Cleanup

Browse Source
This commit is contained in:
retoor 2024-11-26 09:21:46 +01:00
parent 38c13427c6
commit a6697fab58
3 changed files with 29 additions and 56 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@ -6,9 +6,9 @@ Also this data will be used for retoor9b, the newest AI hype! You're still using
## Statistics by last build ## Statistics by last build
Click here for latest [dataset](https://retoor.molodetz.nl/retoor/drstats/src/branch/main/export/0_dataset.txt). Click here for latest [dataset](https://retoor.molodetz.nl/retoor/drstats/src/branch/main/export/0_dataset.txt) optimized for training LLM's like retoor9b.
Click here for latest [graphs compilaiton](https://retoor.molodetz.nl/retoor/drstats/src/branch/main/export/1_graphs_compliation.png). Click here for latest [graphs compilation](https://retoor.molodetz.nl/retoor/drstats/src/branch/main/export/1_graphs_compliation.png).
Click here for all generated [data](https://retoor.molodetz.nl/retoor/drstats/src/branch/main/export). It's a big dataset containing data for LLM's to train on, graphs per user or overal statistics and json files with all made observations. Click here for all generated [data](https://retoor.molodetz.nl/retoor/drstats/src/branch/main/export). It's a big dataset containing data for LLM's to train on, graphs per user or overal statistics and json files with all made observations.

5
setup.cfg
View File

@ -25,8 +25,5 @@ where = src
[options.entry_points] [options.entry_points]
console_scripts = console_scripts =
dr.sync = drstats.sync:sync dr.sync = drstats.sync:sync
dr.rant_stats_per_day = drstats.statistics:rant_stats_per_day dr.stats = drstats.statistics:rant_stats_all
dr.rant_stats_per_weekday = drstats.statistics:rant_stats_per_weekday
dr.rant_stats_per_hour = drstats.statistics:rant_stats_per_hour
dr.stats_all = drstats.statistics:rant_stats_all
dr.dataset = drstats.dataset:dump dr.dataset = drstats.dataset:dump

76
src/drstats/statistics.py
View File

@ -1,38 +1,26 @@
from drstats.db import get_db, Db, get_users from drstats.db import get_db, Db, get_users
from drstats import sync
import asyncio import asyncio
from drstats.duration import Duration from drstats.duration import Duration
from time import sleep
import importlib
def plt_reset():
pass
def new_plot():
_figure_count = 0
def figure_inc():
import matplotlib.pyplot as plt import matplotlib.pyplot as plt
plt.clf() plt.clf()
plt.cla() plt.cla()
plt.close(0) plt.close(0)
plt.figure(figsize=(8, 8)) plt.figure(figsize=(8, 8))
return plt return plt
# return pyplot.figure(_figure_count)
def get_date_range(): def get_date_range():
db = get_db() with Db() as db:
for record in db.query( record = list(db.query(
"SELECT min(date(created)) as start_date, max(date(created)) as end_date FROM rants" "SELECT min(date(created)) as start_date, max(date(created)) as end_date FROM rants"
): ))[0]
db.close()
return record["start_date"], record["end_date"] return record["start_date"], record["end_date"]
db.close()
def get_date_range_str(): def get_date_range_str():
@ -42,7 +30,7 @@ def get_date_range_str():
async def rant_stats_per_day(): async def rant_stats_per_day():
with Duration("Rant stats per day"): with Duration("Rant stats per day"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -62,14 +50,13 @@ async def rant_stats_per_day():
plt.title("Rants per day") plt.title("Rants per day")
plt.legend() plt.legend()
plt.savefig(f"export/rants_per_day_{get_date_range_str()}.png") plt.savefig(f"export/rants_per_day_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def comment_stats_per_day(): async def comment_stats_per_day():
with Duration("Comment stats per day"): with Duration("Comment stats per day"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -89,13 +76,12 @@ async def comment_stats_per_day():
plt.title("Comments per day") plt.title("Comments per day")
plt.legend() plt.legend()
plt.savefig(f"export/comments_per_day_{get_date_range_str()}.png") plt.savefig(f"export/comments_per_day_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def rant_stats_per_weekday(): async def rant_stats_per_weekday():
with Duration("Rant stats per weekday"): with Duration("Rant stats per weekday"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -111,13 +97,12 @@ async def rant_stats_per_weekday():
plt.title("Rants per weekday") plt.title("Rants per weekday")
plt.legend() plt.legend()
plt.savefig(f"export/rants_per_weekday_{get_date_range_str()}.png") plt.savefig(f"export/rants_per_weekday_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def comment_stats_per_weekday(): async def comment_stats_per_weekday():
with Duration("Comment stats per weekday"): with Duration("Comment stats per weekday"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -133,13 +118,12 @@ async def comment_stats_per_weekday():
plt.title("Comments per weekday") plt.title("Comments per weekday")
plt.legend() plt.legend()
plt.savefig(f"export/comments_per_weekday_{get_date_range_str()}.png") plt.savefig(f"export/comments_per_weekday_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def rant_stats_per_hour(): async def rant_stats_per_hour():
with Duration("Rant stats per hour"): with Duration("Rant stats per hour"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -155,13 +139,12 @@ async def rant_stats_per_hour():
plt.title("Rants per hour") plt.title("Rants per hour")
plt.legend() plt.legend()
plt.savefig(f"export/rants_per_hour_{get_date_range_str()}.png") plt.savefig(f"export/rants_per_hour_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def comment_stats_per_hour(): async def comment_stats_per_hour():
with Duration("Comment stats per hour"): with Duration("Comment stats per hour"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -177,13 +160,12 @@ async def comment_stats_per_hour():
plt.title("Comments per hour") plt.title("Comments per hour")
plt.legend() plt.legend()
plt.savefig(f"export/comments_per_hour_{get_date_range_str()}.png") plt.savefig(f"export/comments_per_hour_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def score_most_ignored_last_7_days(): async def score_most_ignored_last_7_days():
with Duration("Score most ignored last 7 days"): with Duration("Score most ignored last 7 days"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -195,7 +177,6 @@ async def score_most_ignored_last_7_days():
y.append(record["userscore"]) y.append(record["userscore"])
plt.bar(x, y, label="Upvotes") plt.bar(x, y, label="Upvotes")
# plt.plot(x, y, label=get_date_range_str(), color='pink')
plt.xticks(rotation=45) plt.xticks(rotation=45)
plt.xlabel("Ranter") plt.xlabel("Ranter")
plt.ylabel("Times ignored") plt.ylabel("Times ignored")
@ -204,13 +185,12 @@ async def score_most_ignored_last_7_days():
plt.savefig( plt.savefig(
f"export/score_ignored_most_last_7_days_{get_date_range_str()}.png" f"export/score_ignored_most_last_7_days_{get_date_range_str()}.png"
) )
plt_reset()
# plt.show()
async def score_last_7_days(): async def score_last_7_days():
with Duration("Upvotes (score) last 7 days"): with Duration("Upvotes (score) last 7 days"):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
x = [] x = []
y = [] y = []
@ -220,20 +200,18 @@ async def score_last_7_days():
y.append(record["userscore"]) y.append(record["userscore"])
plt.bar(x, y, label="Upvotes") plt.bar(x, y, label="Upvotes")
# plt.plot(x, y, label=get_date_range_str(), color='pink')
plt.xticks(rotation=45) plt.xticks(rotation=45)
plt.xlabel("Ranter") plt.xlabel("Ranter")
plt.ylabel("Upvote count") plt.ylabel("Upvote count")
plt.title("Most ignored based on score regarding last 7 days") plt.title("Most ignored based on score regarding last 7 days")
plt.legend() plt.legend()
plt.savefig(f"export/score_last_7_days_{get_date_range_str()}.png") plt.savefig(f"export/score_last_7_days_{get_date_range_str()}.png")
plt_reset()
# plt.show()
async def user_score_per_day(username): async def user_score_per_day(username):
with Duration("User {} score per day".format(username)): with Duration("User {} score per day".format(username)):
plt = figure_inc() plt = new_plot()
async with Db() as db: async with Db() as db:
plt.xticks(rotation=45) plt.xticks(rotation=45)
@ -280,12 +258,10 @@ async def user_score_per_day(username):
plt.legend() plt.legend()
plt.savefig(f"export/score_user_{username}_{get_date_range_str()}.png") plt.savefig(f"export/score_user_{username}_{get_date_range_str()}.png")
plt_reset()
def rant_stats_all(): def stats():
print("Did you sync all rants before running this?")
sleep(3)
with Duration("Complete process"): with Duration("Complete process"):
asyncio.run(rant_stats_per_day()) asyncio.run(rant_stats_per_day())
asyncio.run(rant_stats_per_weekday()) asyncio.run(rant_stats_per_weekday())