"""Module responsibel for plotting statistics."""
import io
import pandas
import matplotlib
import matplotlib.dates as mdates
from sqlalchemy import func, Date, cast, Integer
from stickerfinder.helper.telegram import call_tg_func
from stickerfinder.models import (
InlineQuery,
InlineQueryRequest,
User,
)
matplotlib.use("Agg")
import matplotlib.pyplot as plt # noqa
def send_plots(session, chat):
"""Generate and send plots to the user."""
image = get_inline_queries_statistics(session)
chat.send_document(image, caption="Inline queries")
image = get_inline_query_performance_statistics(session)
chat.send_document(image, caption="Inline query performance statistics")
image = get_user_activity(session)
chat.send_document(image, caption="User statistics")
image.close()
def image_from_figure(fig):
"""Create a pillow image from a figure."""
io_buffer = io.BytesIO()
plt.savefig(io_buffer, format="png")
io_buffer.seek(0)
# from PIL import Image
# image = Image.open(io_buffer)
# image.show()
return io_buffer
def get_inline_queries_statistics(session):
"""Create a plot showing the inline usage statistics."""
# Get all queries over time
all_queries = (
session.query(cast(InlineQuery.created_at, Date), func.count(InlineQuery.id))
.group_by(cast(InlineQuery.created_at, Date))
.all()
)
all_queries = [("all", q[0], q[1]) for q in all_queries]
# Get all successful queries over time
successful_queries = (
session.query(cast(InlineQuery.created_at, Date), func.count(InlineQuery.id))
.filter(InlineQuery.sticker_file_id.isnot(None))
.group_by(cast(InlineQuery.created_at, Date))
.all()
)
successful_queries = [("successful", q[0], q[1]) for q in successful_queries]
# Get all unsuccessful queries over time
unsuccessful_queries = (
session.query(cast(InlineQuery.created_at, Date), func.count(InlineQuery.id))
.filter(InlineQuery.sticker_file_id.is_(None))
.group_by(cast(InlineQuery.created_at, Date))
.all()
)
unsuccessful_queries = [("unsuccessful", q[0], q[1]) for q in unsuccessful_queries]
# Combine the results in a single dataframe and name the columns
inline_queries = all_queries + successful_queries + unsuccessful_queries
dataframe = pandas.DataFrame(inline_queries, columns=["type", "date", "queries"])
# Plot each result set
fig, ax = plt.subplots(figsize=(30, 15), dpi=120)
for key, group in dataframe.groupby(["type"]):
ax = group.plot(ax=ax, kind="line", x="date", y="queries", label=key)
image = image_from_figure(fig)
image.name = "inline_usage.png"
return image
def get_inline_query_performance_statistics(session):
"""Plot statistics regarding performance of inline query requests."""
creation_date = func.cast(InlineQueryRequest.created_at, Date).label(
"creation_date"
)
# Group the started users by date
strict_search_subquery = (
session.query(
creation_date, func.avg(InlineQueryRequest.duration).label("count")
)
.group_by(creation_date)
.order_by(creation_date)
.all()
)
strict_queries = [("strict", q[0], q[1]) for q in strict_search_subquery]
# Combine the results in a single dataframe and name the columns
request_statistics = strict_queries
dataframe = pandas.DataFrame(
request_statistics, columns=["type", "date", "duration"]
)
months = mdates.MonthLocator() # every month
months_fmt = mdates.DateFormatter("%Y-%m")
# Plot each result set
fig, ax = plt.subplots(figsize=(30, 15), dpi=120)
for key, group in dataframe.groupby(["type"]):
ax = group.plot(ax=ax, kind="bar", x="date", y="duration", label=key)
ax.xaxis.set_major_locator(months)
ax.xaxis.set_major_formatter(months_fmt)
image = image_from_figure(fig)
image.name = "request_duration_statistics.png"
return image
def get_user_activity(session):
"""Create a plot showing the user statistics."""
# Create a subquery to ensure that the user fired a inline query
# Group the new users by date
creation_date = cast(User.created_at, Date).label("creation_date")
all_users_subquery = (
session.query(creation_date, func.count(User.id).label("count"))
.filter(User.inline_queries.any())
.group_by(creation_date)
.subquery()
)
# Create a running window which sums all users up to this point for the current millennium ;P
all_users = (
session.query(
all_users_subquery.c.creation_date,
cast(
func.sum(all_users_subquery.c.count).over(
partition_by=func.extract(
"millennium", all_users_subquery.c.creation_date
),
order_by=all_users_subquery.c.creation_date.asc(),
),
Integer,
).label("running_total"),
)
.order_by(all_users_subquery.c.creation_date)
.all()
)
all_users = [("all", q[0], q[1]) for q in all_users]
# Combine the results in a single dataframe and name the columns
user_statistics = all_users
dataframe = pandas.DataFrame(user_statistics, columns=["type", "date", "users"])
# Plot each result set
fig, ax = plt.subplots(figsize=(30, 15), dpi=120)
for key, group in dataframe.groupby(["type"]):
ax = group.plot(ax=ax, kind="line", x="date", y="users", label=key)
image = image_from_figure(fig)
image.name = "user_statistics.png"
return image
