Unlock Data with Precision Key-Value Extraction
Automatically Identify and Extract Critical Information Pairs
1from docprompt import load_document_node
2from docprompt.tasks.factory import AnthropicTaskProviderFactory
3
4document_node = load_document_node("path/to/document")
5
6markerize_provider = AnthropicTaskProviderFactory().get_page_markerization_provider()
7document_markdown = markerize_provder.process_document_node()
8
9print(document_markdown[5]) # Get markdown for page 5Load documents in seconds
Ingest PDF's into your data pipeline with ease.
Convert markdown to HTML and parse the content efficiently.
Extract structured data from documents, including headings and paragraphs.
Customize document processing with flexible provider factories and task-specific providers.
Testimonial
“
I used the parallelized hyperparameter tuning with Prefect and Dask
(incredible) to run about 350 experiments in 30 minutes, which
normally would have taken 2 days
”
Andrew Waterman
Data Scientis,Actium
Testimonial
“
I used the parallelized hyperparameter tuning with Prefect and Dask
(incredible) to run about 350 experiments in 30 minutes, which
normally would have taken 2 days
”
Andrew Waterman
Data Scientis,Actium
test_file.py
1...
2# Get all headings
3@dataclass
4class ParagraphWithHeading:
5 heading: str
6 paragraph: str
7
8document_markdown = markerize_provder.process_document_node()
9
10# Convert markdown to HTML
11html = markdown.markdown(document_markdown[5])
12
13# Parse the HTML tree
14soup = BeautifulSoup(html, 'html.parser')
15
16paragraphs_with_headings = []
17
18# Collect all headings and their following paragraphs
19for heading in soup.find_all(['h1', 'h2', 'h3', 'h4', 'h5', 'h6']):
20 heading_text = heading.get_text(strip=True)
21 paragraph = heading.find_next('p')
22 if paragraph:
23 paragraphs_with_headings.append(ParagraphWithHeading(heading_text, paragraph.get_text()))Unlock Data with Precision Key-Value Extraction
Automatically Identify and Extract Critical Information Pairs
LEARN MOREUnlock Data with Precision Key-Value Extraction
Automatically Identify and Extract Critical Information Pairs
1...
2from langchain.text_splitter import MarkdownTextSplitter
3from itertools import chain
4
5document_markdown = markerize_provder.process_document_node()
6
7markdown_splitter = MarkdownTextSplitter(chunk_size=100, chunk_overlap=0)
8
9docs = markdown_splitter.create_documents(document_markdown.values()))Unlock Data with Precision Key-Value Extraction
Automatically Identify and Extract Critical Information Pairs
LEARN MORE