Polishing repo

tinybird/datasources/analytics_events.datasource

@@ -1,16 +1,16 @@
-TOKEN tracker APPEND
-
 DESCRIPTION >
     Analytics events landing data source
 
+TOKEN "tracker" APPEND
+
 SCHEMA >
     `timestamp` DateTime `json:$.timestamp`,
     `session_id` String `json:$.session_id`,
     `action` LowCardinality(String) `json:$.action`,
     `version` LowCardinality(String) `json:$.version`,
     `payload` String `json:$.payload`
 
-ENGINE "MergeTree"
-ENGINE_PARTITION_KEY "toYYYYMM(timestamp)"
-ENGINE_SORTING_KEY "timestamp"
-ENGINE_TTL "timestamp + toIntervalDay(60)"
+ENGINE MergeTree
+ENGINE_PARTITION_KEY toYYYYMM(timestamp)
+ENGINE_SORTING_KEY timestamp
+ENGINE_TTL timestamp + toIntervalDay(60)

tinybird/datasources/analytics_pages_mv.datasource

@@ -0,0 +1,12 @@
+SCHEMA >
+    `date` Date,
+    `device` String,
+    `browser` String,
+    `location` String,
+    `pathname` String,
+    `visits` AggregateFunction(uniq, String),
+    `hits` AggregateFunction(count)
+
+ENGINE AggregatingMergeTree
+ENGINE_PARTITION_KEY toYYYYMM(date)
+ENGINE_SORTING_KEY date, device, browser, location, pathname

tinybird/datasources/analytics_sessions_mv.datasource

@@ -0,0 +1,13 @@
+SCHEMA >
+    `date` Date,
+    `session_id` String,
+    `device` SimpleAggregateFunction(any, String),
+    `browser` SimpleAggregateFunction(any, String),
+    `location` SimpleAggregateFunction(any, String),
+    `first_hit` SimpleAggregateFunction(min, DateTime),
+    `latest_hit` SimpleAggregateFunction(max, DateTime),
+    `hits` AggregateFunction(count)
+
+ENGINE AggregatingMergeTree
+ENGINE_PARTITION_KEY toYYYYMM(date)
+ENGINE_SORTING_KEY date, session_id

tinybird/datasources/analytics_sources_mv.datasource

@@ -0,0 +1,12 @@
+SCHEMA >
+    `date` Date,
+    `device` String,
+    `browser` String,
+    `location` String,
+    `referrer` String,
+    `visits` AggregateFunction(uniq, String),
+    `hits` AggregateFunction(count)
+
+ENGINE AggregatingMergeTree
+ENGINE_PARTITION_KEY toYYYYMM(date)
+ENGINE_SORTING_KEY date, device, browser, location, referrer

tinybird/pipes/analytics_hits.pipe

@@ -1,52 +1,58 @@
-TOKEN dashboard READ
-
 DESCRIPTION >
     Parsed `page_hit` events, implementing `browser` and `device` detection logic.
 
+TOKEN "dashboard" READ
+
 NODE parsed_hits
 DESCRIPTION >
     Parse raw page_hit events
 
 SQL >
     SELECT
-      timestamp,
-      action,
-      version,
-      coalesce(session_id,'0') as session_id,
-      JSONExtractString(payload, 'locale') as locale,
-      JSONExtractString(payload, 'location') as location,
-      JSONExtractString(payload, 'referrer') as referrer,
-      JSONExtractString(payload, 'pathname') as pathname,
-      JSONExtractString(payload, 'href') as href,
-      lower(JSONExtractString(payload, 'user-agent')) as user_agent
-    FROM
-      analytics_events
+        timestamp,
+        action,
+        version,
+        coalesce(session_id, '0') as session_id,
+        JSONExtractString(payload, 'locale') as locale,
+        JSONExtractString(payload, 'location') as location,
+        JSONExtractString(payload, 'referrer') as referrer,
+        JSONExtractString(payload, 'pathname') as pathname,
+        JSONExtractString(payload, 'href') as href,
+        lower(JSONExtractString(payload, 'user-agent')) as user_agent
+    FROM analytics_events
     where action = 'page_hit'
 
 NODE endpoint
 SQL >
     SELECT
-      timestamp,
-      action,
-      version,
-      session_id,
-      location,
-      referrer,
-      pathname,
-      href,
-      case
-        when match(user_agent, 'wget|ahrefsbot|curl|urllib|bitdiscovery|\+https://|googlebot') then 'bot'
-        when match(user_agent, 'android') then 'mobile-android'
-        when match(user_agent, 'ipad|iphone|ipod') then 'mobile-ios'
-        else 'desktop'
-      END as device,
-      case
-        when match(user_agent, 'firefox') then 'firefox'
-        when match(user_agent, 'chrome|crios') then 'chrome'
-        when match(user_agent, 'opera') then 'opera'
-        when match(user_agent, 'msie|trident') then 'ie'
-        when match(user_agent, 'iphone|ipad|safari') then 'safari'
-        else 'Unknown'
-      END as browser
-    FROM
-      parsed_hits
+        timestamp,
+        action,
+        version,
+        session_id,
+        location,
+        referrer,
+        pathname,
+        href,
+        case
+            when match(user_agent, 'wget|ahrefsbot|curl|urllib|bitdiscovery|\+https://|googlebot')
+            then 'bot'
+            when match(user_agent, 'android')
+            then 'mobile-android'
+            when match(user_agent, 'ipad|iphone|ipod')
+            then 'mobile-ios'
+            else 'desktop'
+        END as device,
+        case
+            when match(user_agent, 'firefox')
+            then 'firefox'
+            when match(user_agent, 'chrome|crios')
+            then 'chrome'
+            when match(user_agent, 'opera')
+            then 'opera'
+            when match(user_agent, 'msie|trident')
+            then 'ie'
+            when match(user_agent, 'iphone|ipad|safari')
+            then 'safari'
+            else 'Unknown'
+        END as browser
+    FROM parsed_hits

tinybird/pipes/analytics_pages.pipe

@@ -3,22 +3,16 @@ DESCRIPTION >
     Aggregate by pathname and calculate session and hits
 
 SQL >
-    select
-      toDate(timestamp) as date,
-      device,
-      browser,
-      location,
-      pathname,
-      uniqState(session_id) as visits,
-      countState() as hits
-    from
-      analytics_hits
-    group by
-      date,
-      device,
-      browser,
-      location,
-      pathname
+    SELECT
+        toDate(timestamp) AS date,
+        device,
+        browser,
+        location,
+        pathname,
+        uniqState(session_id) AS visits,
+        countState() AS hits
+    FROM analytics_hits
+    GROUP BY date, device, browser, location, pathname
 
-TYPE materialized
+TYPE MATERIALIZED
 DATASOURCE analytics_pages_mv

tinybird/pipes/analytics_sessions.pipe

@@ -3,20 +3,17 @@ DESCRIPTION >
     Aggregate by session_id and calculate session metrics
 
 SQL >
-    select
-      toDate(timestamp) as date,
-      session_id,
-      anySimpleState(device) as device,
-      anySimpleState(browser) as browser,
-      anySimpleState(location) as location,
-      minSimpleState(timestamp) as first_hit,
-      maxSimpleState(timestamp) as latest_hit,
-      countState() as hits
-    from
-      analytics_hits
-    group by
-      date,
-      session_id
+    SELECT
+        toDate(timestamp) AS date,
+        session_id,
+        anySimpleState(device) AS device,
+        anySimpleState(browser) AS browser,
+        anySimpleState(location) AS location,
+        minSimpleState(timestamp) AS first_hit,
+        maxSimpleState(timestamp) AS latest_hit,
+        countState() AS hits
+    FROM analytics_hits
+    GROUP BY date, session_id
 
-TYPE materialized
+TYPE MATERIALIZED
 DATASOURCE analytics_sessions_mv

tinybird/pipes/analytics_sources.pipe

@@ -3,25 +3,18 @@ DESCRIPTION >
     Aggregate by referral and calculate session and hits
 
 SQL >
-    with (select domainWithoutWWW(href) from analytics_hits limit 1) as currenct_domain
-    select
-      toDate(timestamp) AS date,
-      device,
-      browser,
-      location,
-      referrer,
-      uniqState(session_id) AS visits,
-      countState() AS hits
-    from
-      analytics_hits
-    where
-      domainWithoutWWW(referrer) <> currenct_domain
-    group by
-      date,
-      device,
-      browser,
-      location,
-      referrer
+    WITH (SELECT domainWithoutWWW(href) FROM analytics_hits LIMIT 1) AS currenct_domain
+    SELECT
+        toDate(timestamp) AS date,
+        device,
+        browser,
+        location,
+        referrer,
+        uniqState(session_id) AS visits,
+        countState() AS hits
+    FROM analytics_hits
+    WHERE domainWithoutWWW(referrer) != currenct_domain
+    GROUP BY date, device, browser, location, referrer
 
-TYPE materialized
+TYPE MATERIALIZED
 DATASOURCE analytics_sources_mv