Rekognita SDK
Rekognita provides official SDKs for fast integration. Python is currently supported, with plans for Node.js, Go, and Java.
Installation
Install the Python SDK via pip:
pip install rekognitaOr via poetry:
poetry add rekognitaInitialization
Import and initialize the client:
from rekognita import RekognitaClient
client = RekognitaClient(api_key="rk_sk_your_key_here")
# Or use the REKOGNITA_API_KEY environment variable
client = RekognitaClient() # automatically reads from envBasic Usage
Converting a Document
# Upload and convert a document
result = client.documents.convert(
file="invoice.pdf",
output_format="markdown", # or "html", "json", "docx"
)
print(result.content) # Markdown/HTML content
print(result.pages) # Number of pages
print(result.metadata) # Document metadataExtracting Structured Data
# Extract data using a JSON Schema
result = client.documents.extract(
file="invoice.pdf",
schema={
"invoice_number": "string",
"total_amount": "number",
"line_items": [{
"description": "string",
"quantity": "number",
"price": "number"
}]
}
)
print(result.data) # {"invoice_number": "INV-2847", ...}Graph Structure
# Get the graph structure of the document
result = client.documents.parse(
file="report.pdf",
output_format="graph",
)
for node in result.nodes:
print(f"{node.type}: {node.text[:50]}...")
for child in node.children:
print(f" └─ {child.type}: {child.text[:30]}...")Asynchronous Client
from rekognita import AsyncRekognitaClient
async_client = AsyncRekognitaClient()
result = await async_client.documents.convert(
file="document.pdf",
output_format="markdown"
)Error Handling
from rekognita.exceptions import (
RekognitaAuthError,
RekognitaRateLimitError,
RekognitaValidationError,
)
try:
result = client.documents.convert(file="doc.pdf")
except RekognitaAuthError:
print("Invalid API Key")
except RekognitaRateLimitError as e:
print(f"Rate limit exceeded. Try again in {e.retry_after}s")
except RekognitaValidationError as e:
print(f"Validation error: {e.message}")Next Steps
- API Overview — full description of endpoints
- Graph Structure — how the document graph works
- Word/PDF — generating documents