administrator/janssen
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

notebook

Browse Source
This commit is contained in:
Vladimir Buzalka
2026-05-30 07:33:06 +02:00
parent d7dbb92dd1
commit 9f955a40fe
29 changed files with 1419 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files
EmailsImport/DockerCustomApp/app.py
+229 -6
View File
@@ -1,17 +1,28 @@
# app.py | v1.0 | 2026-05-29
# FastAPI server pro příjem .msg a .db souborů a upload do Dropboxu.
# Endpointy: /upload (.msg → /msgs), /upload-db (.db → /msgs/db), /upload-dropbox (→ Dropbox /!!!Days/Downloads Z230).
# app.py | v1.3 | 2026-05-29
# FastAPI server pro příjem .msg a .db souborů, upload do Dropboxu a import do Graph API.
# Endpointy: /upload (.msg → /msgs + Graph import), /upload-db (.db → /msgs/db),
# /upload-dropbox (→ Dropbox /!!!Days/Downloads Z230).
from fastapi import FastAPI, UploadFile, File, Header, HTTPException
from fastapi import FastAPI, UploadFile, File, Form, Header, HTTPException
import shutil
import base64
import logging
from pathlib import Path
from typing import Optional
import os
import dropbox
import msal
import requests as http_requests
import extract_msg
from dateutil import parser as dtparser
from datetime import timezone
from dotenv import load_dotenv
load_dotenv(Path(__file__).parent / ".env")
app = FastAPI()
log = logging.getLogger("msgreceiver")
logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
TOKEN = "13e1bb01-9fd5-44a8-8ce9-4ee27133d340"
SAVE_DIR = Path("/msgs")
@@ -24,11 +35,213 @@ DROPBOX_APP_KEY = os.getenv("DROPBOX_APP_KEY", "")
DROPBOX_APP_SECRET = os.getenv("DROPBOX_APP_SECRET", "")
DROPBOX_REFRESH_TOKEN = os.getenv("DROPBOX_APP_REFRESH_TOKEN", "")
# --- Graph API config ---
GRAPH_TENANT_ID = "7d269944-37a4-43a1-8140-c7517dc426e9"
GRAPH_CLIENT_ID = "4b222bfd-78c9-4239-a53f-43006b3ed07f"
GRAPH_CLIENT_SECRET = "Txg8Q~MjhocuopxsJyJBhPmDfMxZ2r5WpTFj1dfk"
GRAPH_MAILBOX = "vladimir.buzalka@buzalka.cz"
GRAPH_ROOT_FOLDER = "JNJ" # subfolder under Inbox — root for imported emails
GRAPH_URL = "https://graph.microsoft.com/v1.0"
# Cache: folder path → Graph folder ID
_folder_id_cache: dict[str, str] = {}
_graph_token: Optional[str] = None
def _get_graph_token() -> str:
global _graph_token
msalapp = msal.ConfidentialClientApplication(
GRAPH_CLIENT_ID,
authority=f"https://login.microsoftonline.com/{GRAPH_TENANT_ID}",
client_credential=GRAPH_CLIENT_SECRET,
)
result = msalapp.acquire_token_for_client(scopes=["https://graph.microsoft.com/.default"])
if "access_token" not in result:
raise RuntimeError(f"Graph auth failed: {result}")
_graph_token = result["access_token"]
return _graph_token
def _graph_headers() -> dict:
token = _graph_token or _get_graph_token()
return {"Authorization": f"Bearer {token}", "Content-Type": "application/json"}
def _ensure_folder(path_parts: list[str]) -> str:
"""Ensure folder hierarchy exists under Inbox, return leaf folder ID."""
cache_key = "/".join(path_parts)
if cache_key in _folder_id_cache:
return _folder_id_cache[cache_key]
headers = _graph_headers()
parent_id = "Inbox"
for i, part in enumerate(path_parts):
partial_key = "/".join(path_parts[: i + 1])
if partial_key in _folder_id_cache:
parent_id = _folder_id_cache[partial_key]
continue
# List children of parent
if parent_id == "Inbox":
url = f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/Inbox/childFolders"
else:
url = f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/{parent_id}/childFolders"
r = http_requests.get(url, headers=headers, timeout=15)
if r.status_code == 401:
_get_graph_token()
headers = _graph_headers()
r = http_requests.get(url, headers=headers, timeout=15)
found = None
for f in r.json().get("value", []):
if f["displayName"].lower() == part.lower():
found = f["id"]
break
if not found:
# Create folder
cr = http_requests.post(url, headers=headers, json={"displayName": part}, timeout=15)
if cr.status_code in (200, 201):
found = cr.json()["id"]
elif cr.status_code == 409:
# Already exists (race condition) — re-fetch
r2 = http_requests.get(url, headers=headers, timeout=15)
for f in r2.json().get("value", []):
if f["displayName"].lower() == part.lower():
found = f["id"]
break
if not found:
raise RuntimeError(f"Cannot create folder '{part}': {cr.text}")
_folder_id_cache[partial_key] = found
parent_id = found
return parent_id
def _map_jnj_folder(folder: str) -> list[str]:
"""Map JNJ folder path to Graph folder parts under JNJ root.
'/vbuzalka@its.jnj.com/Inbox/TMP' → ['JNJ', 'Inbox', 'TMP']
'/Online Archive - vbuzalka@its.jnj.com/Inbox' → ['JNJ', 'Online Archive', 'Inbox']
"""
parts = [p for p in folder.split("/") if p]
if not parts:
return [GRAPH_ROOT_FOLDER]
# First part is mailbox name — strip it but detect Online Archive
mailbox = parts[0]
rest = parts[1:]
prefix = [GRAPH_ROOT_FOLDER]
if "online archive" in mailbox.lower():
prefix.append("Online Archive")
return prefix + rest if rest else prefix
def _make_recipient(addr: str) -> dict:
if "<" in addr and ">" in addr:
name = addr[: addr.index("<")].strip().strip('"')
email = addr[addr.index("<") + 1 : addr.index(">")].strip()
else:
name = addr
email = addr
return {"emailAddress": {"name": name, "address": email}}
def _import_msg_to_graph(msg_path: Path, folder: str) -> Optional[str]:
"""Parse .msg and import into Graph API mailbox. Returns message ID or None."""
try:
msg = extract_msg.Message(str(msg_path))
subject = msg.subject or "(no subject)"
body_html = msg.htmlBody
if isinstance(body_html, bytes):
body_html = body_html.decode("utf-8", errors="replace")
body_text = msg.body or ""
sender_email = msg.sender or ""
sender_name = getattr(msg, "senderName", None) or sender_email
to_raw = msg.to or ""
cc_raw = msg.cc or ""
date_raw = msg.date
att_list = []
for att in msg.attachments:
if att.data and att.longFilename:
att_list.append({
"@odata.type": "#microsoft.graph.fileAttachment",
"name": att.longFilename,
"contentType": getattr(att, "mimetype", None) or "application/octet-stream",
"contentBytes": base64.b64encode(att.data).decode(),
})
msg.close()
to_list = [a.strip() for a in to_raw.split(";") if a.strip()]
cc_list = [a.strip() for a in cc_raw.split(";") if a.strip()]
# Map folder and ensure it exists
folder_parts = _map_jnj_folder(folder)
folder_id = _ensure_folder(folder_parts)
payload = {
"subject": subject,
"body": {
"contentType": "HTML" if body_html else "Text",
"content": body_html or body_text,
},
"from": _make_recipient(f"{sender_name} <{sender_email}>"),
"toRecipients": [_make_recipient(a) for a in to_list],
"ccRecipients": [_make_recipient(a) for a in cc_list],
"isRead": True,
"singleValueExtendedProperties": [
{"id": "Integer 0x0E07", "value": "1"}
],
}
if date_raw:
try:
dt = dtparser.parse(str(date_raw))
payload["receivedDateTime"] = dt.astimezone(timezone.utc).strftime(
"%Y-%m-%dT%H:%M:%SZ"
)
except Exception:
pass
if att_list:
payload["attachments"] = att_list
headers = _graph_headers()
url = f"{GRAPH_URL}/users/{GRAPH_MAILBOX}/mailFolders/{folder_id}/messages"
r = http_requests.post(url, headers=headers, json=payload, timeout=30)
if r.status_code == 401:
_get_graph_token()
headers = _graph_headers()
r = http_requests.post(url, headers=headers, json=payload, timeout=30)
if r.status_code in (200, 201):
msg_id = r.json().get("id", "")
log.info("Graph OK: %s → %s", subject[:60], "/".join(folder_parts))
return msg_id
else:
log.error("Graph FAIL [%d]: %s | %s", r.status_code, subject[:60], r.text[:200])
return None
except Exception as e:
log.error("Graph import error for %s: %s", msg_path.name, e)
return None
@app.post("/upload")
async def upload_msg(
file: UploadFile = File(...),
authorization: str = Header(None)
authorization: str = Header(None),
folder: str = Form(""),
):
if authorization != f"Bearer {TOKEN}":
raise HTTPException(status_code=401, detail="Unauthorized")
@@ -39,7 +252,17 @@ async def upload_msg(
return {"status": "exists", "file": file.filename}
with dest.open("wb") as f:
shutil.copyfileobj(file.file, f)
return {"status": "saved", "file": file.filename}
# Import to Graph API if folder was provided by client
graph_id = None
if folder:
graph_id = _import_msg_to_graph(dest, folder)
return {
"status": "saved",
"file": file.filename,
"graph_imported": graph_id is not None,
}
@app.post("/upload-db")