cyberosa
commited on
Commit
·
73f4a97
1
Parent(s):
4261f54
updating notebook with errors handling. Cleaning
Browse files- nbs/weekly_analysis.ipynb +0 -0
- tabs/error.py +43 -37
nbs/weekly_analysis.ipynb
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tabs/error.py
CHANGED
|
@@ -3,34 +3,38 @@ import gradio as gr
|
|
| 3 |
from typing import List
|
| 4 |
|
| 5 |
|
| 6 |
-
HEIGHT=600
|
| 7 |
-
WIDTH=1000
|
| 8 |
|
| 9 |
-
# def set_error(row: pd.Series) -> bool:
|
| 10 |
-
# """Sets the error for the given row."""
|
| 11 |
-
# if row.error not in [True, False]:
|
| 12 |
-
# if not row.prompt_response:
|
| 13 |
-
# return True
|
| 14 |
-
# return False
|
| 15 |
-
# return row.error
|
| 16 |
|
| 17 |
def get_error_data(tools_df: pd.DataFrame, inc_tools: List[str]) -> pd.DataFrame:
|
| 18 |
"""Gets the error data for the given tools and calculates the error percentage."""
|
| 19 |
-
tools_inc = tools_df[tools_df[
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
return error
|
| 25 |
|
|
|
|
| 26 |
def get_error_data_overall(error_df: pd.DataFrame) -> pd.DataFrame:
|
| 27 |
"""Gets the error data for the given tools and calculates the error percentage."""
|
| 28 |
-
error_total =
|
| 29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
error_total.columns = error_total.columns.astype(str)
|
| 31 |
-
error_total[
|
| 32 |
return error_total
|
| 33 |
|
|
|
|
| 34 |
def plot_error_data(error_all_df: pd.DataFrame) -> gr.BarPlot:
|
| 35 |
"""Plots the error data for the given tools and calculates the error percentage."""
|
| 36 |
return gr.BarPlot(
|
|
@@ -45,35 +49,37 @@ def plot_error_data(error_all_df: pd.DataFrame) -> gr.BarPlot:
|
|
| 45 |
show_actions_button=True,
|
| 46 |
tooltip=["request_month_year_week", "error_perc"],
|
| 47 |
height=HEIGHT,
|
| 48 |
-
width=WIDTH
|
| 49 |
)
|
| 50 |
|
|
|
|
| 51 |
def plot_tool_error_data(error_df: pd.DataFrame, tool: str) -> gr.BarPlot:
|
| 52 |
"""Plots the error data for the given tool."""
|
| 53 |
-
error_tool = error_df[error_df[
|
| 54 |
error_tool.columns = error_tool.columns.astype(str)
|
| 55 |
-
error_tool[
|
| 56 |
-
|
| 57 |
return gr.BarPlot(
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
| 70 |
)
|
| 71 |
|
|
|
|
| 72 |
def plot_week_error_data(error_df: pd.DataFrame, week: str) -> gr.BarPlot:
|
| 73 |
"""Plots the error data for the given week."""
|
| 74 |
-
error_week = error_df[error_df[
|
| 75 |
error_week.columns = error_week.columns.astype(str)
|
| 76 |
-
error_week[
|
| 77 |
return gr.BarPlot(
|
| 78 |
value=error_week,
|
| 79 |
x="tool",
|
|
@@ -86,5 +92,5 @@ def plot_week_error_data(error_df: pd.DataFrame, week: str) -> gr.BarPlot:
|
|
| 86 |
show_actions_button=True,
|
| 87 |
tooltip=["tool", "error_perc"],
|
| 88 |
height=HEIGHT,
|
| 89 |
-
width=WIDTH
|
| 90 |
-
)
|
|
|
|
| 3 |
from typing import List
|
| 4 |
|
| 5 |
|
| 6 |
+
HEIGHT = 600
|
| 7 |
+
WIDTH = 1000
|
| 8 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
|
| 10 |
def get_error_data(tools_df: pd.DataFrame, inc_tools: List[str]) -> pd.DataFrame:
|
| 11 |
"""Gets the error data for the given tools and calculates the error percentage."""
|
| 12 |
+
tools_inc = tools_df[tools_df["tool"].isin(inc_tools)]
|
| 13 |
+
error = (
|
| 14 |
+
tools_inc.groupby(["tool", "request_month_year_week", "error"])
|
| 15 |
+
.size()
|
| 16 |
+
.unstack()
|
| 17 |
+
.fillna(0)
|
| 18 |
+
.reset_index()
|
| 19 |
+
)
|
| 20 |
+
error["error_perc"] = (error[1] / (error[0] + error[1])) * 100
|
| 21 |
+
error["total_requests"] = error[0] + error[1]
|
| 22 |
return error
|
| 23 |
|
| 24 |
+
|
| 25 |
def get_error_data_overall(error_df: pd.DataFrame) -> pd.DataFrame:
|
| 26 |
"""Gets the error data for the given tools and calculates the error percentage."""
|
| 27 |
+
error_total = (
|
| 28 |
+
error_df.groupby("request_month_year_week")
|
| 29 |
+
.agg({"total_requests": "sum", 1: "sum", 0: "sum"})
|
| 30 |
+
.reset_index()
|
| 31 |
+
)
|
| 32 |
+
error_total["error_perc"] = (error_total[1] / error_total["total_requests"]) * 100
|
| 33 |
error_total.columns = error_total.columns.astype(str)
|
| 34 |
+
error_total["error_perc"] = error_total["error_perc"].apply(lambda x: round(x, 4))
|
| 35 |
return error_total
|
| 36 |
|
| 37 |
+
|
| 38 |
def plot_error_data(error_all_df: pd.DataFrame) -> gr.BarPlot:
|
| 39 |
"""Plots the error data for the given tools and calculates the error percentage."""
|
| 40 |
return gr.BarPlot(
|
|
|
|
| 49 |
show_actions_button=True,
|
| 50 |
tooltip=["request_month_year_week", "error_perc"],
|
| 51 |
height=HEIGHT,
|
| 52 |
+
width=WIDTH,
|
| 53 |
)
|
| 54 |
|
| 55 |
+
|
| 56 |
def plot_tool_error_data(error_df: pd.DataFrame, tool: str) -> gr.BarPlot:
|
| 57 |
"""Plots the error data for the given tool."""
|
| 58 |
+
error_tool = error_df[error_df["tool"] == tool]
|
| 59 |
error_tool.columns = error_tool.columns.astype(str)
|
| 60 |
+
error_tool["error_perc"] = error_tool["error_perc"].apply(lambda x: round(x, 4))
|
| 61 |
+
|
| 62 |
return gr.BarPlot(
|
| 63 |
+
title="Error Percentage",
|
| 64 |
+
x_title="Week",
|
| 65 |
+
y_title="Error Percentage",
|
| 66 |
+
show_label=True,
|
| 67 |
+
interactive=True,
|
| 68 |
+
show_actions_button=True,
|
| 69 |
+
tooltip=["request_month_year_week", "error_perc"],
|
| 70 |
+
value=error_tool,
|
| 71 |
+
x="request_month_year_week",
|
| 72 |
+
y="error_perc",
|
| 73 |
+
height=HEIGHT,
|
| 74 |
+
width=WIDTH,
|
| 75 |
)
|
| 76 |
|
| 77 |
+
|
| 78 |
def plot_week_error_data(error_df: pd.DataFrame, week: str) -> gr.BarPlot:
|
| 79 |
"""Plots the error data for the given week."""
|
| 80 |
+
error_week = error_df[error_df["request_month_year_week"] == week]
|
| 81 |
error_week.columns = error_week.columns.astype(str)
|
| 82 |
+
error_week["error_perc"] = error_week["error_perc"].apply(lambda x: round(x, 4))
|
| 83 |
return gr.BarPlot(
|
| 84 |
value=error_week,
|
| 85 |
x="tool",
|
|
|
|
| 92 |
show_actions_button=True,
|
| 93 |
tooltip=["tool", "error_perc"],
|
| 94 |
height=HEIGHT,
|
| 95 |
+
width=WIDTH,
|
| 96 |
+
)
|