采购到付款 - 发票处理 data template

import requests
import pandas as pd
import json
import os
from datetime import datetime

class=class="hl-string">"hl-comment"># --- Configuration Placeholders ---
class=class="hl-string">"hl-comment"># Replace with your actual SAP Ariba credentials and environment details
ARIBE_BASE_URL = class="hl-string">"https://[your-api-gateway-url.com]/api/transactional-data-export/v1/"
API_KEY = class="hl-string">"[Your_API_Key]"
OAUTH_URL = class="hl-string">"https://api.ariba.com/v2/oauth/token"
OAUTH_CLIENT_ID = class="hl-string">"[Your_OAuth_Client_ID]"
OAUTH_CLIENT_SECRET = class="hl-string">"[Your_OAuth_Client_Secret]"
REALM = class="hl-string">"[Your_Company_Realm]"

class=class="hl-string">"hl-comment"># --- Functions ---
def get_oauth_token(client_id, client_secret):
    class="hl-string">"""Fetches OAuth2 token from SAP Ariba."""
    payload = {
        class="hl-string">'grant_type': class="hl-string">'client_credentials',
        class="hl-string">'client_id': client_id,
        class="hl-string">'client_secret': client_secret
    }
    try:
        response = requests.post(OAUTH_URL, data=payload)
        response.raise_for_status()
        return response.json().get(class="hl-string">'access_token')
    except requests.exceptions.RequestException as e:
        print(fclass="hl-string">"Error fetching OAuth token: {e}")
        return None

def fetch_ariba_data(access_token, created_date_from, created_date_to):
    class="hl-string">"""Fetches paginated invoice data from the Ariba API."""
    all_invoices = []
    page_token = None
    headers = {
        class="hl-string">'Authorization': fclass="hl-string">'Bearer {access_token}',
        class="hl-string">'apiKey': API_KEY,
        class="hl-string">'Accept': class="hl-string">'application/json'
    }
    
    while True:
        params = {
            class="hl-string">'realm': REALM,
            class="hl-string">'product': class="hl-string">'ariba.procurement',
            class="hl-string">'view': class="hl-string">'InvoiceReconciliationDocument',
            class="hl-string">'createdDateFrom': created_date_from,
            class="hl-string">'createdDateTo': created_date_to,
            class="hl-string">'pageSize': 100 class=class="hl-string">"hl-comment"># Adjust as needed
        }
        if page_token:
            params[class="hl-string">'pageToken'] = page_token

        print(fclass="hl-string">"Fetching data with page token: {page_token or 'Initial Page'}")
        try:
            response = requests.get(ARIBE_BASE_URL + class="hl-string">"views/InvoiceReconciliationDocument", headers=headers, params=params)
            response.raise_for_status()
            data = response.json()
            all_invoices.extend(data.get(class="hl-string">'records', []))
            page_token = data.get(class="hl-string">'pageToken')
            if not page_token:
                break
        except requests.exceptions.RequestException as e:
            print(fclass="hl-string">"Error fetching data from Ariba API: {e}")
            break

    return all_invoices

def transform_to_event_log(invoices):
    class="hl-string">"""Transforms raw invoice JSON data into a structured event log DataFrame."""
    event_log_records = []
    for inv in invoices:
        case_id = inv.get(class="hl-string">'uniqueName')
        if not case_id:
            continue

        class=class="hl-string">"hl-comment"># Common attributes
        common_attrs = {
            class="hl-string">'InvoiceNumber': case_id,
            class="hl-string">'PurchaseOrderNumber': inv.get(class="hl-string">'poNumber'),
            class="hl-string">'VendorName': inv.get(class="hl-string">'supplier', {}).get(class="hl-string">'name'),
            class="hl-string">'InvoiceAmount': inv.get(class="hl-string">'totalCost', {}).get(class="hl-string">'amount'),
            class="hl-string">'PaymentDueDate': inv.get(class="hl-string">'paymentTerms', {}).get(class="hl-string">'dueDate'),
            class="hl-string">'InvoiceStatus': inv.get(class="hl-string">'statusString')
        }

        class=class="hl-string">"hl-comment"># Activity: Invoice Received
        if inv.get(class="hl-string">'timeCreated'):
            event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Received',
                class="hl-string">'EventTime': inv[class="hl-string">'timeCreated'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })

        class=class="hl-string">"hl-comment"># Activity: Invoice Submitted for Reconciliation
        if inv.get(class="hl-string">'submittedDate'):
            event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Submitted for Reconciliation',
                class="hl-string">'EventTime': inv[class="hl-string">'submittedDate'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })
        
        class=class="hl-string">"hl-comment"># Activity: Matching Discrepancy Identified
        class=class="hl-string">"hl-comment"># This assumes an exception's creation time is available
        if inv.get(class="hl-string">'exceptionStatus') == class="hl-string">'Active' and inv.get(class="hl-string">'timeUpdated'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Matching Discrepancy Identified',
                class="hl-string">'EventTime': inv[class="hl-string">'timeUpdated'], class=class="hl-string">"hl-comment"># Use timeUpdated as a proxy for exception creation
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })

        class=class="hl-string">"hl-comment"># Activity: Discrepancy Resolved
        class=class="hl-string">"hl-comment"># Assumes a specific timestamp exists for this event
        if inv.get(class="hl-string">'exceptionClearedDate'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Discrepancy Resolved',
                class="hl-string">'EventTime': inv[class="hl-string">'exceptionClearedDate'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })

        class=class="hl-string">"hl-comment"># Activity: Invoice Reconciled
        if inv.get(class="hl-string">'reconciledDate'):
            event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Reconciled',
                class="hl-string">'EventTime': inv[class="hl-string">'reconciledDate'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })
            class=class="hl-string">"hl-comment"># Activity: Invoice Sent for Approval (often occurs immediately after reconciliation)
            event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Sent for Approval',
                class="hl-string">'EventTime': inv[class="hl-string">'reconciledDate'], class=class="hl-string">"hl-comment"># Using reconciledDate as the trigger time
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })

        class=class="hl-string">"hl-comment"># Activity: Invoice Approved
        if inv.get(class="hl-string">'approvedDate'):
            class=class="hl-string">"hl-comment"># Approver information may be nested differently depending on configuration
            approver_name = inv.get(class="hl-string">'lastApprover', {}).get(class="hl-string">'name', class="hl-string">'N/A') 
            event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Approved',
                class="hl-string">'EventTime': inv[class="hl-string">'approvedDate'],
                class="hl-string">'Approver': approver_name,
                class="hl-string">'PaymentBlockReason': None
            })

        class=class="hl-string">"hl-comment"># Activity: Invoice Rejected
        if inv.get(class="hl-string">'statusString') == class="hl-string">'Rejected' and inv.get(class="hl-string">'timeUpdated'):
            event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Rejected',
                class="hl-string">'EventTime': inv[class="hl-string">'timeUpdated'],
                class="hl-string">'Approver': inv.get(class="hl-string">'lastApprover', {}).get(class="hl-string">'name'),
                class="hl-string">'PaymentBlockReason': None
            })

        class=class="hl-string">"hl-comment"># Activities related to payment blocks, posting, and payments
        class=class="hl-string">"hl-comment"># These fields may not be standard and depend on ERP integration
        if inv.get(class="hl-string">'paymentBlockTimestamp'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Payment Block Set',
                class="hl-string">'EventTime': inv[class="hl-string">'paymentBlockTimestamp'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': inv.get(class="hl-string">'paymentBlockReason')
            })
        if inv.get(class="hl-string">'paymentBlockReleasedTimestamp'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Payment Block Released',
                class="hl-string">'EventTime': inv[class="hl-string">'paymentBlockReleasedTimestamp'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })
        if inv.get(class="hl-string">'erpPostingDate'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Invoice Posted',
                class="hl-string">'EventTime': inv[class="hl-string">'erpPostingDate'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })
        if inv.get(class="hl-string">'scheduledPaymentDate'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Payment Scheduled',
                class="hl-string">'EventTime': inv[class="hl-string">'scheduledPaymentDate'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })
        if inv.get(class="hl-string">'paidDate'):
             event_log_records.append({
                **common_attrs,
                class="hl-string">'ActivityName': class="hl-string">'Payment Executed',
                class="hl-string">'EventTime': inv[class="hl-string">'paidDate'],
                class="hl-string">'Approver': None,
                class="hl-string">'PaymentBlockReason': None
            })

    return pd.DataFrame(event_log_records)

def main():
    class="hl-string">"""Main function to execute the extraction process."""
    print(class="hl-string">"Starting SAP Ariba invoice data extraction...")
    
    class=class="hl-string">"hl-comment"># Define the extraction time window (e.g., last 3 months)
    created_date_from = class="hl-string">"2023-01-01T00:00:00Z"
    created_date_to = class="hl-string">"2023-03-31T23:59:59Z"

    access_token = get_oauth_token(OAUTH_CLIENT_ID, OAUTH_CLIENT_SECRET)
    if not access_token:
        print(class="hl-string">"Failed to get OAuth token. Exiting.")
        return

    raw_invoices = fetch_ariba_data(access_token, created_date_from, created_date_to)
    if not raw_invoices:
        print(class="hl-string">"No data found for the specified period. Exiting.")
        return
    
    print(fclass="hl-string">"Successfully fetched {len(raw_invoices)} invoice records.")
    
    event_log_df = transform_to_event_log(raw_invoices)
    
    if event_log_df.empty:
        print(class="hl-string">"Transformation resulted in an empty event log.")
        return

    class=class="hl-string">"hl-comment"># Sort by case and time for chronological order
    event_log_df = event_log_df.sort_values(by=[class="hl-string">'InvoiceNumber', class="hl-string">'EventTime'], ascending=[True, True])
    
    output_filename = class="hl-string">'ariba_invoice_event_log.csv'
    event_log_df.to_csv(output_filename, index=False, date_format=class="hl-string">'%Y-%m-%dT%H:%M:%SZ')
    print(fclass="hl-string">"Extraction complete. Event log saved to {output_filename}")

if __name__ == class="hl-string">"__main__":
    main()