2.py

import win32com.client
import pandas as pd
import io
from datetime import datetime

# Connect to Outlook
outlook = win32com.client.Dispatch("Outlook.Application")
namespace = outlook.GetNamespace("MAPI")

# Recherche du dossier principal par son nom
main_folder_name = "NomDuDossierPrincipal"  # Remplacez "NomDuDossierPrincipal" par le nom du dossier principal
main_folder = None

for folder in namespace.Folders:
    if folder.Name == main_folder_name:
        main_folder = folder
        break

if main_folder:
    # Vous avez trouvé le dossier principal, vous pouvez maintenant y accéder.

    # Recherche du premier sous-dossier par son nom
    first_sub_folder_name = "NomDuPremierSousDossier"  # Remplacez "NomDuPremierSousDossier" par le nom du premier sous-dossier
    first_sub_folder = None

    for folder in main_folder.Folders:
        if folder.Name == first_sub_folder_name:
            first_sub_folder = folder
            break

    if first_sub_folder:
        # Vous avez trouvé le premier sous-dossier, vous pouvez maintenant y accéder.

        # Recherche du deuxième sous-dossier par son nom
        second_sub_folder_name = "NomDuDeuxiemeSousDossier"  # Remplacez "NomDuDeuxiemeSousDossier" par le nom du deuxième sous-dossier
        second_sub_folder = None

        for folder in first_sub_folder.Folders:
            if folder.Name == second_sub_folder_name:
                second_sub_folder = folder
                break

        if second_sub_folder:
            # Vous avez trouvé le deuxième sous-dossier, vous pouvez maintenant y accéder
            dataframes = []

            # Define the date range for filtering emails
            start_date = datetime(2023, 1, 1)  # Replace with your start date
            end_date = datetime(2023, 12, 31)  # Replace with your end date

            # Define the date format for the Restrict method
            date_format = "%Y-%m-%d"

            # Format the start and end dates
            start_date_str = start_date.strftime(date_format)
            end_date_str = end_date.strftime(date_format)

            # Filter emails based on received time
            filter_str = f"[ReceivedTime] >= '{start_date_str}' AND [ReceivedTime] <= '{end_date_str}'"
            filtered_emails = second_sub_folder.Items.Restrict(filter_str)

            # Itérer à travers les filtered_emails
            for email in filtered_emails:
                for attachment in email.Attachments:
                    if attachment.FileName.endswith(".csv"):
                        attachment_data = attachment.PropertyAccessor.GetProperty("http://schemas.microsoft.com/mapi/proptag/0x37010102")
                        attachment_data = attachment_data.encode("utf-8")
                        attachment_data = io.BytesIO(attachment_data)

                        # Read the CSV content into a DataFrame
                        df = pd.read_csv(attachment_data)
                        dataframes.append(df)

                    # Print sender, subject, and received time for each email
                    print(f"Sender: {email.SenderName}")
                    print(f"Subject: {email.Subject}")
                    print(f"Received Time: {email.ReceivedTime}")

            # Concatenate all DataFrames into one
            combined_df = pd.concat(dataframes, ignore_index=True)

            # Now you have the combined DataFrame for emails received within the specified date range.
            print(combined_df)

        else:
            print(f"Deuxième sous-dossier '{second_sub_folder_name}' non trouvé dans '{first_sub_folder_name}'.")
    else:
        print(f"Premier sous-dossier '{first_sub_folder_name}' non trouvé dans '{main_folder_name}'.")
else:
    print(f"Dossier principal '{main_folder_name}' non trouvé dans Outlook.")

Email_files.py

import win32com.client
import pandas as pd
import io
from datetime import datetime

# Connect to Outlook
outlook = win32com.client.Dispatch("Outlook.Application")
namespace = outlook.GetNamespace("MAPI")

# Recherche du dossier principal par son nom
main_folder_name = "NomDuDossierPrincipal"  # Remplacez "NomDuDossierPrincipal" par le nom du dossier principal
main_folder = None

for folder in namespace.Folders:
    if folder.Name == main_folder_name:
        main_folder = folder
        break

if main_folder:
    # Vous avez trouvé le dossier principal, vous pouvez maintenant y accéder.

    # Recherche du premier sous-dossier par son nom
    first_sub_folder_name = "NomDuPremierSousDossier"  # Remplacez "NomDuPremierSousDossier" par le nom du premier sous-dossier
    first_sub_folder = None

    for folder in main_folder.Folders:
        if folder.Name == first_sub_folder_name:
            first_sub_folder = folder
            break

    if first_sub_folder:
        # Vous avez trouvé le premier sous-dossier, vous pouvez maintenant y accéder.

        # Recherche du deuxième sous-dossier par son nom
        second_sub_folder_name = "NomDuDeuxiemeSousDossier"  # Remplacez "NomDuDeuxiemeSousDossier" par le nom du deuxième sous-dossier
        second_sub_folder = None

        for folder in first_sub_folder.Folders:
            if folder.Name == second_sub_folder_name:
                second_sub_folder = folder
                break

        if second_sub_folder:
            # Vous avez trouvé le deuxième sous-dossier, vous pouvez maintenant y accéder
            dataframes = []

            # Define the date range for filtering emails
            start_date = datetime(2023, 1, 1)  # Replace with your start date
            end_date = datetime(2023, 12, 31)  # Replace with your end date

            # Define the date format for the Restrict method
            date_format = "%Y-%m-%d"

            # Format the start and end dates
            start_date_str = start_date.strftime(date_format)
            end_date_str = end_date.strftime(date_format)

            # Filter emails based on received time
            filter_str = f"[ReceivedTime] >= '{start_date_str}' AND [ReceivedTime] <= '{end_date_str}'"
            filtered_emails = second_sub_folder.Items.Restrict(filter_str)

            # Itérer à travers les filtered_emails
            for email in filtered_emails:
                if email.Attachments.Count > 0:
                    for attachment in email.Attachments:
                        if attachment.FileName.endswith(".csv"):
                            attachment_content = attachment.Content
                            attachment_data = io.StringIO(attachment_content)
                            df = pd.read_csv(attachment_data)
                            dataframes.append(df)

                # Print sender, subject, and received time for each email
                print(f"Sender: {email.SenderName}")
                print(f"Subject: {email.Subject}")
                print(f"Received Time: {email.ReceivedTime}")

            # Concatenate all DataFrames into one
            combined_df = pd.concat(dataframes, ignore_index=True)

            # Now you have the combined DataFrame for emails received within the specified date range.
            print(combined_df)

        else:
            print(f"Deuxième sous-dossier '{second_sub_folder_name}' non trouvé dans '{first_sub_folder_name}'.")
    else:
        print(f"Premier sous-dossier '{first_sub_folder_name}' non trouvé dans '{main_folder_name}'.")
else:
    print(f"Dossier principal '{main_folder_name}' non trouvé dans Outlook.")