app.py

from fastapi import FastAPI
import os


import pandas as pd
from io import StringIO 
import os
import base64

app = FastAPI()

def get_download_link_dify(df):
    # code to save file in dify framework
    import requests

    # API Configuration
    BASE_URL = "http://redmindgpt.redmindtechnologies.com:81/v1"
    DATASET_ID = "084ae979-d101-414b-8854-9bbf5d3a442e" 
    API_KEY = "dataset-feqz5KrqHkFRdWbh2DInt58L"  

    dataset_name = 'output_dataset'
    # Endpoint URL
    url = f"{BASE_URL}/datasets/{DATASET_ID}/document/create-by-file"
    print(url)
    # Headers
    headers = {
        "Authorization": f"Bearer {API_KEY}"
    }

    # Data payload (form data as a plain text string)
    data_payload = {
        "data": """
        {
            "indexing_technique": "high_quality",
            "process_rule": {
                "rules": {
                    "pre_processing_rules": [
                        {"id": "remove_extra_spaces", "enabled": true},
                        {"id": "remove_urls_emails", "enabled": true}
                    ],
                    "segmentation": {
                        "separator": "###",
                        "max_tokens": 500
                    }
                },
                "mode": "custom"
            }
        }
        """
    }

    # Convert DataFrame to binary (in-memory)
    file_buffer = dataframe_to_binary(df)
    
    files = {
        "file": ("output.xlsx", file_buffer, "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet")
    }

    # Send the POST request
    response = requests.post(url, headers=headers, data=data_payload, files=files)
    print(response)
    data = response.json()
    document_id = data['document']['id']

    # code to get download_url
    url = f"http://redmindgpt.redmindtechnologies.com:81/v1/datasets/{DATASET_ID}/documents/{document_id}/upload-file"
   
    response = requests.get(url, headers=headers)
    print(response)
    
    download_url = response.json().get("download_url")
    download_url = download_url.replace("download/","")
    return download_url

def dataframe_to_binary(df):
    import io
    # Create a BytesIO stream
    output = io.BytesIO()
    
    # Write the DataFrame to this in-memory buffer as an Excel file
    df.to_excel(output, index=False, engine="openpyxl")
    
    # Move the cursor to the beginning of the stream
    output.seek(0)
    
    return output
    


# FastAPI Endpoints
@app.get("/")
def greet_json():
    # Run Data Processing
    #process_and_store(pdf_path=pdf_file, pptx_path=pptx_file)
    return {"Document store": "created!"}

@app.get("/save_file_dify")
def save_file_dify(csv_data: str):    
    
    # Split into lines
    lines = csv_data.split("\n")

    # Find the max number of columns
    max_cols = max(line.count(",") + 1 for line in lines if line.strip())

    # Normalize all rows to have the same number of columns
    fixed_lines = [line + "," * (max_cols - line.count(",") - 1) for line in lines]

    # Reconstruct CSV string
    fixed_csv_data = "\n".join(fixed_lines)
    
    # Convert CSV string to DataFrame
    df = pd.read_csv(StringIO(fixed_csv_data))

    
    #save in dify dataset and return download link
    download_link = get_download_link_dify(df)
        
    return download_link