Extract highlights Leave feedback
On this page
GroupDocs.Parser provides functionality to extract highlights - text fragments with surrounding context. This feature is useful for creating search result previews, text snippets, and contextual excerpts.
- GroupDocs.Parser for Python via .NET installed
- Sample documents for testing
- Understanding of text positioning concepts
Highlights are text fragments extracted from a document at a specific position, including:
- Text: The main text content
- Position: Character position in the document
- Context: Surrounding text before and after
To extract a highlight from a specific position:
from groupdocs.parser import Parser
from groupdocs.parser.options import HighlightOptions
# Create an instance of Parser class
with Parser("./sample.pdf") as parser:
# Create highlight options (extract 20 characters)
options = HighlightOptions(20)
# Extract highlight at position 100
highlight = parser.get_highlight(100, True, options)
if highlight is not None:
print(f"Position: {highlight.position}")
print(f"Text: {highlight.text}")
else:
print("Highlight extraction isn't supported")
The following sample file is used in this example: sample.pdf
Expected behavior: Returns a HighlightItem object containing the text fragment starting at the specified position, or None if extraction is not supported.
To extract highlights with a specific character count:
from groupdocs.parser import Parser
from groupdocs.parser.options import HighlightOptions
# Create an instance of Parser class
with Parser("sample.docx") as parser:
# Define positions to extract highlights from
positions = [0, 100, 500, 1000]
# Create highlight options (50 characters)
options = HighlightOptions(50)
for pos in positions:
# Extract highlight
highlight = parser.get_highlight(pos, True, options)
if highlight is not None:
print(f"\nPosition {pos}:")
print(f"Text: {highlight.text}")
The following sample file is used in this example: sample.docx
Expected behavior: Extracts 50-character text fragments from specified positions.
To create search result previews with context:
from groupdocs.parser import Parser
from groupdocs.parser.options import SearchOptions, HighlightOptions
def get_search_results_with_context(file_path, keyword, context_length=30):
"""
Search for keyword and extract surrounding context.
"""
with Parser(file_path) as parser:
# Create highlight options for context
highlight_opts = HighlightOptions(context_length)
# Create search options with highlights
search_opts = SearchOptions(
match_case=False,
match_whole_word=False,
use_regular_expression=False,
search_by_pages=True,
left_highlight_options=highlight_opts,
right_highlight_options=highlight_opts
)
# Search for keyword
results = parser.search(keyword, search_opts)
if results is None:
print("Search not supported")
return []
# Collect results with context
search_results = []
for result in results:
left_context = result.left_highlight_item.text if result.left_highlight_item is not None else ""
right_context = result.right_highlight_item.text if result.right_highlight_item is not None else ""
search_results.append({
'keyword': result.text,
'position': result.position,
'page': result.page_index + 1 if result.page_index is not None else None,
'left_context': left_context,
'right_context': right_context,
'full_snippet': f"{left_context}[{result.text}]{right_context}"
})
return search_results
# Usage
results = get_search_results_with_context("sample.pdf", "harvest", 40)
print(f"Found {len(results)} occurrences:\n")
for idx, result in enumerate(results, 1):
print(f"{idx}. {result['full_snippet']}")
if result['page']:
print(f" (Page {result['page']}, position {result['position']})\n")
Found 6 occurrences:
1. [Harvest] range and by coffee from our long-stand
(Page 1, position 399)
2. 3. The Autumn [Harvest] range
(Page 2, position 551)
3. The Autumn [Harvest] range launched in September with eight
(Page 2, position 577)
[TRUNCATED]
Expected behavior: Searches for keywords and returns results with surrounding context for better previews.
To extract text snippets from various document sections:
from groupdocs.parser import Parser
from groupdocs.parser.options import HighlightOptions
def extract_multiple_highlights(file_path, positions, length=50):
"""
Extract highlights from multiple positions.
"""
highlights = []
with Parser(file_path) as parser:
options = HighlightOptions(length)
for position in positions:
highlight = parser.get_highlight(position, True, options)
if highlight is not None:
highlights.append({
'position': highlight.position,
'text': highlight.text,
'length': len(highlight.text)
})
return highlights
# Extract highlights at specific positions
positions = [0, 500, 1000, 1500]
highlights = extract_multiple_highlights("sample.pdf", positions, length=80)
print(f"Extracted {len(highlights)} highlights:\n")
for h in highlights:
print(f"Position {h['position']}:")
print(f" {h['text'][:60]}...")
print()
The following sample file is used in this example: sample.pdf
Extracted 4 highlights:
Position 0:
...
Position 500:
s...
Position 1000:
97.4% +1.3 pts...
[TRUNCATED]
Expected behavior: Extracts text fragments from multiple specified positions in the document.
To generate a document preview with highlighted sections:
from groupdocs.parser import Parser
from groupdocs.parser.options import HighlightOptions
def create_document_preview(file_path, num_sections=5, section_length=100):
"""
Create a document preview by extracting highlights from various positions.
"""
with Parser(file_path) as parser:
# Get document text to determine positions
text_reader = parser.get_text()
if text_reader is None:
print("Text extraction not supported")
return None
full_text = text_reader.read_to_end()
text_length = len(full_text)
# Calculate evenly distributed positions
if text_length <= section_length:
positions = [0]
else:
step = text_length // num_sections
positions = [i * step for i in range(num_sections)]
# Extract highlights
preview_sections = []
options = HighlightOptions(section_length)
for pos in positions:
highlight = parser.get_highlight(pos, True, options)
if highlight is not None:
preview_sections.append(highlight.text)
return {
'document_length': text_length,
'preview_sections': preview_sections,
'full_preview': '\n...\n'.join(preview_sections)
}
# Usage
preview = create_document_preview("sample.pdf", num_sections=3, section_length=150)
if preview:
print(f"Document length: {preview['document_length']} characters\n")
print("Preview:")
print(preview['full_preview'])
Document length: 4414 characters
Preview:
Investor relations
...
ed in April and holds 140 pallets at 2 C.
...
8% +9.2%
Expected behavior: Creates a representative preview of the document by extracting highlights from evenly distributed positions.
To extract highlights respecting word boundaries:
from groupdocs.parser import Parser
from groupdocs.parser.options import HighlightOptions
def extract_highlight_with_words(file_path, position, max_length=100):
"""
Extract highlight ensuring complete words.
"""
with Parser(file_path) as parser:
# Extract more than needed
options = HighlightOptions(max_length + 50)
highlight = parser.get_highlight(position, True, options)
if highlight is None:
return None
text = highlight.text
# Find last complete word within max_length
if len(text) > max_length:
truncated = text[:max_length]
last_space = truncated.rfind(' ')
if last_space > 0:
text = truncated[:last_space] + "..."
return {
'position': highlight.position,
'text': text,
'complete_words': True
}
# Usage
highlight = extract_highlight_with_words("sample.pdf", 500, 80)
if highlight:
print(f"Position {highlight['position']}:")
print(highlight['text'])
The following sample file is used in this example: sample.pdf
Position 500:
s
Expected behavior: Extracts highlights that end at word boundaries, avoiding cut-off words.
To create a table of contents with text previews:
from groupdocs.parser import Parser
from groupdocs.parser.options import HighlightOptions
def create_toc_with_previews(file_path, preview_length=60):
"""
Create table of contents with text previews for each section.
"""
with Parser(file_path) as parser:
# Get TOC
toc_items = parser.get_toc()
if not toc_items:
print("TOC extraction not supported")
return []
toc_with_previews = []
options = HighlightOptions(preview_length)
# Get text to find section positions
text_reader = parser.get_text()
if text_reader is None:
return []
full_text = text_reader.read_to_end()
for item in toc_items:
# Try to extract highlight for this TOC item
# Use text search to find the section
section_text = item.text
position = full_text.find(section_text)
if position >= 0:
# Start the preview on the line that follows the heading
next_line = full_text.find("\n", position) + 1
if 0 < next_line < len(full_text):
highlight = parser.get_highlight(next_line, True, options)
preview = highlight.text if highlight is not None else "No preview available"
else:
preview = "No preview available"
else:
preview = "Section not found"
toc_with_previews.append({
'title': item.text,
'depth': item.depth,
'preview': preview
})
return toc_with_previews
# Usage
toc = create_toc_with_previews("SampleWithToc.pdf", preview_length=100)
for item in toc:
indent = " " * item['depth']
print(f"{indent}{item['title']}")
print(f"{indent} Preview: {item['preview'][:50]}...")
The following sample file is used in this example: SampleWithToc.pdf
1. Overview
Preview: Harbor Lane Provisions supplies specialty food to ...
Key figures
Preview: Metric 2024 ...
2. Operations
Preview: Orders shipped rose 13.1% to 55,300 while on-time ...
Warehouse and logistics
Preview: • Cold room 3 opened in April and holds 140 ...
3. The Autumn Harvest range
Preview: The Autumn Harvest range launched in September wit...
[TRUNCATED]
Expected behavior: Generates a table of contents with text previews for each section.
- The
get_highlight()method returnsNoneif highlight extraction is not supported - The second parameter (
is_direct) sets the direction:Trueextracts the text to the right of the position,Falsethe text to the left of it - Highlight length is specified in characters
- Use highlights with search results to show context
- Highlights are useful for creating document previews and snippets
- Consider word boundaries when displaying highlights to users
Was this page helpful?
Any additional feedback you'd like to share with us?
Please tell us how we can improve this page.
Thank you for your feedback!
We value your opinion. Your feedback will help us improve our documentation.
