(user); # dict() | None
# To apply local overrides, set include_local_overrides = True (python sdk v0.32.0+)
values = statsig.get_client_initialize_response(user=user, include_local_overrides=True); # dict() | None
Local Overrides
You can override the values returned by the SDK for testing purposes, which is useful for local development when testing specific scenarios.
# Adding/Removing gate overrides
statsig.override_gate("a_gate_name", true, "a_user_id")
statsig.remove_gate_override("a_gate_name", "a_user_id")
# Adding/Removing config overrides
statsig.override_config("a_config_name", {"key": "value"}, "a_user_id")
statsig.remove_config_override("a_config_name", "a_user_id")
# Adding/Removing experiment overrides
statsig.override_experiment("an_experiment_name", {"key": "value"}, "a_user_id")
statsig.remove_experiment_override("an_experiment_name", "a_user_id")
# Remove All Overrides
statsig.remove_all_overrides()
# You can also override with custom ids
custom_id_user = StatsigUser("a_user_id", custom_ids={"statsigId": "a_statsig_id"})
statsig.override_gate("a_gate_name", true, "a_statsig_id")
# Local overrides will prioritize override with userId, then look up the custom id to override.
# To prevent clashing overrides, it is recommended to not use the same value for userId and customIds for different users.
Multi-instance usage
To create multiple independent instances of the Statsig SDK (for example, to use different API keys or configurations), use the instance-based approach:
sdk_instance = StatsigServer()
sdk_instance.initialize(secret_key, options);
Forward proxy configuration
You can configure the SDK to use a forward proxy for network requests:
Basic setup to stream download config spec from forward proxy:
proxyAddress = "0.0.0.0:50051" // local address update to your address
Statsig.initialize(secret_key, StatsigOptions(proxy_configs={
NetworkEndpoint.DOWNLOAD_CONFIG_SPECS: ProxyConfig(NetworkProtocol.GRPC_WEBSOCKET, proxyAddress)}))
When the SDK disconnects from the forward proxy when using grpc_websocket, the SDK retries the connection with exponential backoff. After push_worker_failover_threshold retries, the SDK starts polling from Statsig until it reconnects to the forward proxy. You can customize Streaming Failover Behavior. You can also define the sources/endpoints SDK poll from, SDK will try from source at index 0, and stops trying if get a response.
statsigOptions = StatsigOptions(
proxy_configs={
NetworkEndpoint.DOWNLOAD_CONFIG_SPECS: ProxyConfig(
protocol=NetworkProtocol.GRPC_WEBSOCKET,
proxy_address=address,
push_worker_failover_threshold=1, # start polling from Statsig endpoint after 1 retry failed
# 1st retry 5000 ms later, 2nd retry 2 * 5000ms = 10 seconds ....
retry_backoff_multiplier=2,
max_retry_attempt=8,
retry_backoff_base_ms=5000
)
},
# Get from network first, which is forward proxy here, if fails, try datastore, if fails try poll from Statsig endpoint
initialize_sources=[
DataSource.NETWORK,
DataSource.DATASTORE,
DataSource.STATSIG_NETWORK,
],
)
FAQs
How can I mock Statsig for testing?
The python server SDK, starting in version 0.5.1+, supports a few features to make testing easier.
The StatsigOptions parameter localMode, when set to true, prevents the SDK from making network requests and causes it to return only default values. This is useful for placeholder or test environments that shouldn't access the network.
overrideGate and overrideConfig APIs on the global statsig interface (refer to Local Overrides) can set a gate or config override for a specific user, or for all users (by not providing a specific user ID).Enable localMode and then override gates/configs/experiments to specific values to test the code flows you are building.
Can I generate the initialize response for a client SDK using the Python server SDK?
Yes. Refer to Client Initialize Response.Reference
StatsigUser
@dataclass
class StatsigUser:
"""An object of properties relating to the current user
user_id or customID is required: /sdks/user#why-is-an-id-always-required-for-server-sdks
Provide as many as possible to take advantage of advanced conditions in the statsig console
A dictionary of additional fields can be provided under the custom field
Set private_attributes for any user property you need for gate evaluation but prefer stripped from logs/metrics
"""
user_id: Optional[str] = None
email: Optional[str] = None
ip: Optional[str] = None
user_agent: Optional[str] = None
country: Optional[str] = None
locale: Optional[str] = None
app_version: Optional[str] = None
custom: Optional[dict] = None # key: string, value: string
private_attributes: Optional[dict] = None # key: string, value: string
custom_ids: Optional[dict] = None # key: string, value: string
StatsigOptions
class StatsigOptions:
"""An object of properties for initializing the sdk with additional parameters"""
def __init__(
self,
api: Optional[str] = None,
api_for_download_config_specs: Optional[str] = None,
api_for_get_id_lists: Optional[str] = None,
api_for_log_event: Optional[str] = None,
tier: Union[str, StatsigEnvironmentTier, None] = None,
init_timeout: Optional[int] = None,
timeout: Optional[int] = None,
rulesets_sync_interval: int = DEFAULT_RULESET_SYNC_INTERVAL,
idlists_sync_interval: int = DEFAULT_IDLIST_SYNC_INTERVAL,
local_mode: bool = False,
bootstrap_values: Optional[str] = None,
rules_updated_callback: Optional[Callable] = None,
event_queue_size: Optional[int] = DEFAULT_EVENT_QUEUE_SIZE,
data_store: Optional[IDataStore] = None,
idlists_thread_limit: int = DEFAULT_IDLISTS_THREAD_LIMIT,
logging_interval: int = DEFAULT_LOGGING_INTERVAL, #deprecated
disable_diagnostics: bool = False,
custom_logger: Optional[OutputLogger] = None,
enable_debug_logs = False,
disable_all_logging = False,
evaluation_callback: Optional[Callable[[Union[Layer, DynamicConfig, FeatureGate]], None]] = None,
retry_queue_size: int = DEFAULT_RETRY_QUEUE_SIZE,
proxy_configs: Optional[Dict[NetworkEndpoint, ProxyConfig]] = None,
fallback_to_statsig_api: Optional[bool] = False,
initialize_sources: Optional[List[DataSource]] = None,
config_sync_sources: Optional[List[DataSource]] = None,
):
FeatureGate
class FeatureGate:
def get_value(self):
"""Returns the underlying value of this FeatureGate"""
def get_name(self):
"""Returns the name of this FeatureGate"""
def get_evaluation_details(self):
"""Returns the evaluation detail of this FeatureGate"""
DynamicConfig
class DynamicConfig:
def get(self, key, default=None):
"""Returns the value of the config at the given key
or the provided default if the key is not found
"""
def get_typed(self, key, default=None):
"""Returns the value of the config at the given key
iff the type matches the type of the provided default.
Otherwise, returns the default value
"""
def get_value(self):
"""Returns the underlying value of this DynamicConfig"""
def get_name(self):
"""Returns the name of this DynamicConfig"""
def get_evaluation_details(self):
"""Returns the evaluation detail of this DynamicConfig"""
Layer
class Layer:
def get(self, key, default=None):
"""Returns the value of the layer at the given key
or the provided default if the key is not found
"""
def get_typed(self, key, default=None):
"""Returns the value of the layer at the given key
iff the type matches the type of the provided default.
Otherwise, returns the default value
"""
def get_name(self):
"""Returns the name of this Layer"""
def get_values(self):
"""Returns all the values in this Layer but does not trigger an exposure log"""
def get_evaluation_details(self):
"""Returns the evaluation detail of this Layer"""
EvaluationDetails
class EvaluationDetails:
reason: EvaluationReason
config_sync_time: int
init_time: int
server_time: int
class EvaluationReason(str, Enum):
network = "Network"
local_override = "LocalOverride"
unrecognized = "Unrecognized"
uninitialized = "Uninitialized"
bootstrap = "Bootstrap"
data_adapter = "DataAdapter"
unsupported = "Unsupported"
error = "error"
DataStore
class IDataStore:
def get(self, key: str) -> Optional[str]:
return None
def set(self, key: str, value: str):
pass
def shutdown(self):
pass
def should_be_used_for_querying_updates(self, key: str) -> bool:
return False
ForwardProxy - ProxyConfig
class NetworkProtocol(Enum):
HTTP = "http"
GRPC = "grpc"
GRPC_WEBSOCKET = "grpc_websocket"
class NetworkEndpoint(Enum):
LOG_EVENT = "log_event"
DOWNLOAD_CONFIG_SPECS = "download_config_specs"
GET_ID_LISTS = "get_id_lists"
ALL = "all"
class ProxyConfig:
def __init__(
self,
protocol: NetworkProtocol,
proxy_address: str,
# Websocket worker failover config
max_retry_attempt: Optional[int] = None, # default is 10
retry_backoff_multiplier: Optional[int] = None, # default is # default is 5
retry_backoff_base_ms: Optional[int] = None, # default is 10,000 ms
# Push worker failback to polling threshold, fallback immediate set 0,
# n means fallback after n retry failed
push_worker_failover_threshold: Optional[int] = None, # default is 4, about 30 minutes
):
self.proxy_address = proxy_address
self.protocol = protocol
self.max_retry_attempt = max_retry_attempt
self.retry_backoff_multiplier = retry_backoff_multiplier
self.retry_backoff_base_ms = retry_backoff_base_ms
self.push_worker_failover_threshold = push_worker_failover_threshold
Was this helpful?