pyramid_analytics 0.2.1

要使用pyramid_analytics，您应该在配置文件中包含以下指令

pyramid.includes =
    pyramid_analytics

以下配置选项可用。注意，analytics.api_token和analytics.secret是必需的。

analytics.api_token = mytoken
analytics.async = True
analytics.flush_after = 10
analytics.flush_at = 20
analytics.max_queue_size = 100000
analytics.secret = mysekret
analytics.send = True

flush_after指定在服务器刷新之前没有刷新的时间（以秒为单位）。与flush_at大小策略一起使用。

要管理由Segment.io的analytics-python指定的log_level和log选项，您应该在您的.ini文件中设置以下配置

[logger_analytics]
level = WARN
handlers = console  # The config needs to include this handler
qualname = analytics
propagate = 0  # This "disables" logging
formatter = generic  # The config needs to include this formatter

此外，您还需要注册此记录器

在底层，analytics-python通过将日志级别设置为CRITICAL来关闭日志记录。设置propagate = 0可能是一个更优雅的解决方案。

所有其他选项的详细信息请参阅Segment.io的Python文档。

此库将一个属性analytics附加到请求对象上。pyramid_analytics会检查Pyramid api中的authenticated_userid，并为认证用户识别或跟踪数据。

要更新此analytics包代表其发起调用的用户ID，您应该发送一个新的UpdatedAnalyticsUserId事件。示例

from pyramid_analytics.events import UpdatedAnalyticsUserId

# ...
# User logged in
event = UpdatedAnalyticsUserId(request, userid, alias=False)
request.registry.notify(event)

此事件将更新分析数据。如果用户注册了您的服务，您应设置alias=True。这不会向segment.io发送任何代码，而是与下面描述的Jinja2模板良好地协作。

此分析包尽可能地直接从您的服务器发送数据，并回退到通过Jinja2模板向浏览器提供事件。以下是一些标识和跟踪的示例

# identify the authenticated user with certain traits (timestamp and
# context are optional and described by the segment.io documentation)
request.analytics.identify(traits, timestamp=None, context=None)

# identify a (not identified) user with certain traits (timestamp and
# context are optional and described by the segment.io documentation)
request.analytics.identify_foreign(user_id, traits, timestamp=None,
                                   context=None)

# track an authenticated user with an event and properties (timestamp and
# context are optional and described by the segment.io documentation)
request.analytics.track(event, properties, timestamp=None, context=None)

# track a (not identified) user with an event and properties (timestamp and
# context are optional and described by the segment.io documentation)
request.analytics.track_foreign(user_id, event, properties, timestamp=None,
                                context=None)

当Pyramid触发BeforeRender事件时，pyramid_analytics会将以下内容添加到渲染器的全局变量中

analytics: {
   api_token: 'mytoken',
   tracking: {
        alias: userid, // or null if analytics hasn't been told to alias
        events: [{event: 'myevent',
                    properties: {property1: 'pvalue1'},
                    context: {context1: 'cvalue1'}}]
        identify: userid, // or null
    }
    tracking_json: '...' // the tracking item as json
}

提供了一个易于使用的Jinja2模板。要使用此模板，您必须将pyramid_analytics添加到您的jinja2搜索路径。如果您使用Pyramid的pyramid_jinja2，请按照以下示例更新您的config.ini文件

jinja2.directories =
    pyramid_analytics:templates
    myapp:templates

要使用模板，请在包含segment.io的javascript后将其放置在您的模板中

{# Segment.io's javascript precedes this... #}
{% include 'analytics.jinja2' %}