lzy/sci-gui-agent-benchmark
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge branch 'zdy'

Browse Source
This commit is contained in:
David Chang
2023-12-30 20:53:54 +08:00
parent 2d493759e3 aaca06ba40
commit 19b99a13e2
8 changed files with 210 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
desktop_env/evaluators/metrics/__init__.py
View File

@@ -1,3 +1,3 @@
from .table import compare_table, compare_with_sparklines, compare_with_charts
from .table import check_sheet_list, check_xlsx_freeze
from .docs import find_default_font, contains_page_break, compare_docx_files
from .table import compare_table
from .table import check_sheet_list, check_xlsx_freeze, check_zoom
from .docs import find_default_font, contains_page_break, compare_docx_files

132
desktop_env/evaluators/metrics/table.py
View File

@@ -1,56 +1,72 @@
import pandas as pd
import openpyxl
from openpyxl import Workbook
from openpyxl.worksheet.worksheet import Worksheet
from .utils import load_charts, load_sparklines
import operator
from typing import Dict, List
from typing import Any
from typing import Any, Union
from numbers import Number
def compare_table(actual, expected):
df1 = pd.read_excel(expected)
df2 = pd.read_excel(actual)
# Compare the DataFrames
return 1 if df1.equals(df2) else 0
def compare_with_sparklines(actual: str, expected: str) -> float:
df1 = pd.read_excel(actual)
df2 = pd.read_excel(expected)
normal_content_metric: bool = df1.equals(df2)
print("Normal Contents Metric: {:}".format(normal_content_metric))
sp1 = load_sparklines(actual)
sp2 = load_sparklines(expected)
sparkline_metric: bool = sp1 == sp2
print("Sparkline Metric: {:}".format(sparkline_metric))
return float(normal_content_metric and sparkline_metric)
def compare_with_charts(actual: str, expected: str, **options) -> float:
def compare_table(actual: str, expected: str, **options) -> float:
"""
Args:
actual (str): path to result xlsx
expected (str): path to gold xlsx
options (Dict[str, List[str]]): dict like {"chart_props": list of str}
giving the concerned chart properties
options (Dict[str, List[str]]): dict like
{
"features": list of str for other features, supports:
* sparkline
* chart
* number_format
"chart_props": list of str, giving the converned chart properties
}
Return:
float: the score
"""
df1 = pd.read_excel(actual)
df2 = pd.read_excel(expected)
normal_content_metric: bool = df1.equals(df2)
print("Normal Contents Metric: {:}".format(normal_content_metric))
df1 = pd.read_excel(expected)
df2 = pd.read_excel(actual)
metric: bool = df1.equals(df2)
print("Normal Contents Metric: {:}".format(metric))
charts1 = load_charts(actual, **options)
charts2 = load_charts(expected, **options)
chart_metric: bool = charts1 == charts2
print("Chart Metric: {:}".format(chart_metric))
features: List[str] = options.get("features", [])
for ftr in features:
workbook1: Workbook = openpyxl.load_workbook(actual)
workbook2: Workbook = openpyxl.load_workbook(expected)
return float(normal_content_metric and chart_metric)
if ftr=="sparkline":
sp1 = load_sparklines(actual)
sp2 = load_sparklines(expected)
new_metric: bool = sp1 == sp2
print("Sparkline Metric: {:}".format(new_metric))
elif ftr=="chart":
charts1 = load_charts(workbook1, **options)
charts2 = load_charts(workbook2, **options)
new_metric: bool = charts1 == charts2
print("Chart Metric: {:}".format(new_metric))
elif ftr=="number_format":
number_formats1: List[str] = [ c.number_format.lower()\
for col in workbook1.active.iter_cols()\
for c in col\
if c.data_type=="n"
]
number_formats2: List[str] = [ c.number_format.lower()\
for col in workbook2.active.iter_cols()\
for c in col\
if c.data_type=="n"
]
new_metric: bool = number_formats1==number_formats2
print("Number Format Metric: {:}".format(new_metric))
else:
raise NotImplementedError("Unsupported xlsx feature: {:}".format(ftr))
metric = metric and new_metric
return float(metric)
def check_sheet_list(result: str, rules: List[Dict[str, Any]]) -> float:
# workbook: Workbook = openpyxl.load_workbook(filename=result)
@@ -90,11 +106,17 @@ def check_sheet_list(result: str, rules: List[Dict[str, Any]]) -> float:
return float(passes)
def check_xlsx_freeze(result: str, rules: Dict[str, str]) -> float:
worksheet: Worksheet = openpyxl.load_workbook(filename=result).active
return float(worksheet.freeze_panes == rules["position"])
def check_zoom(result: str, rules: Dict[str, Union[str, Number]]) -> float:
worksheet = openpyxl.load_workbook(filename=result).active
zoom_scale: Number = worksheet.sheet_view.zoomScale or 100.
return float( getattr(operator, rules["relation"])( zoom_scale
, rules["ref_value"]
)
)
if __name__ == '__main__':
# path1 = ""
@@ -132,6 +154,38 @@ if __name__ == '__main__':
# ]
# print(check_sheet_list(path1, rule))
path1 = "../../../../../任务数据/LibreOffice Calc/Create_column_charts_using_statistics_gold.xlsx"
path2 = "../../../../../任务数据/LibreOffice Calc/Create_column_charts_using_statistics_gold2.xlsx"
print(compare_with_charts(path1, path2, chart_props=["type", "direction"]))
#path1 = "../../任务数据/LibreOffice Calc/Create_column_charts_using_statistics_gold.xlsx"
#path2 = "../../任务数据/LibreOffice Calc/Create_column_charts_using_statistics_gold2.xlsx"
#print(compare_table(path1, path2, features=["chart"], chart_props=["type", "direction"]))
#path1 = "../../任务数据/LibreOffice Calc/Represent_in_millions_billions_gold.xlsx"
#path2 = "../../任务数据/LibreOffice Calc/Represent_in_millions_billions_gold3.xlsx"
#workbook1: Workbook = openpyxl.load_workbook(filename=path1)
#worksheet1: Worksheet = workbook1.active
#
#import itertools
#for col, r in itertools.product( ['A', 'B', 'C']
#, range(1, 9)
#):
#position: str = "{:}{:d}".format(col, r)
#print(worksheet1[position])
#print(worksheet1[position].value)
#print(worksheet1[position].number_format)
#print(compare_table(path1, path2, features=["number_format"]))
path1 = "../../任务数据/LibreOffice Calc/Zoom_Out_Oversized_Cells_gold.xlsx"
path2 = "../../任务数据/LibreOffice Calc/Zoom_Out_Oversized_Cells.xlsx"
#workbook1: Workbook = openpyxl.load_workbook(filename=path1)
#worksheet1: Worksheet = workbook1.active
#print(worksheet1.sheet_view.zoomScale)
#print(type(worksheet1.sheet_view.zoomScale))
#
#import os
#import os.path
#for wb in filter( lambda f: f.endswith(".xlsx")
#, os.listdir("../../任务数据/LibreOffice Calc/")
#):
#path = os.path.join("../../任务数据/LibreOffice Calc/", wb)
#print(wb, openpyxl.load_workbook(filename=path).active.sheet_view.zoomScale)
print(check_zoom(path1, {"relation": "lt", "ref_value": 100}))
print(check_zoom(path2, {"relation": "lt", "ref_value": 100}))

8
desktop_env/evaluators/metrics/utils.py
View File

@@ -56,10 +56,10 @@ def load_sparklines(xlsx_file: str) -> Dict[str, str]:
# type: "scatterChart" | "lineChart" | "barChart"
# direction: "bar" (hori) | "col" (vert)
# xtitle, ytitle, ztitle: str
def load_charts(xlsx_file: str, **options) -> Dict[str, Any]:
def load_charts(xlsx_file: Workbook, **options) -> Dict[str, Any]:
"""
Args:
xlsx_file (str): path to xlsx
xlsx_file (Workbook): concerned excel book
options (Dict[str, List[str]]): dict like {"chart_props": list of str}
giving the concerned chart properties
@@ -67,8 +67,8 @@ def load_charts(xlsx_file: str, **options) -> Dict[str, Any]:
Dict[str, Any]: information of charts
"""
workbook: Workbook = openpyxl.load_workbook(filename=xlsx_file)
worksheet: Worksheet = workbook.active
#workbook: Workbook = openpyxl.load_workbook(filename=xlsx_file)
worksheet: Worksheet = xlsx_file.active
charts: List[ChartBase] = worksheet._charts
chart_set: Dict[str, Any] = {}

44
evaluation_examples/examples/libreoffice_calc/1334ca3e-f9e3-4db8-9ca7-b4c653be7d17.json Normal file
View File

@@ -0,0 +1,44 @@
{
"id": "1334ca3e-f9e3-4db8-9ca7-b4c653be7d17",
"snapshot": "libreoffice_calc",
"instruction": "The cells are so big that I can not click on the cell I want, zoom out a little bit.",
"source": "https://techcommunity.microsoft.com/t5/excel/excel-workbook-top-way-too-big-can-t-see-rows-and-columns/m-p/4014694",
"config": [
{
"type": "download",
"parameters": {
"file": [
{
"url": "https://drive.usercontent.google.com/download?id=1Wkepf_vic9o7CZFiosZ4jZT_Hy2WbRPZ&export=download&authuser=0&confirm=t&uuid=bc2ce901-a6bb-433f-bcce-cbe42d813f18&at=APZUnTVQcGTcXjwqenmtSH6IMFkM:1703858853235",
"path": "Desktop/Zoom_Out_Oversized_Cells.xlsx"
}
]
}
},
{
"type": "open",
"parameters": {
"path": "Desktop/Zoom_Out_Oversized_Cells.xlsx"
}
}
],
"trajectory": "trajectories/1334ca3e-f9e3-4db8-9ca7-b4c653be7d17",
"related_apps": [
"libreoffice_calc"
],
"evaluator": {
"func": "check_zoom",
"result": {
"type": "vm_file",
"path": "Desktop/Zoom_Out_Oversized_Cells.xlsx",
"dest": "Zoom_Out_Oversized_Cells.xlsx"
},
"expected": {
"type": "rule",
"rules": {
"relation": "lt",
"ref_value": 260
}
}
}
}

47
evaluation_examples/examples/libreoffice_calc/21df9241-f8d7-4509-b7f1-37e501a823f7.json Normal file
View File

@@ -0,0 +1,47 @@
{
"id": "21df9241-f8d7-4509-b7f1-37e501a823f7",
"snapshot": "libreoffice_calc",
"instruction": "Change the representation of colum \"Parameter\" and show in Millions (M) and Billions (B). Keep one decimal and place a white space between the digits and the unit.",
"source": "https://www.youtube.com/watch?v=p5C4V_AO1UU",
"config": [
{
"type": "download",
"parameters": {
"file": [
{
"url": "https://drive.usercontent.google.com/download?id=16PowrQA4E71xUoJmpXPHy0dr9HBcTRmo&export=download&authuser=0&confirm=t&uuid=9a6265f7-585c-4cf8-b321-3b859aec1e68&at=APZUnTWzzOw85wws0ojXNPsIwnjE:1703858126178",
"path": "Desktop/Represent_in_millions_billions.xlsx"
}
]
}
},
{
"type": "open",
"parameters": {
"path": "Desktop/Represent_in_millions_billions.xlsx"
}
}
],
"trajectory": "trajectories/21df9241-f8d7-4509-b7f1-37e501a823f7",
"related_apps": [
"libreoffice_calc"
],
"evaluator": {
"func": "compare_table",
"result": {
"type": "vm_file",
"path": "Desktop/Represent_in_millions_billions.xlsx",
"dest": "Represent_in_millions_billions.xlsx"
},
"expected": {
"type": "cloud_file",
"path": "https://drive.usercontent.google.com/download?id=1Jy6lZexhU5t0eW1GwXJ_csnwIe0Xiy9-&export=download&authuser=0&confirm=t&uuid=601701e7-9eb8-4ce8-83d5-35916094a15d&at=APZUnTW4WE-plIC5MmWTuFu24qLL:1703857882995",
"dest": "Represent_in_millions_billions_gold.xlsx"
},
"options": {
"features": [
"number_format"
]
}
}
}

10
evaluation_examples/examples/libreoffice_calc/2bd59342-0664-4ccb-ba87-79379096cc08.json
View File

@@ -27,7 +27,7 @@
"libreoffice calc"
],
"evaluator": {
"func": "compare_with_sparklines",
"func": "compare_table",
"expected": {
"type": "cloud_file",
"path": "https://drive.usercontent.google.com/download?id=1KQJJLVPGtTL_7ArEWvwwbFbJSiA3cgSE&export=download&authuser=0&confirm=t&uuid=6b11c721-caad-439a-b369-4c13c7a485df&at=APZUnTV5-1isKrDKSHV9NeJ6TDeS:1703509054094",
@@ -37,6 +37,10 @@
"type": "vm_file",
"path": "Desktop/OrderId_Month_Chart.xlsx",
"dest": "OrderId_Month_Chart.xlsx"
}
},
"options": {
"features": [
"sparkline"
]
}
}
}

5
evaluation_examples/examples/libreoffice_calc/347ef137-7eeb-4c80-a3bb-0951f26a8aff.json
View File

@@ -27,7 +27,7 @@
"libreoffice_calc"
],
"evaluator": {
"func": "compare_with_charts",
"func": "compare_table",
"result": {
"type": "vm_file",
"path": "Desktop/Create_column_charts_using_statistics.xlsx",
@@ -39,6 +39,9 @@
"dest": "Create_column_charts_using_statistics_gold.xlsx"
},
"options": {
"features": [
"chart"
],
"chart_props": [
"type",
"direction"

16
evaluation_examples/examples/libreoffice_calc/f9584479-3d0d-4c79-affa-9ad7afdd8850.json
View File

@@ -21,13 +21,13 @@
"path": "Desktop/Quarterly_Product_Sales_by_Zone.xlsx"
}
}
],
"trajectory": "trajectories/f9584479-3d0d-4c79-affa-9ad7afdd8850",
"related_apps": [
"libreoffice calc"
],
"evaluator": {
"func": "compare_table",
],
"trajectory": "trajectories/f9584479-3d0d-4c79-affa-9ad7afdd8850",
"related_apps": [
"libreoffice calc"
],
"evaluator": {
"func": "compare_table",
"expected": {
"type": "cloud_file",
"path": "https://drive.usercontent.google.com/download?id=17f1wZuJPvUEc5at_Fy3c18VFdOk0x7xz&export=download&authuser=0&confirm=t&uuid=6d2edffd-0ce0-426e-9820-8af25b4667f3&at=APZUnTVh7JS85dwZBaV2hytWQgDK:1702361510956",
@@ -38,5 +38,5 @@
"path": "Desktop/Quarterly_Product_Sales_by_Zone.xlsx",
"dest": "Quarterly_Product_Sales_by_Zone.xlsx"
}
}
}
}