omg la ram

This commit is contained in:
Evann Regnault 2024-05-31 02:04:24 +02:00
parent 4addf5448d
commit 3ef595e195
4 changed files with 90 additions and 3 deletions

View file

@ -113,3 +113,53 @@ class Elastic_Testing(DB_Testing):
}
}
return f"Got {driver.search(index=INDEX, body=query).body['aggregations']['attachment_count']['value']}"
def update_add_replies_per_last_name(self, driver: Elasticsearch) -> str:
agg_query = {
"size": 0,
"aggs": {
"reply_count_by_last_name": {
"terms": {
"script": {
"source": "String name = doc['sender_name.keyword'].value; String[] parts = /\\s+/.split(name); return parts[parts.length - 1]",
"lang": "painless"
},
"size": 1000
},
"aggs": {
"total_reply_count": {
"sum": {
"field": "reply_count"
}
}
}
}
}
}
agg_results = driver.search(index=INDEX, body=agg_query)
buckets = agg_results['aggregations']['reply_count_by_last_name']['buckets']
reply_counts = {bucket['key']: bucket['total_reply_count']['value'] for bucket in buckets}
update_query = {
"script": {
"source": """
String name = ctx._source.sender_name;
String[] parts = /\\s+/.split(name);
String lastName = parts[parts.length - 1];
if (params.replyCounts.containsKey(lastName)) {
ctx._source.total_reply_count = params.replyCounts.get(lastName);
}
""",
"lang": "painless",
"params": {
"replyCounts": reply_counts
}
},
"query": {
"match_all": {}
}
}
driver.update_by_query(index=INDEX, body=update_query, conflicts="proceed")

View file

@ -29,6 +29,8 @@ class PSQL_Thread(Data_Thread):
class PSQL_Testing(DB_Testing):
driver_name = "PSQL"
def __init__(self) -> None:
super().__init__()
self.data = [PSQL_Thread(x) for x in Threads]
@ -110,3 +112,28 @@ class PSQL_Testing(DB_Testing):
driver.commit()
return f"Got {cursor.fetchone()[0]}"
def update_add_replies_per_last_name(self, driver: connection) -> str:
cursor = driver.cursor()
alterQuery = """
ALTER TABLE thread ADD COLUMN last_name_replies INT;
"""
updateQuery = """
WITH last_name_sums AS (
SELECT
recipient_name,
SUM(reply_count) OVER (PARTITION BY SPLIT_PART(recipient_name, ' ', -1)) AS total_replies
FROM
thread
)
UPDATE thread AS t
SET last_name_replies = ln_sums.total_replies
FROM last_name_sums AS ln_sums
WHERE t.recipient_name = ln_sums.recipient_name;
"""
unalterQuery = """
ALTER TABLE thread DROP COLUMN last_name_replies;
"""
cursor.execute(alterQuery)
cursor.execute(updateQuery)
cursor.execute(unalterQuery)
driver.commit()

View file

@ -5,7 +5,7 @@ from utils.utils import time_func
from time import sleep
T = TypeVar('T')
NUM_RUNS = 10000
NUM_RUNS = 10
class DB_Testing(Generic[T]):
driver_name: str
@ -50,6 +50,10 @@ class DB_Testing(Generic[T]):
def get_sum_attachment_less_5(self, driver: T) -> str:
pass
@abstractmethod
def update_add_replies_per_last_name(self, driver: T) -> str:
pass
def do_tests(self, driver: T) -> Dict[str, float]:
stats = {}
@ -63,9 +67,15 @@ class DB_Testing(Generic[T]):
sleep(1) # To wait for any indexing to avoid race conditions somehow
# Multiple
# Search
stats.update(time_func(f"[{self.driver_name}] Count mails with more than 5 attachments and Mr/Mrs", "attach5_mr_mrs", lambda : self.attach5_mr_mrs(driver), NUM_RUNS))
stats.update(time_func(f"[{self.driver_name}] Sum the number of attachment from mail that have less than 5 attachments", "get_sum_attachment_less_5", lambda : self.get_sum_attachment_less_5(driver), NUM_RUNS))
stats.update(time_func(f"[{self.driver_name}] Count mails that have a lorem sentence", "search_lorem", lambda : self.search_lorem(driver), NUM_RUNS))
stats.update(time_func(f"[{self.driver_name}] Count mails that have a mail that ends with microsoft.com", "search_mails_ends_microsoftcom", lambda : self.search_mails_ends_microsoftcom(driver), NUM_RUNS))
# Update
stats.update(time_func(f"[{self.driver_name}] Update all thread to add a field with the total replies per last name", "update_add_replies_per_last_name", lambda : self.update_add_replies_per_last_name(driver), NUM_RUNS))
print("\n")
return stats

BIN
plot.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 65 KiB