def calculate_paths(self, filter: PathFilter, team: Team):
parsed_date_from, parsed_date_to, _ = parse_timestamps(filter=filter,
team_id=team.pk)
event, path_type, start_comparator = self._determine_path_type(
filter.path_type if filter else None)
prop_filters, prop_filter_params = parse_prop_clauses(
filter.properties,
team.pk,
filter_test_accounts=filter.filter_test_accounts)
# Step 0. Event culling subexpression for step 1.
# Make an expression that removes events in a session that are definitely unused.
# For example the 4th, 5th, etc row after a "new_session = 1" or "marked_session_start = 1" row gets removed
excess_row_filter = "("
for i in range(4):
if i > 0:
excess_row_filter += " or "
excess_row_filter += "neighbor(new_session, {}, 0) = 1".format(-i)
if filter and filter.start_point:
excess_row_filter += " or neighbor(marked_session_start, {}, 0) = 1".format(
-i)
excess_row_filter += ")"
paths_query = PATHS_QUERY_FINAL.format(
event_query="event = %(event)s" if event else
"event NOT IN ('$autocapture', '$pageview', '$identify', '$pageleave', '$screen')",
path_type=path_type,
parsed_date_from=parsed_date_from,
parsed_date_to=parsed_date_to,
filters=prop_filters,
marked_session_start="{} = %(start_point)s".format(
start_comparator)
if filter and filter.start_point else "new_session",
excess_row_filter=excess_row_filter,
select_elements_chain=", events.elements_chain as elements_chain"
if event == AUTOCAPTURE_EVENT else "",
group_by_elements_chain=", events.elements_chain"
if event == AUTOCAPTURE_EVENT else "",
)
params: Dict = {
"team_id": team.pk,
"property": "$current_url",
"event": event,
"start_point": filter.start_point,
}
params = {**params, **prop_filter_params}
rows = sync_execute(paths_query, params)
resp: List[Dict[str, str]] = []
for row in rows:
resp.append({
"source": row[0],
"source_id": row[1],
"target": row[2],
"target_id": row[3],
"value": row[4],
})
resp = sorted(resp, key=lambda x: x["value"], reverse=True)
return resp
def list(self, request: Request, *args: Any, **kwargs: Any) -> Response:
team = self.team
filter = Filter(request=request)
if request.GET.get("after"):
filter._date_from = request.GET["after"]
if request.GET.get("before"):
filter._date_to = request.GET["before"]
limit = "LIMIT 101"
conditions, condition_params = determine_event_conditions(
request.GET.dict())
prop_filters, prop_filter_params = parse_prop_clauses(
filter.properties, team.pk)
if request.GET.get("action_id"):
action = Action.objects.get(pk=request.GET["action_id"])
if action.steps.count() == 0:
return Response({"next": False, "results": []})
action_query, params = format_action_filter(action)
prop_filters += " AND {}".format(action_query)
prop_filter_params = {**prop_filter_params, **params}
if prop_filters != "":
query_result = sync_execute(
SELECT_EVENT_WITH_PROP_SQL.format(conditions=conditions,
limit=limit,
filters=prop_filters),
{
"team_id": team.pk,
**condition_params,
**prop_filter_params
},
)
else:
query_result = sync_execute(
SELECT_EVENT_WITH_ARRAY_PROPS_SQL.format(conditions=conditions,
limit=limit),
{
"team_id": team.pk,
**condition_params
},
)
result = ClickhouseEventSerializer(
query_result[0:100],
many=True,
context={
"people": self._get_people(query_result, team),
},
).data
if len(query_result) > 100:
path = request.get_full_path()
reverse = request.GET.get("orderBy", "-timestamp") != "-timestamp"
next_url: Optional[str] = request.build_absolute_uri(
"{}{}{}={}".format(
path,
"&" if "?" in path else "?",
"after" if reverse else "before",
query_result[99][3].strftime("%Y-%m-%dT%H:%M:%S.%fZ"),
))
else:
next_url = None
return Response({"next": next_url, "results": result})
def _serialize_lifecycle(self, entity: Entity, filter: Filter,
team_id: int) -> List[Dict[str, Any]]:
date_from = filter.date_from
if not date_from:
date_from = get_earliest_timestamp(team_id)
interval = filter.interval or "day"
num_intervals, seconds_in_interval = get_time_diff(
interval, filter.date_from, filter.date_to, team_id)
interval_increment, interval_string, sub_interval_string = self.get_interval(
interval)
trunc_func = get_trunc_func_ch(interval)
event_query = ""
event_params: Dict[str, Any] = {}
props_to_filter = [*filter.properties, *entity.properties]
prop_filters, prop_filter_params = parse_prop_clauses(
props_to_filter, team_id)
_, _, date_params = parse_timestamps(filter=filter, team_id=team_id)
if entity.type == TREND_FILTER_TYPE_ACTIONS:
try:
action = Action.objects.get(pk=entity.id)
event_query, event_params = format_action_filter(action)
except:
return []
else:
event_query = "event = %(event)s"
event_params = {"event": entity.id}
result = sync_execute(
LIFECYCLE_SQL.format(
interval=interval_string,
trunc_func=trunc_func,
event_query=event_query,
filters=prop_filters,
sub_interval=sub_interval_string,
),
{
"team_id":
team_id,
"prev_date_from":
(date_from - interval_increment).strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.interval == "hour"
or filter.interval == "minute" else " 00:00:00")),
"num_intervals":
num_intervals,
"seconds_in_interval":
seconds_in_interval,
**event_params,
**date_params,
**prop_filter_params,
},
)
res = []
for val in result:
label = "{} - {}".format(entity.name, val[2])
additional_values = {"label": label, "status": val[2]}
parsed_result = parse_response(val, filter, additional_values)
res.append(parsed_result)
return res
def _retrieve_people_in_period(self, filter: RetentionFilter, team: Team):
period = filter.period
is_first_time_retention = filter.retention_type == RETENTION_FIRST_TIME
trunc_func = get_trunc_func_ch(period)
prop_filters, prop_filter_params = parse_prop_clauses(
filter.properties, team.pk)
returning_entity = filter.returning_entity if filter.selected_interval > 0 else filter.target_entity
target_query, target_params = self._get_condition(filter.target_entity,
table="e")
target_query_formatted = "AND {target_query}".format(
target_query=target_query)
return_query, return_params = self._get_condition(returning_entity,
table="e",
prepend="returning")
return_query_formatted = "AND {return_query}".format(
return_query=return_query)
first_event_sql = (REFERENCE_EVENT_UNIQUE_PEOPLE_PER_PERIOD_SQL
if is_first_time_retention else
REFERENCE_EVENT_PEOPLE_PER_PERIOD_SQL).format(
target_query=target_query_formatted,
filters=prop_filters,
trunc_func=trunc_func,
)
default_event_query = (
DEFAULT_REFERENCE_EVENT_UNIQUE_PEOPLE_PER_PERIOD_SQL
if is_first_time_retention else
DEFAULT_REFERENCE_EVENT_PEOPLE_PER_PERIOD_SQL).format(
target_query=target_query_formatted,
filters=prop_filters,
trunc_func=trunc_func,
)
date_from = filter.date_from + filter.selected_interval * filter.period_increment
date_to = filter.date_to
new_data = filter._data
new_data.update({
"total_intervals":
filter.total_intervals - filter.selected_interval
})
filter = RetentionFilter(data=new_data)
query_result = sync_execute(
RETENTION_PEOPLE_PER_PERIOD_SQL.format(
returning_query=return_query_formatted,
filters=prop_filters,
first_event_sql=first_event_sql,
first_event_default_sql=default_event_query,
trunc_func=trunc_func,
),
{
"team_id":
team.pk,
"start_date":
date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"end_date":
date_to.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"offset":
filter.offset,
"limit":
100,
"period":
period,
**target_params,
**return_params,
**prop_filter_params,
},
)
people_dict = {}
from posthog.api.person import PersonSerializer
for person in Person.objects.filter(
team_id=team.pk, uuid__in=[val[0] for val in query_result]):
people_dict.update(
{str(person.uuid): PersonSerializer(person).data})
result = self.process_people_in_period(filter, query_result,
people_dict)
return result
def execute_query(query: str) -> Any:
from ee.clickhouse.client import sync_execute
return sync_execute(query)
def get_property_values_for_key(key: str, team: Team, value: Optional[str] = None):
if value:
return sync_execute(
SELECT_PROP_VALUES_SQL_WITH_FILTER, {"team_id": team.pk, "key": key, "value": "%{}%".format(value)}
)
return sync_execute(SELECT_PROP_VALUES_SQL, {"team_id": team.pk, "key": key})
def run(self, *args, **kwargs) -> SessionRecordingQueryResult:
query, query_params = self.get_query()
query_results = sync_execute(query, query_params)
session_recordings = self._data_to_return(query_results)
return self._paginate_results(session_recordings)
def get_persons(team_id: int):
result = sync_execute(GET_PERSON_SQL, {"team_id": team_id})
return ClickhousePersonSerializer(result, many=True).data
def get_person_distinct_ids(team_id: int):
result = sync_execute(GET_DISTINCT_IDS_SQL, {"team_id": team_id})
return ClickhousePersonDistinctIdSerializer(result, many=True).data
def _exec_query(self) -> List[Tuple]:
query = self.get_query()
return sync_execute(query, self.params)
def distinct_ids_exist(team_id: int, ids: List[str]) -> bool:
return bool(
sync_execute(
PERSON_DISTINCT_ID_EXISTS_SQL.format([str(id) for id in ids]),
{"team_id": team_id})[0][0])
def _team_ids(self):
return list(
sorted(row[0] for row in sync_execute(
"SELECT DISTINCT team_id FROM person_distinct_id")))
def _get_people():
return [Person(p) for p in sync_execute("select * from person")]
def _get_events():
return sync_execute("select * from events")
def perform_query(self):
sql = self._configure_sql()
results = sync_execute(sql, self.params)
summary = self._summarize_data(results)
return summary
def update_person_properties(team_id: int, id: str, properties: Dict) -> None:
sync_execute(UPDATE_PERSON_PROPERTIES, {
"team_id": team_id,
"id": id,
"properties": json.dumps(properties)
})
def _run_formula_query(self, filter: Filter, team_id: int):
letters = [chr(65 + i) for i in range(0, len(filter.entities))]
queries = []
params: Dict[str, Any] = {}
for idx, entity in enumerate(filter.entities):
sql, entity_params, _ = self._get_sql_for_entity(
filter, entity, team_id) # type: ignore
sql = sql.replace("%(", "%({}_".format(idx))
entity_params = {
"{}_{}".format(idx, key): value
for key, value in entity_params.items()
}
queries.append(sql)
params = {**params, **entity_params}
breakdown_value = (", sub_A.breakdown_value"
if filter.breakdown_type == "cohort" else
", trim(BOTH '\"' FROM sub_A.breakdown_value)")
is_aggregate = filter.display in [TRENDS_TABLE, TRENDS_PIE]
sql = """SELECT
{date_select}
arrayMap(({letters_select}) -> {formula}, {selects})
{breakdown_value}
FROM ({first_query}) as sub_A
{queries}
""".format(
date_select="'' as date," if is_aggregate else "sub_A.date,",
letters_select=", ".join(letters),
formula=filter.
formula, # formula is properly escaped in the filter
# Need to wrap aggregates in arrays so we can still use arrayMap
selects=", ".join([
("[sub_{}.data]" if is_aggregate else "sub_{}.data").format(
letters[i]) for i in range(0, len(filter.entities))
]),
breakdown_value=breakdown_value if filter.breakdown else "",
first_query=queries[0],
queries="".join([
"FULL OUTER JOIN ({query}) as sub_{letter} ON sub_A.breakdown_value = sub_{letter}.breakdown_value "
.format(query=query, letter=letters[i + 1])
for i, query in enumerate(queries[1:])
]) if filter.breakdown else "".join([
" CROSS JOIN ({}) as sub_{}".format(query, letters[i + 1])
for i, query in enumerate(queries[1:])
]),
)
result = sync_execute(sql, params)
response = []
for item in result:
additional_values: Dict[str, Any] = {
"label": self._label(filter, item, team_id),
}
if is_aggregate:
additional_values["data"] = []
additional_values["aggregated_value"] = item[1][0]
else:
additional_values["data"] = [
round(number, 2) if not math.isnan(number)
and not math.isinf(number) else 0.0 for number in item[1]
]
if filter.display == TRENDS_CUMULATIVE:
additional_values["data"] = list(
accumulate(additional_values["data"]))
additional_values["count"] = float(sum(additional_values["data"]))
response.append(parse_response(item, filter, additional_values))
return response
def get_events():
events = sync_execute(GET_EVENTS_SQL)
return ClickhouseEventSerializer(events, many=True, context={"elements": None, "people": None}).data
def test_person_properties_filter(self):
filter = Filter(
data={
"date_from":
"2021-05-01 00:00:00",
"date_to":
"2021-05-07 00:00:00",
"events": [
{
"id": "viewed",
"order": 0
},
],
"properties": [
{
"key": "email",
"value": "@posthog.com",
"operator": "not_icontains",
"type": "person"
},
{
"key": "key",
"value": "val"
},
],
})
entity = Entity({"id": "viewed", "type": "events"})
global_prop_query, global_prop_query_params = TrendsEventQuery(
filter=filter, entity=entity, team_id=self.team.pk).get_query()
sync_execute(global_prop_query, global_prop_query_params)
filter = Filter(
data={
"date_from": "2021-05-01 00:00:00",
"date_to": "2021-05-07 00:00:00",
"events": [
{
"id": "viewed",
"order": 0
},
],
})
entity = Entity({
"id":
"viewed",
"type":
"events",
"properties": [
{
"key": "email",
"value": "@posthog.com",
"operator": "not_icontains",
"type": "person"
},
{
"key": "key",
"value": "val"
},
],
})
entity_prop_query, entity_prop_query_params = TrendsEventQuery(
filter=filter, entity=entity, team_id=self.team.pk).get_query()
# global queries and enttiy queries should be the same
self.assertEqual(sqlparse.format(global_prop_query, reindent=True),
sqlparse.format(entity_prop_query, reindent=True))
sync_execute(entity_prop_query, entity_prop_query_params)
def get_events_by_team(team_id: Union[str, int]):
events = sync_execute(GET_EVENTS_BY_TEAM_SQL, {"team_id": str(team_id)})
return ClickhouseEventSerializer(events, many=True, context={"elements": None, "people": None}).data
def _retrieve_people(self, filter: RetentionFilter, team: Team):
period = filter.period
is_first_time_retention = filter.retention_type == RETENTION_FIRST_TIME
trunc_func = get_trunc_func_ch(period)
prop_filters, prop_filter_params = parse_prop_clauses(
filter.properties, team.pk)
returning_entity = filter.returning_entity if filter.selected_interval > 0 else filter.target_entity
target_query, target_params = self._get_condition(filter.target_entity,
table="e")
target_query_formatted = "AND {target_query}".format(
target_query=target_query)
return_query, return_params = self._get_condition(returning_entity,
table="e",
prepend="returning")
return_query_formatted = "AND {return_query}".format(
return_query=return_query)
reference_event_query = (REFERENCE_EVENT_UNIQUE_SQL
if is_first_time_retention else
REFERENCE_EVENT_SQL).format(
target_query=target_query_formatted,
filters=prop_filters,
trunc_func=trunc_func,
)
reference_date_from = filter.date_from
reference_date_to = filter.date_from + filter.period_increment
date_from = filter.date_from + filter.selected_interval * filter.period_increment
date_to = date_from + filter.period_increment
result = sync_execute(
RETENTION_PEOPLE_SQL.format(
reference_event_query=reference_event_query,
target_query=return_query_formatted,
filters=prop_filters),
{
"team_id":
team.pk,
"start_date":
date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"end_date":
date_to.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"reference_start_date":
reference_date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"reference_end_date":
reference_date_to.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"offset":
filter.offset,
**target_params,
**return_params,
**prop_filter_params,
},
)
people = Person.objects.filter(team_id=team.pk,
uuid__in=[val[0] for val in result])
from posthog.api.person import PersonSerializer
return PersonSerializer(people, many=True).data
def _format_breakdown_query(self, entity: Entity, filter: Filter,
team: Team) -> List[Dict[str, Any]]:
params = {"team_id": team.pk}
interval_annotation = get_interval_annotation_ch(filter.interval)
num_intervals, seconds_in_interval = get_time_diff(
filter.interval or "day", filter.date_from, filter.date_to)
parsed_date_from, parsed_date_to = parse_timestamps(filter=filter)
props_to_filter = [*filter.properties, *entity.properties]
prop_filters, prop_filter_params = parse_prop_clauses(
"uuid", props_to_filter, team)
aggregate_operation, join_condition, math_params = self._process_math(
entity)
action_query = ""
action_params: Dict = {}
if entity.type == TREND_FILTER_TYPE_ACTIONS:
action = Action.objects.get(pk=entity.id)
action_query, action_params = format_action_filter(action)
null_sql = NULL_BREAKDOWN_SQL.format(
interval=interval_annotation,
seconds_in_interval=seconds_in_interval,
num_intervals=num_intervals,
date_to=((filter.date_to
or timezone.now())).strftime("%Y-%m-%d %H:%M:%S"),
)
params = {**params, **math_params, **prop_filter_params}
top_elements_array = []
if filter.breakdown_type == "cohort":
breakdown = filter.breakdown if filter.breakdown and isinstance(
filter.breakdown, list) else []
if "all" in breakdown:
params = {**params, "event": entity.id, **action_params}
null_sql = NULL_SQL.format(
interval=interval_annotation,
seconds_in_interval=seconds_in_interval,
num_intervals=num_intervals,
date_to=((filter.date_to or
timezone.now())).strftime("%Y-%m-%d %H:%M:%S"),
)
conditions = BREAKDOWN_CONDITIONS_SQL.format(
parsed_date_from=parsed_date_from,
parsed_date_to=parsed_date_to,
actions_query="AND uuid IN ({})".format(action_query)
if action_query else "",
event_filter="AND event = %(event)s"
if not action_query else "",
filters="{filters}".format(
filters=prop_filters) if props_to_filter else "",
)
breakdown_query = BREAKDOWN_DEFAULT_SQL.format(
null_sql=null_sql,
conditions=conditions,
event_join=join_condition,
aggregate_operation=aggregate_operation,
interval_annotation=interval_annotation,
)
else:
cohort_queries, cohort_ids, cohort_params = self._format_breakdown_cohort_join_query(
breakdown, team)
params = {
**params, "values": cohort_ids,
"event": entity.id,
**action_params,
**cohort_params
}
breakdown_filter = BREAKDOWN_COHORT_JOIN_SQL.format(
cohort_queries=cohort_queries,
parsed_date_from=parsed_date_from,
parsed_date_to=parsed_date_to,
actions_query="AND uuid IN ({})".format(action_query)
if action_query else "",
event_filter="AND event = %(event)s"
if not action_query else "",
filters="{filters}".format(
filters=prop_filters) if props_to_filter else "",
)
breakdown_query = BREAKDOWN_QUERY_SQL.format(
null_sql=null_sql,
breakdown_filter=breakdown_filter,
event_join=join_condition,
aggregate_operation=aggregate_operation,
interval_annotation=interval_annotation,
)
elif filter.breakdown_type == "person":
top_elements_array = self._get_top_elements(
TOP_PERSON_PROPS_ARRAY_OF_KEY_SQL, filter, parsed_date_from,
parsed_date_to, team)
params = {
**params,
"values": top_elements_array,
"key": filter.breakdown,
"event": entity.id,
**action_params,
}
breakdown_filter = BREAKDOWN_PERSON_PROP_JOIN_SQL.format(
parsed_date_from=parsed_date_from,
parsed_date_to=parsed_date_to,
actions_query="AND uuid IN ({})".format(action_query)
if action_query else "",
event_filter="AND event = %(event)s"
if not action_query else "",
)
breakdown_query = BREAKDOWN_QUERY_SQL.format(
null_sql=null_sql,
breakdown_filter=breakdown_filter,
event_join=join_condition,
aggregate_operation=aggregate_operation,
interval_annotation=interval_annotation,
)
else:
top_elements_array = self._get_top_elements(
TOP_ELEMENTS_ARRAY_OF_KEY_SQL, filter, parsed_date_from,
parsed_date_to, team)
params = {
**params,
"values": top_elements_array,
"key": filter.breakdown,
"event": entity.id,
**action_params,
}
breakdown_filter = BREAKDOWN_PROP_JOIN_SQL.format(
parsed_date_from=parsed_date_from,
parsed_date_to=parsed_date_to,
actions_query="AND uuid IN ({})".format(action_query)
if action_query else "",
event_filter="AND event = %(event)s"
if not action_query else "",
filters="{filters}".format(
filters=prop_filters) if props_to_filter else "",
)
breakdown_query = BREAKDOWN_QUERY_SQL.format(
null_sql=null_sql,
breakdown_filter=breakdown_filter,
event_join=join_condition,
aggregate_operation=aggregate_operation,
interval_annotation=interval_annotation,
)
try:
result = sync_execute(breakdown_query, params)
except:
result = []
parsed_results = []
for idx, stats in enumerate(result):
breakdown_value = stats[
2] if not filter.breakdown_type == "cohort" else ""
stripped_value = breakdown_value.strip('"') if isinstance(
breakdown_value, str) else breakdown_value
extra_label = self._determine_breakdown_label(
idx, filter.breakdown_type, filter.breakdown, stripped_value)
label = "{} - {}".format(entity.name, extra_label)
additional_values = {
"label":
label,
"breakdown_value":
filter.breakdown[idx]
if isinstance(filter.breakdown, list) else filter.breakdown
if filter.breakdown_type == "cohort" else stripped_value,
}
parsed_result = self._parse_response(stats, filter,
additional_values)
parsed_results.append(parsed_result)
return parsed_results
def _execute_sql(
self,
filter: RetentionFilter,
team: Team,
) -> Dict[Tuple[int, int], Dict[str, Any]]:
period = filter.period
prop_filters, prop_filter_params = parse_prop_clauses(
filter.properties, team.pk)
target_entity = filter.target_entity
returning_entity = filter.returning_entity
is_first_time_retention = filter.retention_type == RETENTION_FIRST_TIME
date_from = filter.date_from
date_to = filter.date_to
target_query = ""
target_params: Dict = {}
trunc_func = get_trunc_func_ch(period)
target_query, target_params = self._get_condition(target_entity,
table="e")
returning_query, returning_params = self._get_condition(
returning_entity, table="e", prepend="returning")
target_query_formatted = "AND {target_query}".format(
target_query=target_query)
returning_query_formatted = "AND {returning_query}".format(
returning_query=returning_query)
reference_event_sql = (REFERENCE_EVENT_UNIQUE_SQL
if is_first_time_retention else
REFERENCE_EVENT_SQL).format(
target_query=target_query_formatted,
filters=prop_filters,
trunc_func=trunc_func,
)
target_condition, _ = self._get_condition(target_entity,
table="reference_event")
if is_first_time_retention:
target_condition = target_condition.replace(
"reference_event.uuid", "reference_event.min_uuid")
target_condition = target_condition.replace(
"reference_event.event", "reference_event.min_event")
returning_condition, _ = self._get_condition(returning_entity,
table="event",
prepend="returning")
result = sync_execute(
RETENTION_SQL.format(
target_query=target_query_formatted,
returning_query=returning_query_formatted,
filters=prop_filters,
trunc_func=trunc_func,
extra_union="UNION ALL {} ".format(reference_event_sql),
reference_event_sql=reference_event_sql,
target_condition=target_condition,
returning_condition=returning_condition,
),
{
"team_id":
team.pk,
"start_date":
date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"end_date":
date_to.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"reference_start_date":
date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"reference_end_date":
((date_from + filter.period_increment) if filter.display
== TRENDS_LINEAR else date_to).strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
**prop_filter_params,
**target_params,
**returning_params,
"period":
period,
},
)
initial_interval_result = sync_execute(
INITIAL_INTERVAL_SQL.format(
reference_event_sql=reference_event_sql),
{
"team_id":
team.pk,
"start_date":
date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"end_date":
date_to.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"reference_start_date":
date_from.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
"reference_end_date":
((date_from + filter.period_increment) if filter.display
== TRENDS_LINEAR else date_to).strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.period == "Hour" else " 00:00:00")),
**prop_filter_params,
**target_params,
**returning_params,
"period":
period,
},
)
result_dict = {}
for initial_res in initial_interval_result:
result_dict.update({
(initial_res[0], 0): {
"count": initial_res[1],
"people": []
}
})
for res in result:
result_dict.update({
(res[0], res[1]): {
"count": res[2],
"people": []
}
})
return result_dict
def _format_normal_query(self, entity: Entity, filter: Filter,
team: Team) -> List[Dict[str, Any]]:
interval_annotation = get_interval_annotation_ch(filter.interval)
num_intervals, seconds_in_interval = get_time_diff(
filter.interval or "day", filter.date_from, filter.date_to)
parsed_date_from, parsed_date_to = parse_timestamps(filter=filter)
props_to_filter = [*filter.properties, *entity.properties]
prop_filters, prop_filter_params = parse_prop_clauses(
"uuid", props_to_filter, team)
aggregate_operation, join_condition, math_params = self._process_math(
entity)
params: Dict = {"team_id": team.pk}
params = {**params, **prop_filter_params, **math_params}
if entity.type == TREND_FILTER_TYPE_ACTIONS:
try:
action = Action.objects.get(pk=entity.id)
action_query, action_params = format_action_filter(action)
params = {**params, **action_params}
content_sql = VOLUME_ACTIONS_SQL.format(
interval=interval_annotation,
timestamp="timestamp",
team_id=team.pk,
actions_query=action_query,
parsed_date_from=(parsed_date_from or ""),
parsed_date_to=(parsed_date_to or ""),
filters="{filters}".format(
filters=prop_filters) if props_to_filter else "",
event_join=join_condition,
aggregate_operation=aggregate_operation,
)
except:
return []
else:
content_sql = VOLUME_SQL.format(
interval=interval_annotation,
timestamp="timestamp",
team_id=team.pk,
parsed_date_from=(parsed_date_from or ""),
parsed_date_to=(parsed_date_to or ""),
filters="{filters}".format(
filters=prop_filters) if props_to_filter else "",
event_join=join_condition,
aggregate_operation=aggregate_operation,
)
params = {**params, "event": entity.id}
null_sql = NULL_SQL.format(
interval=interval_annotation,
seconds_in_interval=seconds_in_interval,
num_intervals=num_intervals,
date_to=((filter.date_to
or timezone.now())).strftime("%Y-%m-%d %H:%M:%S"),
)
final_query = AGGREGATE_SQL.format(null_sql=null_sql,
content_sql=content_sql)
try:
result = sync_execute(final_query, params)
except:
result = []
parsed_results = []
for _, stats in enumerate(result):
parsed_result = self._parse_response(stats, filter)
parsed_results.append(parsed_result)
return parsed_results
def _format_normal_query(self, entity: Entity, filter: Filter,
team_id: int) -> List[Dict[str, Any]]:
interval_annotation = get_trunc_func_ch(filter.interval)
num_intervals, seconds_in_interval = get_time_diff(filter.interval
or "day",
filter.date_from,
filter.date_to,
team_id=team_id)
parsed_date_from, parsed_date_to, _ = parse_timestamps(filter=filter,
team_id=team_id)
props_to_filter = [*filter.properties, *entity.properties]
prop_filters, prop_filter_params = parse_prop_clauses(
props_to_filter, team_id)
aggregate_operation, join_condition, math_params = process_math(entity)
params: Dict = {"team_id": team_id}
params = {**params, **prop_filter_params, **math_params}
content_sql_params = {
"interval": interval_annotation,
"timestamp": "timestamp",
"team_id": team_id,
"parsed_date_from": parsed_date_from,
"parsed_date_to": parsed_date_to,
"filters": prop_filters,
"event_join": join_condition,
"aggregate_operation": aggregate_operation,
}
entity_params, entity_format_params = self._populate_entity_params(
entity)
params = {**params, **entity_params}
content_sql_params = {**content_sql_params, **entity_format_params}
if filter.display == TRENDS_TABLE or filter.display == TRENDS_PIE:
agg_query = self._determine_single_aggregate_query(filter, entity)
content_sql = agg_query.format(**content_sql_params)
try:
result = sync_execute(content_sql, params)
except:
result = []
return [{
"aggregated_value":
result[0][0] if result and len(result) else 0
}]
else:
content_sql = self._determine_trend_aggregate_query(filter, entity)
content_sql = content_sql.format(**content_sql_params)
null_sql = NULL_SQL.format(
interval=interval_annotation,
seconds_in_interval=seconds_in_interval,
num_intervals=num_intervals,
date_to=filter.date_to.strftime("%Y-%m-%d %H:%M:%S"),
)
final_query = AGGREGATE_SQL.format(null_sql=null_sql,
content_sql=content_sql)
try:
result = sync_execute(final_query, params)
except:
result = []
parsed_results = []
for _, stats in enumerate(result):
parsed_result = parse_response(stats, filter)
parsed_results.append(parsed_result)
return parsed_results
def get_properties(self, request: Request):
rows = sync_execute(GET_PERSON_PROPERTIES_COUNT,
{"team_id": self.team.pk})
return [{"name": name, "count": count} for name, count in rows]
def get_people(
self,
filter: Filter,
team_id: int,
target_date: datetime,
lifecycle_type: str,
limit: int = 100,
):
entity = filter.entities[0]
date_from = filter.date_from
if not date_from:
date_from = get_earliest_timestamp(team_id)
interval = filter.interval or "day"
num_intervals, seconds_in_interval = get_time_diff(interval,
filter.date_from,
filter.date_to,
team_id=team_id)
interval_increment, interval_string, sub_interval_string = self.get_interval(
interval)
trunc_func = get_trunc_func_ch(interval)
event_query = ""
event_params: Dict[str, Any] = {}
_, _, date_params = parse_timestamps(filter=filter, team_id=team_id)
if entity.type == TREND_FILTER_TYPE_ACTIONS:
try:
action = Action.objects.get(pk=entity.id)
event_query, event_params = format_action_filter(action)
except:
return []
else:
event_query = "event = %(event)s"
event_params = {"event": entity.id}
props_to_filter = [*filter.properties, *entity.properties]
prop_filters, prop_filter_params = parse_prop_clauses(
props_to_filter, team_id)
result = sync_execute(
LIFECYCLE_PEOPLE_SQL.format(
interval=interval_string,
trunc_func=trunc_func,
event_query=event_query,
filters=prop_filters,
sub_interval=sub_interval_string,
),
{
"team_id":
team_id,
"prev_date_from":
(date_from - interval_increment).strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.interval == "hour"
or filter.interval == "minute" else " 00:00:00")),
"num_intervals":
num_intervals,
"seconds_in_interval":
seconds_in_interval,
**event_params,
**date_params,
**prop_filter_params,
"status":
lifecycle_type,
"target_date":
target_date.strftime("%Y-%m-%d{}".format(
" %H:%M:%S" if filter.interval == "hour"
or filter.interval == "minute" else " 00:00:00")),
"offset":
filter.offset,
"limit":
limit,
},
)
people = get_persons_by_uuids(team_id=team_id,
uuids=[p[0] for p in result])
people = people.prefetch_related(
Prefetch("persondistinctid_set", to_attr="distinct_ids_cache"))
from posthog.api.person import PersonSerializer
return PersonSerializer(people, many=True).data
def test_create_trends_cohort(self, _insert_cohort_from_query):
_create_person(team_id=self.team.pk, distinct_ids=["blabla"])
with freeze_time("2021-01-01 00:06:34"):
_create_event(
team=self.team,
event="$pageview",
distinct_id="blabla",
properties={"$math_prop": 1},
timestamp="2021-01-01T12:00:00Z",
)
with freeze_time("2021-01-02 00:06:34"):
_create_event(
team=self.team,
event="$pageview",
distinct_id="blabla",
properties={"$math_prop": 4},
timestamp="2021-01-01T12:00:00Z",
)
response = self.client.post(
f"/api/projects/{self.team.id}/cohorts/?interval=day&display=ActionsLineGraph&events=%5B%7B%22id%22%3A%22%24pageview%22%2C%22name%22%3A%22%24pageview%22%2C%22type%22%3A%22events%22%2C%22order%22%3A0%7D%5D&properties=%5B%5D&entity_id=%24pageview&entity_type=events&date_from=2021-01-01&date_to=2021-01-01&label=%24pageview",
{"name": "test", "is_static": True},
).json()
cohort_id = response["id"]
_insert_cohort_from_query.assert_called_once_with(
cohort_id,
"TRENDS",
{
"date_from": "2021-01-01",
"date_to": "2021-01-01",
"display": "ActionsLineGraph",
"events": [
{
"id": "$pageview",
"type": "events",
"order": 0,
"name": "$pageview",
"custom_name": None,
"math": None,
"math_property": None,
"math_group_type_index": None,
"properties": [],
}
],
"entity_id": "$pageview",
"entity_type": "events",
"insight": "TRENDS",
"interval": "day",
},
entity_data={
"id": "$pageview",
"type": "events",
"order": None,
"name": "$pageview",
"custom_name": None,
"math": None,
"math_property": None,
"math_group_type_index": None,
"properties": [],
},
)
insert_cohort_from_query(
cohort_id,
"TRENDS",
{
"date_from": "2021-01-01",
"date_to": "2021-01-01",
"display": "ActionsLineGraph",
"events": [
{
"id": "$pageview",
"type": "events",
"order": 0,
"name": "$pageview",
"math": None,
"math_property": None,
"math_group_type_index": None,
"properties": [],
}
],
"entity_id": "$pageview",
"entity_type": "events",
"insight": "TRENDS",
"interval": "day",
},
entity_data={
"id": "$pageview",
"type": "events",
"order": 0,
"name": "$pageview",
"math": None,
"math_property": None,
"math_group_type_index": None,
"properties": [],
},
)
cohort = Cohort.objects.get(pk=cohort_id)
people = Person.objects.filter(cohort__id=cohort.pk)
self.assertEqual(cohort.errors_calculating, 0)
self.assertEqual(
len(people),
1,
{
"a": sync_execute(
"select person_id from person_static_cohort where team_id = {} and cohort_id = {} ".format(
self.team.id, cohort.pk
)
),
"b": sync_execute(
"select person_id from person_static_cohort FINAL where team_id = {} and cohort_id = {} ".format(
self.team.id, cohort.pk
)
),
},
)
def test_prop_cohort_basic_event_days(self):
_create_person(distinct_ids=["some_other_id"],
team_id=self.team.pk,
properties={"$some_prop": "something"})
_create_person(
distinct_ids=["some_id"],
team_id=self.team.pk,
properties={
"$some_prop": "something",
"$another_prop": "something"
},
)
_create_event(
event="$pageview",
team=self.team,
distinct_id="some_id",
properties={"attr": "some_val"},
timestamp=datetime(2020, 1, 9, 12, 0, 1),
)
_create_event(
event="$pageview",
team=self.team,
distinct_id="some_other_id",
properties={"attr": "some_val"},
timestamp=datetime(2020, 1, 5, 12, 0, 1),
)
with freeze_time("2020-01-10"):
cohort1 = Cohort.objects.create(
team=self.team,
groups=[{
"event_id": "$pageview",
"days": 1
}],
name="cohort1",
)
filter = Filter(
data={
"properties": [{
"key": "id",
"value": cohort1.pk,
"type": "cohort"
}],
})
query, params = parse_prop_clauses(filter.properties, self.team.pk)
final_query = "SELECT uuid FROM events WHERE team_id = %(team_id)s {}".format(
query)
result = sync_execute(final_query, {
**params, "team_id": self.team.pk
})
self.assertEqual(len(result), 1)
cohort2 = Cohort.objects.create(
team=self.team,
groups=[{
"event_id": "$pageview",
"days": 7
}],
name="cohort2",
)
filter = Filter(
data={
"properties": [{
"key": "id",
"value": cohort2.pk,
"type": "cohort"
}],
})
query, params = parse_prop_clauses(filter.properties, self.team.pk)
final_query = "SELECT uuid FROM events WHERE team_id = %(team_id)s {}".format(
query)
result = sync_execute(final_query, {
**params, "team_id": self.team.pk
})
self.assertEqual(len(result), 2)
def get_elements(event_id: Union[int, UUID]) -> List[Element]:
return chain_to_elements(
sync_execute("select elements_chain from events where uuid = %(id)s", {"id": event_id})[0][0]
)
