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

Merge remote-tracking branch 'upstream/main' into fix_chrome

Browse Source
This commit is contained in:
yuanmengqi
2025-07-11 12:16:23 +00:00
parent fe3bb2fd92 6f0382c0c2
commit 37c56533f0
31 changed files with 3853 additions and 206 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
desktop_env/evaluators/metrics/chrome.py
View File

@@ -115,6 +115,11 @@ def is_expected_tabs(open_tabs: List[Dict[str, str]], rule: Dict[str, Any]) -> f
if match_type == "url":
expected_urls = rule['urls']
actual_urls = [tab['url'] for tab in open_tabs]
if not are_lists_equal(expected_urls, actual_urls, compare_urls):
logger.error("list not match")
logger.error(expected_urls)
logger.error(actual_urls)
return 0
return 1 if are_lists_equal(expected_urls, actual_urls, compare_urls) else 0
else:
logger.error(f"Unknown type: {match_type}")
@@ -343,7 +348,7 @@ def compare_archive(pred_path: str, gold_path: str, **kwargs) -> float:
return score / len(pred_files)
def compare_htmls(html_path1: str, html_path2: str) -> float:
def compare_htmls(html_path1: str, html_path2: str, **options) -> float:
"""
Compare two HTML files.
"""
@@ -351,20 +356,33 @@ def compare_htmls(html_path1: str, html_path2: str) -> float:
soup1 = BeautifulSoup(inf, 'lxml')
with open(html_path2, 'r', encoding='utf-8') as inf:
soup2 = BeautifulSoup(inf, 'lxml')
ignore_sdnum = options.get("ignore_sdnum", None)
def compare_elements(elem1, elem2):
if not (isinstance(elem1, Tag) and isinstance(elem2, Tag)):
if elem1 != elem2:
logger.info("not the same")
return elem1 == elem2
if elem1.name != elem2.name:
logger.info("html name not match")
return False
if elem1.text.strip() != elem2.text.strip():
logger.info("html text not match")
return False
if elem1.attrs != elem2.attrs:
if ignore_sdnum:
attrs1 = {k: v for k, v in elem1.attrs.items() if k != 'sdnum'}
attrs2 = {k: v for k, v in elem2.attrs.items() if k != 'sdnum'}
return attrs1 == attrs2
logger.info("html attrs not match")
logger.info(f"{elem1.attrs}")
logger.info(f"{elem2.attrs}")
return False
return True
for elem1, elem2 in zip(soup1.recursiveChildGenerator(), soup2.recursiveChildGenerator()):
if not compare_elements(elem1, elem2):
logger.info("html not match")
return .0
return 1.

25
desktop_env/evaluators/metrics/docs.py
View File

@@ -297,8 +297,24 @@ def compare_image_text(image_path, rule):
reader = easyocr.Reader(['en'])
result = reader.readtext(image_path)
extracted_text = ' '.join([entry[1] for entry in result])
# Log OCR results
logger.info(f"OCR extracted texts: {[entry[1] for entry in result]}")
logger.info(f"Combined extracted text: {extracted_text}")
if rule['type'] == 'text':
return 1 if rule['text'] in extracted_text else 0
target_text = rule['text']
match_found = target_text in extracted_text
# Log matching results
logger.info(f"Target text: '{target_text}'")
logger.info(f"Match found: {match_found}")
if match_found:
logger.info("✅ Text matching successful!")
else:
logger.info("❌ Text matching failed!")
return 1 if match_found else 0
else:
raise ValueError("Unsupported rule type")
@@ -986,3 +1002,10 @@ def compare_unique_train_records(processed_file, expected_files, **kwargs):
return 0
return 1
if __name__ == "__main__":
image_path = "/home/ubuntu/OSWorld/cache/02ce9a50-7af2-47ed-8596-af0c230501f8/ls.png"
print(compare_image_text(image_path, {
"type": "text",
"text": "ls"
}))

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

@@ -213,7 +213,6 @@ _accessibility_ns_map = {
}
def check_accessibility_tree(result: str, rules: List[Dict[str, Any]], osname: str = "ubuntu") -> float:
"""
Args:
@@ -402,12 +401,17 @@ def check_direct_json_object(result, rules) -> float:
expected_value_list = expected_json.get(key)
logger.info(f"[DEBUG] Checking list key '{key}': expected_list={expected_value_list}, actual='{result.get(key)}'")
for each_expected_value in expected_value_list:
# Handle both list and string cases
if isinstance(result.get(key), list) and each_expected_value in result.get(key):
flag = 1
logger.info(f"[DEBUG] Found expected value '{each_expected_value}' in result list for key '{key}'")
break
elif isinstance(result.get(key), str) and each_expected_value == result.get(key):
flag = 1
logger.info(f"[DEBUG] Found expected value '{each_expected_value}' matches result string for key '{key}'")
break
if flag == 0:
logger.info(f"[DEBUG] No expected values found in result list for key '{key}', returning 0.0")
logger.info(f"[DEBUG] No expected values found in result for key '{key}', returning 0.0")
return 0.
elif isinstance(expected_json.get(key), str):
expected_str = expected_json.get(key)

310
desktop_env/evaluators/metrics/slides.py
View File

@@ -9,6 +9,31 @@ from pptx.enum.shapes import MSO_SHAPE_TYPE
logger = logging.getLogger("desktopenv.metric.slides")
# Add a new logger specifically for debugging PPTX comparisons
debug_logger = logging.getLogger("desktopenv.metric.slides.debug")
def enable_debug_logging():
"""Enable debug logging for PPTX comparison"""
debug_logger.setLevel(logging.DEBUG)
if not debug_logger.handlers:
handler = logging.StreamHandler()
handler.setLevel(logging.DEBUG)
formatter = logging.Formatter('[PPTX_DEBUG] %(message)s')
handler.setFormatter(formatter)
debug_logger.addHandler(handler)
# Add debug logger for detailed comparison output
debug_logger = logging.getLogger("desktopenv.metric.slides.debug")
def enable_debug_logging():
"""Enable detailed debug logging for PPTX comparison"""
debug_logger.setLevel(logging.DEBUG)
if not debug_logger.handlers:
handler = logging.StreamHandler()
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
handler.setFormatter(formatter)
debug_logger.addHandler(handler)
def check_presenter_console_disable(config_file_path):
try:
@@ -135,11 +160,45 @@ def check_slide_numbers_color(pptx_file_path):
# return similarity_index
def get_all_text_shapes(slide):
"""递归获取slide中所有包含文本的shapes，包括GROUP内部的"""
def extract_text_shapes(shape):
results = []
# 检查当前shape是否有文本
if hasattr(shape, "text") and hasattr(shape, "text_frame"):
results.append(shape)
# 如果是GROUP，递归检查内部shapes
if hasattr(shape, 'shapes'):
for sub_shape in shape.shapes:
results.extend(extract_text_shapes(sub_shape))
return results
all_text_shapes = []
for shape in slide.shapes:
all_text_shapes.extend(extract_text_shapes(shape))
return all_text_shapes
def compare_pptx_files(file1_path, file2_path, **options):
# todo: not strictly match since not all information is compared because we cannot get the info through pptx
prs1 = Presentation(file1_path)
prs2 = Presentation(file2_path)
# Enable debug logging if requested
enable_debug = options.get("enable_debug", True)
if enable_debug:
enable_debug_logging()
debug_logger.debug(f"=== COMPARING PPTX FILES ===")
debug_logger.debug(f"File 1: {file1_path}")
debug_logger.debug(f"File 2: {file2_path}")
debug_logger.debug(f"File 1 slides: {len(prs1.slides)}")
debug_logger.debug(f"File 2 slides: {len(prs2.slides)}")
approximately_tolerance = options.get("approximately_tolerance", 0.005)
def is_approximately_equal(val1, val2, tolerance=approximately_tolerance):
"""Compare two values with a tolerance of 0.1% (0.005)"""
@@ -176,12 +235,17 @@ def compare_pptx_files(file1_path, file2_path, **options):
# compare the number of slides
if len(prs1.slides) != len(prs2.slides) and examine_number_of_slides:
if enable_debug:
debug_logger.debug(f"MISMATCH: Number of slides differ - File1: {len(prs1.slides)}, File2: {len(prs2.slides)}")
return 0
slide_idx = 0
# compare the content of each slide
for slide1, slide2 in zip(prs1.slides, prs2.slides):
slide_idx += 1
if enable_debug:
debug_logger.debug(f"--- Comparing Slide {slide_idx} ---")
debug_logger.debug(f"Slide {slide_idx} - Shapes count: File1={len(slide1.shapes)}, File2={len(slide2.shapes)}")
def get_slide_background_color(slide):
# background = slide.background
@@ -212,14 +276,35 @@ def compare_pptx_files(file1_path, file2_path, **options):
return None
if get_slide_notes(slide1).strip() != get_slide_notes(slide2).strip() and examine_note:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx} - Notes differ:")
debug_logger.debug(f" Notes1: '{get_slide_notes(slide1).strip()}'")
debug_logger.debug(f" Notes2: '{get_slide_notes(slide2).strip()}'")
return 0
# Get all text shapes including those inside GROUPs
text_shapes1 = get_all_text_shapes(slide1)
text_shapes2 = get_all_text_shapes(slide2)
if enable_debug:
debug_logger.debug(f"Slide {slide_idx} - Text shapes found: File1={len(text_shapes1)}, File2={len(text_shapes2)}")
# check if the number of slides is the same
if len(slide1.shapes) != len(slide2.shapes):
if enable_debug:
debug_logger.debug(f"MISMATCH: Slide {slide_idx} - Different number of shapes: File1={len(slide1.shapes)}, File2={len(slide2.shapes)}")
return 0
# check if the shapes are the same
shape_idx = 0
for shape1, shape2 in zip(slide1.shapes, slide2.shapes):
shape_idx += 1
if enable_debug:
debug_logger.debug(f" Shape {shape_idx} - Type: {shape1.shape_type} vs {shape2.shape_type}")
if hasattr(shape1, "text") and hasattr(shape2, "text"):
debug_logger.debug(f" Shape {shape_idx} - Text: '{shape1.text.strip()}' vs '{shape2.text.strip()}'")
debug_logger.debug(f" Shape {shape_idx} - Position: ({shape1.left}, {shape1.top}) vs ({shape2.left}, {shape2.top})")
debug_logger.debug(f" Shape {shape_idx} - Size: ({shape1.width}, {shape1.height}) vs ({shape2.width}, {shape2.height})")
if examine_title_bottom_position:
if hasattr(shape1, "text") and hasattr(shape2, "text") and shape1.text == shape2.text:
if shape1.text == "Product Comparison" and (shape1.top <= shape2.top or shape1.top < 3600000):
@@ -249,11 +334,7 @@ def compare_pptx_files(file1_path, file2_path, **options):
if slide_idx == 2 and shape1.shape_type == 13 and shape2.shape_type == 13:
if shape1.top >= shape2.top or shape1.top > 1980000:
return 0
elif (not is_approximately_equal(shape1.left, shape2.left) or
not is_approximately_equal(shape1.top, shape2.top) or
not is_approximately_equal(shape1.width, shape2.width) or
not is_approximately_equal(shape1.height, shape2.height)):
return 0
if examine_shape_for_shift_size:
if (not is_approximately_equal(shape1.left, shape2.left) or
@@ -264,11 +345,23 @@ def compare_pptx_files(file1_path, file2_path, **options):
"text") and shape1.text == shape2.text and shape1.text == "Elaborate on what you want to discuss."):
return 0
# CRITICAL: examine_shape check happens BEFORE examine_modify_height!
# If examine_shape=True (default), any shape dimension mismatch will cause immediate return 0,
# preventing examine_modify_height from ever being executed.
# For height modification tasks, you MUST set examine_shape=False to allow examine_modify_height to work.
if (
not is_approximately_equal(shape1.left, shape2.left) or
not is_approximately_equal(shape1.top, shape2.top) or
not is_approximately_equal(shape1.width, shape2.width) or
not is_approximately_equal(shape1.height, shape2.height)) and examine_shape:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx} - Shape dimensions differ:")
debug_logger.debug(f" Left: {shape1.left} vs {shape2.left} (equal: {is_approximately_equal(shape1.left, shape2.left)})")
debug_logger.debug(f" Top: {shape1.top} vs {shape2.top} (equal: {is_approximately_equal(shape1.top, shape2.top)})")
debug_logger.debug(f" Width: {shape1.width} vs {shape2.width} (equal: {is_approximately_equal(shape1.width, shape2.width)})")
debug_logger.debug(f" Height: {shape1.height} vs {shape2.height} (equal: {is_approximately_equal(shape1.height, shape2.height)})")
if hasattr(shape1, "text") and hasattr(shape2, "text"):
debug_logger.debug(f" Shape text: '{shape1.text.strip()}' vs '{shape2.text.strip()}'")
return 0
if examine_image_size:
@@ -281,6 +374,11 @@ def compare_pptx_files(file1_path, file2_path, **options):
not is_approximately_equal(shape1.height, shape2.height)):
return 0
# examine_modify_height: Special logic for height modification tasks
# - For non-text shapes and FREEFORM shapes (type 5): Only check height differences
# - For other shapes: Check all dimensions (left, top, width, height)
# WARNING: This check only works if examine_shape=False, otherwise examine_shape will
# terminate the comparison before this code is reached!
if examine_modify_height:
if not hasattr(shape1, "text") and not hasattr(shape2,
"text") or shape1.shape_type == 5 and shape2.shape_type == 5:
@@ -295,24 +393,92 @@ def compare_pptx_files(file1_path, file2_path, **options):
if shape1.shape_type == MSO_SHAPE_TYPE.TABLE:
table1 = shape1.table
table2 = shape2.table
if enable_debug:
debug_logger.debug(f" Shape {shape_idx} - Comparing TABLE with {len(table1.rows)} rows and {len(table1.columns)} columns")
for row_idx in range(len(table1.rows)):
for col_idx in range(len(table1.columns)):
cell1 = table1.cell(row_idx, col_idx)
cell2 = table2.cell(row_idx, col_idx)
for para1, para2 in zip(cell1.text_frame.paragraphs, cell2.text_frame.paragraphs):
for run1, run2 in zip(para1.runs, para2.runs):
if run1.font.color.rgb != run2.font.color.rgb:
return 0
for para_idx, (para1, para2) in enumerate(zip(cell1.text_frame.paragraphs, cell2.text_frame.paragraphs)):
for run_idx, (run1, run2) in enumerate(zip(para1.runs, para2.runs)):
# Check font color
if hasattr(run1.font.color, "rgb") and hasattr(run2.font.color, "rgb"):
if run1.font.color.rgb != run2.font.color.rgb and examine_color_rgb:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx} (TABLE) - Cell [{row_idx},{col_idx}], Para {para_idx}, Run {run_idx} - Font color differs:")
debug_logger.debug(f" Color1: {run1.font.color.rgb} vs Color2: {run2.font.color.rgb}")
debug_logger.debug(f" Cell text: '{cell1.text.strip()}' vs '{cell2.text.strip()}'")
debug_logger.debug(f" Run text: '{run1.text}' vs '{run2.text}'")
return 0
# Check font bold
if run1.font.bold != run2.font.bold:
if not ((run1.font.bold is None or run1.font.bold is False) and
(run2.font.bold is None or run2.font.bold is False)):
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx} (TABLE) - Cell [{row_idx},{col_idx}], Para {para_idx}, Run {run_idx} - Font bold differs:")
debug_logger.debug(f" Bold1: {run1.font.bold} vs Bold2: {run2.font.bold}")
debug_logger.debug(f" Run text: '{run1.text}' vs '{run2.text}'")
return 0
# Check font italic
if run1.font.italic != run2.font.italic:
if not ((run1.font.italic is None or run1.font.italic is False) and
(run2.font.italic is None or run2.font.italic is False)):
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx} (TABLE) - Cell [{row_idx},{col_idx}], Para {para_idx}, Run {run_idx} - Font italic differs:")
debug_logger.debug(f" Italic1: {run1.font.italic} vs Italic2: {run2.font.italic}")
debug_logger.debug(f" Run text: '{run1.text}' vs '{run2.text}'")
return 0
# Check font underline
if run1.font.underline != run2.font.underline:
if run1.font.underline is not None and run2.font.underline is not None:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx} (TABLE) - Cell [{row_idx},{col_idx}], Para {para_idx}, Run {run_idx} - Font underline differs:")
debug_logger.debug(f" Underline1: {run1.font.underline} vs Underline2: {run2.font.underline}")
debug_logger.debug(f" Run text: '{run1.text}' vs '{run2.text}'")
return 0
if (run1.font.underline is None and run2.font.underline is True) or (run1.font.underline is True and run2.font.underline is None):
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx} (TABLE) - Cell [{row_idx},{col_idx}], Para {para_idx}, Run {run_idx} - Font underline differs (None vs True):")
debug_logger.debug(f" Underline1: {run1.font.underline} vs Underline2: {run2.font.underline}")
debug_logger.debug(f" Run text: '{run1.text}' vs '{run2.text}'")
return 0
if hasattr(shape1, "text") and hasattr(shape2, "text"):
if shape1.text.strip() != shape2.text.strip() and examine_text:
return 0
# check if the paragraphs are the same
para_idx = 0
for para1, para2 in zip(shape1.text_frame.paragraphs, shape2.text_frame.paragraphs):
if para1.alignment != para2.alignment and examine_alignment:
return 0
para_idx += 1
# Handle alignment comparison - treat None and LEFT (1) as equivalent
if examine_alignment:
from pptx.enum.text import PP_ALIGN
align1 = para1.alignment
align2 = para2.alignment
if enable_debug:
align1_name = "None" if align1 is None else getattr(align1, 'name', str(align1))
align2_name = "None" if align2 is None else getattr(align2, 'name', str(align2))
debug_logger.debug(f" Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Alignment: '{align1_name}' vs '{align2_name}'")
debug_logger.debug(f" Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Text: '{para1.text}' vs '{para2.text}'")
# Convert None to LEFT for comparison since None means default left alignment
if align1 is None:
align1 = PP_ALIGN.LEFT # LEFT alignment
if align2 is None:
align2 = PP_ALIGN.LEFT # LEFT alignment
if align1 != align2:
if enable_debug:
align1_final = getattr(align1, 'name', str(align1))
align2_final = getattr(align2, 'name', str(align2))
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Alignment differs: '{align1_final}' vs '{align2_final}'")
return 0
# check if the runs are the same
if para1.text != para2.text and examine_text:
@@ -325,26 +491,59 @@ def compare_pptx_files(file1_path, file2_path, **options):
# check if the font properties are the same
if run1.font.name != run2.font.name and examine_font_name:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font name differs:")
debug_logger.debug(f" Name1: '{run1.font.name}' vs Name2: '{run2.font.name}'")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if run1.font.size != run2.font.size and examine_font_size:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font size differs:")
debug_logger.debug(f" Size1: {run1.font.size} vs Size2: {run2.font.size}")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if run1.font.bold != run2.font.bold and examine_font_bold:
return 0
# Special handling for None vs False - both mean "not bold"
if not ((run1.font.bold is None or run1.font.bold is False) and
(run2.font.bold is None or run2.font.bold is False)):
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font bold differs:")
debug_logger.debug(f" Bold1: {run1.font.bold} vs Bold2: {run2.font.bold}")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if run1.font.italic != run2.font.italic and examine_font_italic:
if run1.font.italic is not None and run2.font.italic is not None:
# Special handling for None vs False - both mean "not italic"
if not ((run1.font.italic is None or run1.font.italic is False) and
(run2.font.italic is None or run2.font.italic is False)):
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font italic differs:")
debug_logger.debug(f" Italic1: {run1.font.italic} vs Italic2: {run2.font.italic}")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if hasattr(run1.font.color, "rgb") and hasattr(run2.font.color, "rgb"):
if run1.font.color.rgb != run2.font.color.rgb and examine_color_rgb:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font color differs:")
debug_logger.debug(f" Color1: {run1.font.color.rgb} vs Color2: {run2.font.color.rgb}")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if run1.font.underline != run2.font.underline and examine_font_underline:
if run1.font.underline is not None and run2.font.underline is not None:
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font underline differs:")
debug_logger.debug(f" Underline1: {run1.font.underline} vs Underline2: {run2.font.underline}")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if (run1.font.underline is None and run2.font.underline is True) or (run1.font.underline is True and run2.font.underline is None):
if enable_debug:
debug_logger.debug(f" MISMATCH: Slide {slide_idx}, Shape {shape_idx}, Para {para_idx} - Font underline differs (None vs True):")
debug_logger.debug(f" Underline1: {run1.font.underline} vs Underline2: {run2.font.underline}")
debug_logger.debug(f" Text: '{run1.text}' vs '{run2.text}'")
return 0
if run1.font._element.attrib.get('strike', 'noStrike') != run2.font._element.attrib.get(
@@ -382,16 +581,89 @@ def compare_pptx_files(file1_path, file2_path, **options):
return bullets
if examine_bullets:
bullets1 = _extract_bullets(run1.part.blob.decode('utf-8'))
bullets2 = _extract_bullets(run2.part.blob.decode('utf-8'))
def _compare_bullets_with_tolerance(bullets1, bullets2):
"""Compare bullets with tolerance for minor differences"""
if len(bullets1) != len(bullets2):
return False
# Compare only non-empty bullets
if bullets1 != bullets2:
return 0
for (lvl1, char1, text1, color1), (lvl2, char2, text2, color2) in zip(bullets1, bullets2):
# Compare text (most important)
if text1 != text2:
return False
# Compare bullet character
if char1 != char2:
return False
# Compare level with tolerance (None and '0' are equivalent)
normalized_lvl1 = '0' if lvl1 is None else lvl1
normalized_lvl2 = '0' if lvl2 is None else lvl2
if normalized_lvl1 != normalized_lvl2:
return False
# Color comparison is more lenient - we don't fail on color differences
# since they might be due to theme or formatting differences
# if color1 != color2:
# return False
return True
if examine_bullets:
try:
bullets1 = _extract_bullets(run1.part.blob.decode('utf-8'))
bullets2 = _extract_bullets(run2.part.blob.decode('utf-8'))
# Compare bullets with tolerance for minor differences
if not _compare_bullets_with_tolerance(bullets1, bullets2):
return 0
except:
# If bullet extraction fails, skip bullet comparison
pass
# fixme: Actually there are more properties to be compared, we can add them later via parsing the xml data
# Additional check: compare all text shapes including those in GROUPs
if examine_alignment and len(text_shapes1) == len(text_shapes2):
for idx, (tshape1, tshape2) in enumerate(zip(text_shapes1, text_shapes2)):
if enable_debug:
debug_logger.debug(f" Additional text shape check {idx+1}: '{tshape1.text.strip()[:30]}' vs '{tshape2.text.strip()[:30]}'")
# Compare text content
if tshape1.text.strip() != tshape2.text.strip() and examine_text:
if enable_debug:
debug_logger.debug(f" MISMATCH: Text differs - '{tshape1.text.strip()}' vs '{tshape2.text.strip()}'")
return 0
# Compare alignment of each paragraph
for para_idx, (para1, para2) in enumerate(zip(tshape1.text_frame.paragraphs, tshape2.text_frame.paragraphs)):
from pptx.enum.text import PP_ALIGN
align1 = para1.alignment
align2 = para2.alignment
if enable_debug:
align1_name = "None" if align1 is None else getattr(align1, 'name', str(align1))
align2_name = "None" if align2 is None else getattr(align2, 'name', str(align2))
debug_logger.debug(f" Para {para_idx+1}: Alignment '{align1_name}' vs '{align2_name}'")
# Convert None to LEFT for comparison
if align1 is None:
align1 = PP_ALIGN.LEFT
if align2 is None:
align2 = PP_ALIGN.LEFT
if align1 != align2:
if enable_debug:
align1_final = getattr(align1, 'name', str(align1))
align2_final = getattr(align2, 'name', str(align2))
debug_logger.debug(f" MISMATCH: Alignment differs - '{align1_final}' vs '{align2_final}'")
return 0
elif len(text_shapes1) != len(text_shapes2):
if enable_debug:
debug_logger.debug(f"MISMATCH: Different number of text shapes - {len(text_shapes1)} vs {len(text_shapes2)}")
return 0
if enable_debug:
debug_logger.debug(f"=== COMPARISON SUCCESSFUL - Files match ===")
return 1

43
evaluation_examples/examples/libreoffice_impress/04578141-1d42-4146-b9cf-6fab4ce5fd74.json
View File

@@ -1,7 +1,7 @@
{
"id": "04578141-1d42-4146-b9cf-6fab4ce5fd74",
"snapshot": "libreoffice_impress",
"instruction":"Color the first three textboxes on slide 1 yellow, red, and green, respectively, in top-to-bottom order. Use exactly these colors—no variations (e.g., no dark red, light green, etc.).",
"instruction": "Color the first three textboxes on slide 1 yellow, red, and green, respectively, in top-to-bottom order. Use exactly these colors—no variations (e.g., no dark red, light green, etc.).",
"source": "https://arxiv.org/pdf/2311.01767.pdf",
"config": [
{
@@ -64,18 +64,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/04578141-1d42-4146-b9cf-6fab4ce5fd74/45_2_Gold.pptx",
"dest": "45_2_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/45_2.pptx",
"dest": "45_2.pptx"
},
"options": {}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/04578141-1d42-4146-b9cf-6fab4ce5fd74/45_2_Gold.pptx",
"dest": "45_2_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/04578141-1d42-4146-b9cf-6fab4ce5fd74/45_2_Gold_all_fonts.pptx",
"dest": "45_2_Gold_All_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/45_2.pptx",
"dest": "45_2.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/45_2.pptx",
"dest": "45_2.pptx"
}
]
},
"proxy": false
}

35
evaluation_examples/examples/libreoffice_impress/05dd4c1d-c489-4c85-8389-a7836c4f0567.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/05dd4c1d-c489-4c85-8389-a7836c4f0567/38_1_Gold.pptx",
"dest": "38_1_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/38_1.pptx",
"dest": "38_1.pptx"
},
"options": {}
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/05dd4c1d-c489-4c85-8389-a7836c4f0567/38_1_Gold_all_fonts.pptx",
"dest": "38_1_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/38_1.pptx",
"dest": "38_1.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/38_1.pptx",
"dest": "38_1.pptx"
}
]
},
"proxy": false
}

41
evaluation_examples/examples/libreoffice_impress/08aced46-45a2-48d7-993b-ed3fb5b32302.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/08aced46-45a2-48d7-993b-ed3fb5b32302/22_6_Gold.pptx",
"dest": "22_6_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/22_6.pptx",
"dest": "22_6.pptx"
},
"options": {}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/08aced46-45a2-48d7-993b-ed3fb5b32302/22_6_Gold.pptx",
"dest": "22_6_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/08aced46-45a2-48d7-993b-ed3fb5b32302/22_6_Gold2.pptx",
"dest": "22_6_Gold2.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/22_6.pptx",
"dest": "22_6.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/22_6.pptx",
"dest": "22_6.pptx"
}
]
},
"proxy": false
}

3
evaluation_examples/examples/libreoffice_impress/3161d64e-3120-47b4-aaad-6a764a92493b.json
View File

@@ -70,8 +70,7 @@
"dest": "45_1.pptx"
},
"options": {
"examine_shape": false,
"examine_shape_for_shift_size": true
"examine_shape": false
}
},
"proxy": false

43
evaluation_examples/examples/libreoffice_impress/4ed5abd0-8b5d-47bd-839f-cacfa15ca37a.json
View File

@@ -58,20 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/4ed5abd0-8b5d-47bd-839f-cacfa15ca37a/4_1_Gold.pptx",
"dest": "4_1_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/4_1.pptx",
"dest": "4_1.pptx"
},
"options": {
"examine_shape": false
}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/4ed5abd0-8b5d-47bd-839f-cacfa15ca37a/4_1_Gold.pptx",
"dest": "4_1_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/4ed5abd0-8b5d-47bd-839f-cacfa15ca37a/4_1_Gold_all_fonts.pptx",
"dest": "4_1_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/4_1.pptx",
"dest": "4_1.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/4_1.pptx",
"dest": "4_1.pptx"
}
]
},
"proxy": false
}

40
evaluation_examples/examples/libreoffice_impress/550ce7e7-747b-495f-b122-acdc4d0b8e54.json
View File

@@ -74,17 +74,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/550ce7e7-747b-495f-b122-acdc4d0b8e54/New_Club_Spring_2018_Training_with_strike.data",
"dest": "New_Club_Spring_2018_Training_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/New_Club_Spring_2018_Training.pptx",
"dest": "New_Club_Spring_2018_Training.pptx"
}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/550ce7e7-747b-495f-b122-acdc4d0b8e54/New_Club_Spring_2018_Training_with_strike.data",
"dest": "New_Club_Spring_2018_Training_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/550ce7e7-747b-495f-b122-acdc4d0b8e54/New_Club_Spring_2018_Training_Gold_all_fonts.pptx",
"dest": "New_Club_Spring_2018_Training_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/New_Club_Spring_2018_Training.pptx",
"dest": "New_Club_Spring_2018_Training.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/New_Club_Spring_2018_Training.pptx",
"dest": "New_Club_Spring_2018_Training.pptx"
}
]
},
"proxy": false
}

41
evaluation_examples/examples/libreoffice_impress/57667013-ea97-417c-9dce-2713091e6e2a.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/57667013-ea97-417c-9dce-2713091e6e2a/1_2_Gold.pptx",
"dest": "1_2_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/1_2.pptx",
"dest": "1_2.pptx"
},
"options": {}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/57667013-ea97-417c-9dce-2713091e6e2a/1_2_Gold.pptx",
"dest": "1_2_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/57667013-ea97-417c-9dce-2713091e6e2a/1_2_Gold_all_fonts.pptx",
"dest": "1_2_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/1_2.pptx",
"dest": "1_2.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/1_2.pptx",
"dest": "1_2.pptx"
}
]
},
"proxy": false
}

41
evaluation_examples/examples/libreoffice_impress/5c1a6c3d-c1b3-47cb-9b01-8d1b7544ffa1.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/5c1a6c3d-c1b3-47cb-9b01-8d1b7544ffa1/39_2_Gold.pptx",
"dest": "39_2_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/39_2.pptx",
"dest": "39_2.pptx"
},
"options": {}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/5c1a6c3d-c1b3-47cb-9b01-8d1b7544ffa1/39_2_Gold.pptx",
"dest": "39_2_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/5c1a6c3d-c1b3-47cb-9b01-8d1b7544ffa1/39_2_Gold_all_fonts.pptx",
"dest": "39_2_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/39_2.pptx",
"dest": "39_2.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/39_2.pptx",
"dest": "39_2.pptx"
}
]
},
"proxy": false
}

35
evaluation_examples/examples/libreoffice_impress/5cfb9197-e72b-454b-900e-c06b0c802b40.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/5cfb9197-e72b-454b-900e-c06b0c802b40/33_1_Gold.pptx",
"dest": "33_1_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/33_1.pptx",
"dest": "33_1.pptx"
},
"options": {}
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/5cfb9197-e72b-454b-900e-c06b0c802b40/33_1_Gold_all_fonts.pptx",
"dest": "33_1_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/33_1.pptx",
"dest": "33_1.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/33_1.pptx",
"dest": "33_1.pptx"
}
]
},
"proxy": false
}

59
evaluation_examples/examples/libreoffice_impress/7ae48c60-f143-4119-b659-15b8f485eb9a.json
View File

@@ -1,7 +1,7 @@
{
"id": "7ae48c60-f143-4119-b659-15b8f485eb9a",
"snapshot": "libreoffice_impress",
"instruction": "Change picture's height to 20,30,25cm on slide 3,4,6 respectively.",
"instruction": "Change picture's height to 20, 30, 25cm on slide 3, 4, 6 respectively.",
"hint": "The input unit is cm",
"source": "https://arxiv.org/pdf/2311.01767.pdf",
"config": [
@@ -28,6 +28,7 @@
"libreoffice_impress"
],
"evaluator": {
"_config_note": "CRITICAL: examine_shape must be False for examine_modify_height to work. Due to execution order, examine_shape check happens before examine_modify_height, so any shape dimension mismatch will cause immediate failure before examine_modify_height logic is reached.",
"postconfig": [
{
"type": "activate_window",
@@ -59,21 +60,47 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/7ae48c60-f143-4119-b659-15b8f485eb9a/30_1_Gold.pptx",
"dest": "30_1_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/30_1.pptx",
"dest": "30_1.pptx"
},
"options": {
"examine_shape": false,
"examine_modify_height": true
}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/7ae48c60-f143-4119-b659-15b8f485eb9a/30_1_Gold.pptx",
"dest": "30_1_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/7ae48c60-f143-4119-b659-15b8f485eb9a/30_1_Gold_all_fonts.pptx",
"dest": "30_1_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/30_1.pptx",
"dest": "30_1.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/30_1.pptx",
"dest": "30_1.pptx"
}
],
"options": [
{
"_note": "examine_shape=False is REQUIRED for examine_modify_height to work - see slides.py for execution order details",
"examine_modify_height": true,
"examine_shape": false
},
{
"_note": "examine_shape=False is REQUIRED for examine_modify_height to work - see slides.py for execution order details",
"examine_modify_height": true,
"examine_shape": false
}
]
},
"proxy": false
}

42
evaluation_examples/examples/libreoffice_impress/9ec204e4-f0a3-42f8-8458-b772a6797cab.json
View File

@@ -1,7 +1,7 @@
{
"id": "9ec204e4-f0a3-42f8-8458-b772a6797cab",
"snapshot": "libreoffice_impress",
"instruction":"Please duplicate the last two slides and insert the copies in alternating order, so the sequence becomes: original slide A, original slide B, then duplicated slide A, duplicated slide B.",
"instruction": "Please duplicate the last two slides and insert the copies in alternating order, so the sequence becomes: original slide A, original slide B, then duplicated slide A, duplicated slide B.",
"source": "https://www.tiktok.com/@lil.d1rt_/video/7247574148887629083",
"config": [
{
@@ -58,17 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/9ec204e4-f0a3-42f8-8458-b772a6797cab/MLA_Workshop_061X_Works_Cited_Gold.pptx",
"dest": "MLA_Workshop_061X_Works_Cited_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/MLA_Workshop_061X_Works_Cited.pptx",
"dest": "MLA_Workshop_061X_Works_Cited.pptx"
}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/9ec204e4-f0a3-42f8-8458-b772a6797cab/MLA_Workshop_061X_Works_Cited_Gold.pptx",
"dest": "MLA_Workshop_061X_Works_Cited_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/9ec204e4-f0a3-42f8-8458-b772a6797cab/MLA_Workshop_061X_Works_Cited_Gold_all_fonts.pptx",
"dest": "MLA_Workshop_061X_Works_Cited_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/MLA_Workshop_061X_Works_Cited.pptx",
"dest": "MLA_Workshop_061X_Works_Cited.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/MLA_Workshop_061X_Works_Cited.pptx",
"dest": "MLA_Workshop_061X_Works_Cited.pptx"
}
]
},
"proxy": false
}

41
evaluation_examples/examples/libreoffice_impress/a53f80cd-4a90-4490-8310-097b011433f6.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/a53f80cd-4a90-4490-8310-097b011433f6/21_0_Gold.pptx",
"dest": "21_0_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/21_0.pptx",
"dest": "21_0.pptx"
},
"options": {}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/a53f80cd-4a90-4490-8310-097b011433f6/21_0_Gold.pptx",
"dest": "21_0_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/a53f80cd-4a90-4490-8310-097b011433f6/21_0_Gold_all_fonts_3.pptx",
"dest": "21_0_Gold_all_fonts_3.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/21_0.pptx",
"dest": "21_0.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/21_0.pptx",
"dest": "21_0.pptx"
}
]
},
"proxy": false
}

4
evaluation_examples/examples/libreoffice_impress/b8adbc24-cef2-4b15-99d5-ecbe7ff445eb.json
View File

@@ -72,7 +72,9 @@
"options": {
"examine_alignment": false,
"examine_font_name": false,
"examine_shape": false
"examine_shape": false,
"examine_font_size": false
}
},
"proxy": false

2
evaluation_examples/examples/libreoffice_impress/bf4e9888-f10f-47af-8dba-76413038b73c.json
View File

@@ -1,7 +1,7 @@
{
"id": "bf4e9888-f10f-47af-8dba-76413038b73c",
"snapshot": "libreoffice_impress",
"instruction": "I have a series of .png images named pic1.png to pic6.png saved on the Desktop. I want to create a new presentation with six blank slides, and insert one image per slide in order—pic1.png on the first slide, pic2.png on the second, and so on—following the numerical order. The goal is to prepare a presentation suitable for continuous playback in a kiosk or multimedia show. Could you help me with that?",
"instruction": "I have a series of .png images named pic1.png to pic6.png saved on the Desktop. I want to create a new presentation with six blank slides(which means no textbox or any other stuff on the page), and insert one image per slide in order—pic1.png on the first slide, pic2.png on the second, and so on—following the numerical order. The goal is to prepare a presentation suitable for continuous playback in a kiosk or multimedia show. Could you help me with that?",
"source": "https://help.libreoffice.org/6.4/en-US/text/simpress/guide/photo_album.html?DbPAR=IMPRESS",
"config": [
{

41
evaluation_examples/examples/libreoffice_impress/e4ef0baf-4b52-4590-a47e-d4d464cca2d7.json
View File

@@ -58,18 +58,35 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/e4ef0baf-4b52-4590-a47e-d4d464cca2d7/42_2_Gold.pptx",
"dest": "42_2_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/42_2.pptx",
"dest": "42_2.pptx"
},
"options": {}
"func": [
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/e4ef0baf-4b52-4590-a47e-d4d464cca2d7/42_2_Gold.pptx",
"dest": "42_2_Gold.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/e4ef0baf-4b52-4590-a47e-d4d464cca2d7/42_2_Gold_all_fonts.pptx",
"dest": "42_2_Gold_all_fonts.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/42_2.pptx",
"dest": "42_2.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/42_2.pptx",
"dest": "42_2.pptx"
}
]
},
"proxy": false
}

57
evaluation_examples/examples/libreoffice_impress/f23acfd2-c485-4b7c-a1e7-d4303ddfe864.json
View File

@@ -58,18 +58,57 @@
}
}
],
"func": "compare_pptx_files",
"expected": {
"func": [
"compare_pptx_files",
"compare_pptx_files",
"compare_pptx_files",
"compare_pptx_files"
],
"conj": "or",
"expected": [
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/f23acfd2-c485-4b7c-a1e7-d4303ddfe864/69_4_Gold.pptx",
"dest": "69_4_Gold.pptx"
},
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/69_4.pptx",
"dest": "69_4.pptx"
},
"options": {}
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/f23acfd2-c485-4b7c-a1e7-d4303ddfe864/69_4_Gold2.pptx",
"dest": "69_4_Gold2.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/f23acfd2-c485-4b7c-a1e7-d4303ddfe864/69_4_Gold_all_fonts_1.pptx",
"dest": "69_4_Gold_all_fonts_1.pptx"
},
{
"type": "cloud_file",
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/libreoffice_impress/f23acfd2-c485-4b7c-a1e7-d4303ddfe864/69_4_Gold_all_fonts_2.pptx",
"dest": "69_4_Gold_all_fonts_2.pptx"
}
],
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/69_4.pptx",
"dest": "69_4.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/69_4.pptx",
"dest": "69_4.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/69_4.pptx",
"dest": "69_4.pptx"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/69_4.pptx",
"dest": "69_4.pptx"
}
]
},
"proxy": false
}

57
evaluation_examples/examples/multi_apps/02ce9a50-7af2-47ed-8596-af0c230501f8.json
View File

@@ -44,19 +44,52 @@
"os"
],
"evaluator": {
"func": "compare_image_text",
"result": {
"type": "vm_file",
"path": "/home/user/Desktop/ls.png",
"dest": "ls.png"
},
"expected": {
"type": "rule",
"rules": {
"type": "text",
"text": "ls"
"func": [
"compare_image_text",
"compare_image_text",
"compare_image_text"
],
"conj": "or",
"result": [
{
"type": "vm_file",
"path": "/home/user/Desktop/ls.png",
"dest": "ls.png"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/ls.png",
"dest": "ls.png"
},
{
"type": "vm_file",
"path": "/home/user/Desktop/ls.png",
"dest": "ls.png"
}
}
],
"expected": [
{
"type": "rule",
"rules": {
"type": "text",
"text": "ls"
}
},
{
"type": "rule",
"rules": {
"type": "text",
"text": "1s"
}
},
{
"type": "rule",
"rules": {
"type": "text",
"text": "1s"
}
}
]
},
"proxy": false
}

93
evaluation_examples/examples/multi_apps/ce2b64a2-ddc1-4f91-8c7d-a88be7121aac.json
View File

@@ -69,19 +69,102 @@
"rules": {
"expected": {
"ec076282f61ba74642e94b5a6a1250c6988204d59d9b02936606b6b8ef1e4433": [
"Kilimanjaro"
"Kili",
"kili",
"Kili.jpg",
"kili.jpg",
"Kilimanjaro",
"kilimanjaro",
"Kilimanjaro.jpg",
"kilimanjaro.jpg",
"Mount Kilimanjaro",
"mount kilimanjaro",
"Mount Kilimanjaro.jpg",
"mount kilimanjaro.jpg",
"Kilimanjaro Mountain",
"kilimanjaro mountain",
"Kilimanjaro Mountain.jpg",
"kilimanjaro mountain.jpg"
],
"6ed4239ecc2be3ec15ad65a78c5c823b9004d640b8cc83a6a7af5930f354de91": [
"Himalayas",
"Everest",
"Sagarmatha"
"everest",
"Everest.jpg",
"everest.jpg",
"Mount Everest",
"mount everest",
"Mount Everest.jpg",
"mount everest.jpg",
"Everest Mountain",
"everest mountain",
"Everest Mountain.jpg",
"everest mountain.jpg",
"Sagarmatha",
"sagarmatha",
"Sagarmatha.jpg",
"sagarmatha.jpg",
"Sagarmatha Mountain",
"sagarmatha mountain",
"Sagarmatha Mountain.jpg",
"sagarmatha mountain.jpg",
"Chomolungma",
"chomolungma",
"Chomolungma.jpg",
"chomolungma.jpg",
"Qomolangma",
"qomolangma",
"Qomolangma.jpg",
"qomolangma.jpg",
"Himalayas",
"himalayas",
"Himalayas.jpg",
"himalayas.jpg",
"Himalayas Mountain",
"himalayas mountain",
"Himalayas Mountain.jpg",
"himalayas mountain.jpg",
"Himalaya",
"himalaya",
"Himalaya.jpg",
"himalaya.jpg",
"Himalaya Mountain",
"himalaya mountain",
"Himalaya Mountain.jpg",
"himalaya mountain.jpg"
],
"79f45d40d8413d4e81f1b9734ea39e58622cafd79e12bab32959643fc245147c": [
"Hua",
"hua"
"hua",
"Hua.jpg",
"hua.jpg",
"Mount Hua",
"mount hua",
"Mount Hua.jpg",
"mount hua.jpg",
"Hua Mountain",
"hua mountain",
"Hua Mountain.jpg",
"hua mountain.jpg",
"Huashan",
"huashan",
"Huashan.jpg",
"huashan.jpg",
"Hua Shan",
"hua shan",
"Hua Shan.jpg",
"hua shan.jpg",
"Huashan Mountain",
"huashan mountain",
"Huashan Mountain.jpg",
"huashan mountain.jpg",
"Hua Shan Mountain",
"hua shan mountain",
"Hua Shan Mountain.jpg",
"hua shan mountain.jpg"
]
},
"expect_in_result": true
"expect_in_result": true,
"result_not_list": true
}
}
},

2
evaluation_examples/examples/multi_apps/da922383-bfa4-4cd3-bbad-6bebab3d7742.json
View File

@@ -1,7 +1,7 @@
{
"id": "da922383-bfa4-4cd3-bbad-6bebab3d7742",
"snapshot": "multiapps",
"instruction": "I browsed a lot of interesting blog articles today. I hope to store these articles in my local designated folder just like zotero stores papers. Please download the blogs opening now in pdf format and save them in their tile to /home/user/Documents/Blog.",
"instruction": "I browsed a lot of interesting blog articles today. I hope to store these articles in my local designated folder just like zotero stores papers. Please download the blogs opening now in pdf format and save them in their title to /home/user/Documents/Blog.",
"source": "authors",
"config": [
{

7
evaluation_examples/examples/multi_apps/e135df7c-7687-4ac0-a5f0-76b74438b53e.json
View File

@@ -89,7 +89,14 @@
"path": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/multi_apps/e135df7c-7687-4ac0-a5f0-76b74438b53e/annual-enterprise-survey-2021-financial-year-provisional.html",
"dest": "annual-enterprise-survey-2021-financial-year-provisional_gold.html"
}
],
"options": [
{},
{
"ignore_sdnum": true
}
]
},
"proxy": true
}

4
evaluation_examples/examples/multi_apps/e8172110-ec08-421b-a6f5-842e6451911f.json
View File

@@ -36,8 +36,8 @@
],
"evaluator": {
"func": [
"check_structure_sim",
"check_structure_sim"
"check_structure_sim_resized",
"check_structure_sim_resized"
],
"result": [
{

6
evaluation_examples/examples/multi_apps/f8cfa149-d1c1-4215-8dac-4a0932bad3c2.json
View File

@@ -10,7 +10,7 @@
"files": [
{
"url": "https://huggingface.co/datasets/xlangai/ubuntu_osworld_file_cache/resolve/main/multi_apps/f8cfa149-d1c1-4215-8dac-4a0932bad3c2/file.xls",
"path": "/home/user/cell_search.xlsx"
"path": "/home/user/cell_search.xls"
}
]
}
@@ -47,7 +47,7 @@
{
"type": "open",
"parameters": {
"path": "/home/user/cell_search.xlsx"
"path": "/home/user/cell_search.xls"
}
}
],
@@ -65,7 +65,7 @@
"type": "rule",
"rules": {
"expect": {
"pattern": "www\\.google\\.com.*?/search\\?q=Nereida&"
"pattern": "https?://(www\\.?)?google\\.com/search\\?q=nereida(&|$)"
}
}
}

4
evaluation_examples/settings/proxy/dataimpulse.json
View File

@@ -2,8 +2,8 @@
{
"host": "gw.dataimpulse.com",
"port": 823,
"username": "fba5ac061fe18be70c6c",
"password": "e225c50bf56bdd6c",
"username": "your_username",
"password": "your_password",
"protocol": "http",
"provider": "dataimpulse",
"type": "residential",

2
main.py
View File

@@ -83,4 +83,4 @@ def human_agent():
if __name__ == "__main__":
human_agent()
human_agent()

1387
mm_agents/gat1_agent.py Normal file
View File

File diff suppressed because it is too large Load Diff

1387
mm_agents/gta1_agent.py Normal file
View File

File diff suppressed because it is too large Load Diff

179
mm_agents/prompts.py
View File

@@ -1338,3 +1338,182 @@ Here are some guidelines for you:
My computer's password is 'password', feel free to use it when you need sudo rights.
First give the current screenshot and previous things we did a short reflection, then RETURN ME THE CODE OR SPECIAL CODE I ASKED FOR NEVER EVER RETURN ME ANYTHING ELSE.
"""
GTA1_PLANNER_SYSTEM_PROMPT = """You are an agent which follow my instruction and perform desktop computer tasks as instructed.
You have good knowledge of computer and good internet connection and assume your code will run on a computer for controlling the mouse and keyboard.
You are on Ubuntu operating system and the resolution of the screen is 1920x1080.
For each step, you will get:
- An observation of an image, which is the screenshot of the computer screen and you will predict the action of the computer based on the image.
- Access to the following class and methods to interact with the UI:
class Agent:
def click(self, instruction: str, num_clicks: int = 1, button_type: str = 'left', hold_keys: List = []):
'''Click on the element
Args:
instruction:str, decribe the element you want to interact with in detail including the visual description and function description. And make it clear and concise. For example you can describe what the element looks like, and what will be the expected result when you interact with it.
num_clicks:int, number of times to click the element
button_type:str, which mouse button to press can be "left", "middle", or "right"
hold_keys:List, list of keys to hold while clicking
'''
def done(self, return_value: Union[Dict, str, List, Tuple, int, float, bool, NoneType] = None):
'''End the current task with a success and the required return value'''
def drag_and_drop(self, starting_description: str, ending_description: str, hold_keys: List = []):
'''Drag from the starting description to the ending description
Args:
starting_description:str, a very detailed description of where to start the drag action. This description should be at least a full sentence. And make it clear and concise.
ending_description:str, a very detailed description of where to end the drag action. This description should be at least a full sentence. And make it clear and concise.
hold_keys:List list of keys to hold while dragging
'''
def fail(self):
'''End the current task with a failure, and replan the whole task.'''
def highlight_text_span(self, starting_phrase: str, ending_phrase: str):
'''Highlight a text span between a provided starting phrase and ending phrase. Use this to highlight words, lines, and paragraphs.
Args:
starting_phrase:str, the phrase that denotes the start of the text span you want to highlight. If you only want to highlight one word, just pass in that single word.
ending_phrase:str, the phrase that denotes the end of the text span you want to highlight. If you only want to highlight one word, just pass in that single word.
'''
def hold_and_press(self, hold_keys: List, press_keys: List):
'''Hold a list of keys and press a list of keys
Args:
hold_keys:List, list of keys to hold
press_keys:List, list of keys to press in a sequence
'''
def hotkey(self, keys: List):
'''Press a hotkey combination
Args:
keys:List the keys to press in combination in a list format (e.g. ['ctrl', 'c'])
'''
def open(self, app_or_filename: str):
'''Open any application or file with name app_or_filename. Use this action to open applications or files on the desktop, do not open manually.
Args:
app_or_filename:str, the name of the application or filename to open
'''
def scroll(self, instruction: str, clicks: int, shift: bool = False):
'''Scroll the element in the specified direction
Args:
instruction:str, a very detailed description of which element to enter scroll in. This description should be at least a full sentence. And make it clear and concise.
clicks:int, the number of clicks to scroll can be positive (up) or negative (down).
shift:bool, whether to use shift+scroll for horizontal scrolling
'''
def set_cell_values(self, cell_values: Dict[str, Any], app_name: str, sheet_name: str):
'''Use this to set individual cell values in a spreadsheet. For example, setting A2 to "hello" would be done by passing {"A2": "hello"} as cell_values. The sheet must be opened before this command can be used.
Args:
cell_values: Dict[str, Any], A dictionary of cell values to set in the spreadsheet. The keys are the cell coordinates in the format "A1", "B2", etc.
Supported value types include: float, int, string, bool, formulas.
app_name: str, The name of the spreadsheet application. For example, "Some_sheet.xlsx".
sheet_name: str, The name of the sheet in the spreadsheet. For example, "Sheet1".
'''
def switch_applications(self, app_code):
'''Switch to a different application that is already open
Args:
app_code:str the code name of the application to switch to from the provided list of open applications
'''
def type(self, element_description: Optional[str] = None, text: str = '', overwrite: bool = False, enter: bool = False):
'''Type text into a specific element
Args:
element_description:str, a detailed description of which element to enter text in. This description should be at least a full sentence.
text:str, the text to type
overwrite:bool, Assign it to True if the text should overwrite the existing text, otherwise assign it to False. Using this argument clears all text in an element.
enter:bool, Assign it to True if the enter key should be pressed after typing the text, otherwise assign it to False.
'''
def wait(self, time: float):
'''Wait for a specified amount of time
Args:
time:float the amount of time to wait in seconds
'''
The following rules are IMPORTANT:
- If previous actions didn't achieve the expected result, do not repeat them, especially the last one. Try to adjust either the coordinate or the action based on the new screenshot.
- Do not predict multiple clicks at once. Base each action on the current screenshot; do not predict actions for elements or events not yet visible in the screenshot.
- You cannot complete the task by outputting text content in your response. You must use mouse and keyboard to interact with the computer. Call ```agent.fail()``` function when you think the task can not be done.
- You must use only the available methods provided above to interact with the UI, do not invent new methods.
You should provide a detailed observation of the current computer state based on the full screenshot in detail in the "Observation:" section.
Provide any information that is possibly relevant to achieving the task goal and any elements that may affect the task execution, such as pop-ups, notifications, error messages, loading states, etc..
You MUST return the observation before the thought.
You should think step by step and provide a detailed thought process before generating the next action:
Thought:
- Step by Step Progress Assessment:
- Analyze completed task parts and their contribution to the overall goal
- Reflect on potential errors, unexpected results, or obstacles
- If previous action was incorrect, predict a logical recovery step
- Next Action Analysis:
- List possible next actions based on current state
- Evaluate options considering current state and previous actions
- Propose most logical next action
- Anticipate consequences of the proposed action
Your thought should be returned in "Thought:" section. You MUST return the thought before the code.
You are required to use `agent` class methods to perform the action grounded to the observation.
Return exactly ONE line of python code to perform the action each time. At each step (example: ```agent.click('Click \"Yes, I trust the authors\" button', 1, 'left')\n```)
Remember you should only return ONE line of code, DO NOT RETURN more. You should return the code inside a code block, like this:
```python
agent.click('Click \"Yes, I trust the authors\" button', 1, "left")
```
For your reference, you have maximum of 100 steps, and current step is {current_step} out of {max_steps}.
If you are in the last step, you should return ```agent.done()``` or ```agent.fail()``` according to the result.
Here are some guidelines for you:
1. Remember to generate the corresponding instruction to the code before a # in a comment and only return ONE line of code.
2. `agent.click` can have multiple clicks. For example, agent.click('Click \"Yes, I trust the authors\" button', 2, "left") is double click.
3. Return ```agent.done()``` in the code block when you think the task is done (Be careful when evaluating whether the task has been successfully completed). Return ```agent.fail()``` in the code block when you think the task can not be done.
4. Whenever possible, your grounded action should use hot-keys with the agent.hotkey() action instead of clicking or dragging.
5. Save modified files before returning ```agent.done()```. When you finish modifying a file, always save it before proceeding using ```agent.hotkey(['ctrl', 's'])``` or equivalent. Tasks may involve multiple files. Save each after finishing modification.
6. If you meet "Authentication required" prompt, you can continue to click "Cancel" to close it.
My computer's password is 'password', feel free to use it when you need sudo rights.
First give the current screenshot and previous things we did a short reflection, then RETURN ME THE CODE I ASKED FOR NEVER EVER RETURN ME ANYTHING ELSE."""
GTA1_GROUNDING_SYSTEM_PROMPT = '''
You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. The image resolution is height {height} and width {width}. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
'''.strip()
GTA1_JUDGE_SYSTEM_PROMPT='''
You are an expert at evaluating the planning and reasoning of UI agents working toward achieving a goal.
My computer's password is 'password', feel free to use it when you need sudo rights or login.
Each time, I will provide you with:
- The current screenshot of the UI of width {width} and height {height}
- The goal of the task
- Past histories of planning and actions that have been taken
- A list of {N_PLANNING} different planning approaches toward achieving the goal in the current state in this form:
Observation: <screenshot caption>
Thought: <planning and reasoning>
Action: <UI action>
Your task is to select the single most effective planning approach that best advances toward the goal.
Evaluation criteria:
- Correctness: Does the action move closer to the goal?
- Effectiveness: Does it make meaningful progress immediately?
- Alignment: Does it support both immediate steps and long-term objectives?
- Planning quality: Is the thought process clear, concise, and logical?
- Appropriateness: Is the action valid and executable in the current UI context?
Note that some planning approaches may be similar - do not let the number of similar approaches dominate your decision. Evaluate each planning on its own merits.
Respond **only** with valid JSON (no extra keys or comments):
```json
{{
"explaining": "Your explanation of why this planning is best using the evaluation criteria",
"index": The index of the best planning (0, 1, ..., {N_INDEX})
}}
```
'''.strip()