Hangul API Documentation

Overview

This documentation covers the usage of two API endpoints of Hangul provided by https://d4gumsi.pythonanywhere.com for handling document processing and summarization. Below is an outline of the endpoints, payload structure, and the sequence of API calls.

To learn more about Hangul here.

Jump to sections:

Document Processing Endpoint

Payload Parameters
Code Example
Response Keys

Endpoints

1. Document Processing Endpoint

URL: https://d4gumsi.pythonanywhere.com/api/v2/products/hangul

Method: POST

Description: This endpoint processes a given PDF document based on the specified payload parameters.

File: A PDF file to be sent along with the payload.

Payload Parameters

kw_num (str): Number of keywords to extract.

Return_ALL (bool): A flag to indicate if all available data should be returned. If this one True, it overrides all the rest.

Document-related flags (bool): Flags to specify if certain document attributes should be extracted:

document_language: The language of the document.
document_title: A selection of the most likely titles of the document with the weight of each title likelihood.
document_summary: The summary of the document and the keywords of the summary.
content: Content to be displayed, the text of up to 4 pages of the document.
report_type: Type of report.
locations: Locations mentioned in the report.
full_content: Clean text of the pdf.
markdown_text: Markdown of the pdf text.
document_theme: Theme of the document.
new_detected_disasters: Disasters detected in the document.

pdf Metadata information: This information is extracted from the metadata of the pdf file. It is available in the metadata dictionary of the response (All are bool type).

author: Author of the pdf document.
doc_created_date: The date of the document creation.
doc_modified_date: The date of the document modification.
num_of_pages: Number of pages in the document.
charsPerPage: Characters per page in the document.

Code example:

# Define API URL
        api_url_1 = 'https://d4gumsi.pythonanywhere.com/api/v2/products/hangul'

        # Define Payload
        payload = {
            'kw_num': "5",
            'Return_ALL': False,
            'document_language': False,
            'document_title': True,
            'document_summary': False,
            'content': False,
            'report_type': True,
            'locations': True,
            'full_content': True,
            'markdown_text': True,
            'document_theme': False,
            'new_detected_disasters': True,
            "Author": False,
            "doc_created_date": False,
            "doc_modified_date": False,
            "num_of_pages": False,
            "charsPerPage": False,
        }
        
        # Define File to be uploaded
        files = {
            "file": ("filename.pdf", open("filepath/filename.pdf", "rb")),
        }
        
        # Make the Request
        response = requests.post(api_url_1, files=files, data=payload)

        # Print the response
        print(response.json())

Response

The response will be a dictionary which contains various keys, one of which is document_summary_parameters that will be used for the next API call.

The response keys are the following:

'metadata': Returns a dictionary with metadata.
'document_language': Returns the language and the language score.
'document_title': List of title – Likelihood pairs
'document_summary_parameters': Dictionary with inputs to generate the summary.
'content': List of strings with the text of the first 4 pages of the document.
'report_type': string with the report-type.
'locations': Dictionary with locations and number of occurrences.
'full_content': Clean string with all the text of the document.
'markdown_text': String with the text of the document in markdown type.
'document_theme': Themes of the document in a list.
'new_detected_disasters': List of detected disasters in the document.